<%BANNER%>

Late Pleistocene and Holocene Climate Change in the Maya Lowlands

Permanent Link: http://ufdc.ufl.edu/UFE0041371/00001

Material Information

Title: Late Pleistocene and Holocene Climate Change in the Maya Lowlands
Physical Description: 1 online resource (135 p.)
Language: english
Creator: Escobar, Jaime
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: deglacial, guatelama, isotopes, lgm, maya, mis3, paleoclimate, paleolimnology, peten, petenitza
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre: Interdisciplinary Ecology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Late Pleistocene and Holocene Climate Change in the Maya Lowlands I present the first continuous, well-dated, high resolution (~decadal), continental isotope (?18O) record from the northern, lowland Neotropics that spans the Last Glacial Maximum and last deglaciation (25.0 - 10.0 cal ka BP). The record comes from measurements on ostracod shells (Limnocythere sp.) in a lake sediment sequence from deep Lake Pete acuten Itza acute, Guatemala. Results show a transition from the Last Glacial Maximum (LGM) to the deglacial period similar to paleoclimate reconstructions from western North America, with a cold, wet LGM and a cold and/or dry early deglacial. A wetter LGM in the lowlands of Central America might be explained by (1) greater summer precipitation due to a more northerly position of the Intertropical Converge Zone (ITCZ), or (2) an increase in winter precipitation due to the splitting of the jet stream and/or movement of moist eastern tropical Pacific air to the northeast. ?18O results during the deglacial period show a link between North Atlantic and tropical Atlantic climate. Periods of reduced Atlantic Meridional Overturning Circulation (AMOC) and cool sea surface temperatures in the North Atlantic coincide with a more southerly position of the Intertropical Convergence Zone (ITCZ), and cold and/or dry climate in the lowlands of northern Central America. The deglacial climate of lowland Central America, reconstructed from the Pete acuten Itza acute sediment sequence, differs from climate reconstructions derived from ice and marine sediment cores, which show a ?two-step? deglaciation. The Pete acuten Itza acute oxygen isotope record displays three large, abrupt climate transitions, at ~16.5-16.3 cal ka BP, ~15.0-14.4 cal ka BP, and ~10.6-10.3 cal ka BP. Individual ?18O measurements on C. ilosvayi in sediments from Lake Punta Laguna show that samples from core depths that have high mean ?18O values, indicative of low effective moisture, display lower variability, whereas samples with low mean ?18O values, reflecting times of higher effective moisture, display higher variability. Relatively dry periods were thus consistently dry, whereas relatively wet periods had both wet and dry years. This interpretation of data from the cores applies to an important period of the late Holocene, the Maya Terminal Classic. ?18O variability during the ancient Maya Terminal Classic Period (ca. 910-990 AD) indicates not only the driest mean conditions in the last 3,000 years, but consistently dry climate. Variability of ?13C measurements in single stratigraphic layers displayed no relationship with climate conditions inferred from ?18O measurements.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jaime Escobar.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Brenner, Mark.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041371:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041371/00001

Material Information

Title: Late Pleistocene and Holocene Climate Change in the Maya Lowlands
Physical Description: 1 online resource (135 p.)
Language: english
Creator: Escobar, Jaime
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: deglacial, guatelama, isotopes, lgm, maya, mis3, paleoclimate, paleolimnology, peten, petenitza
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre: Interdisciplinary Ecology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Late Pleistocene and Holocene Climate Change in the Maya Lowlands I present the first continuous, well-dated, high resolution (~decadal), continental isotope (?18O) record from the northern, lowland Neotropics that spans the Last Glacial Maximum and last deglaciation (25.0 - 10.0 cal ka BP). The record comes from measurements on ostracod shells (Limnocythere sp.) in a lake sediment sequence from deep Lake Pete acuten Itza acute, Guatemala. Results show a transition from the Last Glacial Maximum (LGM) to the deglacial period similar to paleoclimate reconstructions from western North America, with a cold, wet LGM and a cold and/or dry early deglacial. A wetter LGM in the lowlands of Central America might be explained by (1) greater summer precipitation due to a more northerly position of the Intertropical Converge Zone (ITCZ), or (2) an increase in winter precipitation due to the splitting of the jet stream and/or movement of moist eastern tropical Pacific air to the northeast. ?18O results during the deglacial period show a link between North Atlantic and tropical Atlantic climate. Periods of reduced Atlantic Meridional Overturning Circulation (AMOC) and cool sea surface temperatures in the North Atlantic coincide with a more southerly position of the Intertropical Convergence Zone (ITCZ), and cold and/or dry climate in the lowlands of northern Central America. The deglacial climate of lowland Central America, reconstructed from the Pete acuten Itza acute sediment sequence, differs from climate reconstructions derived from ice and marine sediment cores, which show a ?two-step? deglaciation. The Pete acuten Itza acute oxygen isotope record displays three large, abrupt climate transitions, at ~16.5-16.3 cal ka BP, ~15.0-14.4 cal ka BP, and ~10.6-10.3 cal ka BP. Individual ?18O measurements on C. ilosvayi in sediments from Lake Punta Laguna show that samples from core depths that have high mean ?18O values, indicative of low effective moisture, display lower variability, whereas samples with low mean ?18O values, reflecting times of higher effective moisture, display higher variability. Relatively dry periods were thus consistently dry, whereas relatively wet periods had both wet and dry years. This interpretation of data from the cores applies to an important period of the late Holocene, the Maya Terminal Classic. ?18O variability during the ancient Maya Terminal Classic Period (ca. 910-990 AD) indicates not only the driest mean conditions in the last 3,000 years, but consistently dry climate. Variability of ?13C measurements in single stratigraphic layers displayed no relationship with climate conditions inferred from ?18O measurements.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jaime Escobar.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Brenner, Mark.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041371:00001


This item has the following downloads:


Full Text

PAGE 1

1 LATE PLEISTOCENE AND HOLOCENE CLIMATE CHANGE IN THE MAYA LOWLANDS By JAIME ESCOBAR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Jaime Escobar

PAGE 3

3 To Natalia Hoyos, Antonia Escobar, and my family in Colombia

PAGE 4

4 ACKNOWLEDGMENTS First and foremost I would like to thank Dr. Mark Brenner. I had the opportunity to work w ith the best advisor. His experience and care for students go far beyond academia into the personal qualities that make him a true mentor and friend. I will d efinitively miss our trips to "Hear Again" to buy 3 CDs for 10 bucks Many thanks go to Dr. David Hodell for providing critical advice on every aspect of this dissertation and for his funny comments such as "J ust maybe the benthic forams are lying to us ". Special thanks also go to the other members of my committee, Dr. Ellen Martin and Dr. Thomas Whitm ore, for their assistance and constructive suggestions towards my dissertation. I thank Gianna Browne and Dr. Jason Curtis for their assistance in the Light Stable Isotope Mass Spectrometry laboratory and the Land Use and Environmental Change Institute's P aleoenvironmental Research Laboratory. I thank the people at the School of Natural Resources and Environment, specially Dr. Stephen R. Humphrey Meisha S. Wade, and Cathy L. Ritchie. I also thank professors, staff and students from the Department of Geolog ical Sciences. I am extremely grateful to the many agencies and individuals in Guatemala who assisted this project (Chapters 2 and 3) They include: Universidad del Valle, Universidad San Carlos, Ministerio de Ambiente y Recursos Naturales, Consejo Nacion al de Areas Protegidas, Instituto de Antropolog’a e Historia, Autoridad Para el Manejo y Desarrollo Sostenible de la Cuenca del Lago PetŽn Itz‡, Wildlife Conservation Society, Alex Arrivillaga, Cathy Lopez, Margaret Dix, Michael Dix, Margarita Palmieri, Da vid, Rosita, & Kelsey Kuhn, and the staff at La Casa de Don David I also thank our many collab orators from University of Minnesota (Minneapolis/Duluth), Geoforschungszentrum (Potsdam), Swiss Federal Institute of Technology (Zurich),

PAGE 5

5 UniversitŽ de Genve, as well as the personnel of DOSECC (Drilling, Observation and Sampling of the Earth's Continental Crust). This project was funded by grants from the US National Science Foundation, the International Continental Scientific Drilling Program, the Swiss Feder al Institute of Technology, and the Swiss N ational Science Foundation. Coring on Lakes Punta Laguna and Chichancanab (Chapter 4) was funded by NOAA Grant NA36GP0304. Coring at Lake PetŽn Itz‡ (Chapter 4) was supported by NSF grant EAR 9709314 and a grant f rom the National Geographic Society. I received support from the University of Florida School of Natural Resources and Environment, Gulf Coast Association of Geological Societies Student Grant Program, the DOSECC Internship for Student Research, and a Geol ogical Society of America Graduate Student Research Grant. I thank all the friends and family members who supported me during this great endeavor. Friends David Buck, Elli Harrison Buck, Eliza Buck, Pablo Solano, Diana "Tita" Alvira, and Simon Solano welco med me into their homes where I spent some of the most enjoyable moments while pursuing my PhD degree. I also thank the "Parce" family, Cathy Venz Curtis and the "girls", Susan Milbrath, Susanna Blair, William Kenney, "Cocoya", and Carlos Jaramillo. Thanks go to my father, Jaime Escobar, my mother, Gloria Jaramillo and my sister Maria Teresa Escobar for always being there to share their strength and love. Last but certainly not least, I would like to give infinite thanks to my wife, Natalia Hoyos and daught er, Antonia Escobar, who represents all the good things worth looking for in life.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ... 4 LIST OF TABLES ................................ ................................ ................................ ............. 8 LIST OF FI GURES ................................ ................................ ................................ ........... 9 ABSTRACT. ...... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ..... 15 The PetŽn Itz‡ Scientific Drilling Project: A Historical Perspective .......................... 15 The Lake PetŽn I tz‡ Scientific Drilling Project (PISDP) ................................ ........... 17 Lake PetŽn Itz‡'s Modern Limnology and Regional Climate ................................ ... 18 Paleoclimate Proxies in Lake PetŽn Itz‡ Sediments ................................ ............... 19 Lithology ................................ ................................ ................................ ............ 19 Stable Isotopes of Oxygen ................................ ................................ ................ 20 Chronology ................................ ................................ ................................ .............. 23 A 45 ka Climate Record from Lake PetŽn Itz‡ ................................ ........................ 25 Statement of Objectives ................................ ................................ ........................... 26 Dissertation Organization ................................ ................................ ........................ 27 2 CLIMATE CONDITIONS IN THE LOWLANDS OF CENTRAL AMERICA DURING THE LAST GLACIAL MAXIUM (LGM) ................................ ...................... 42 Introduction ................................ ................................ ................................ .............. 42 Study Site ................................ ................................ ................................ ................ 44 Methods ................................ ................................ ................................ ................... 45 Results ................................ ................................ ................................ ..................... 46 Discussion ................................ ................................ ................................ ............... 47 Conclusions ................................ ................................ ................................ ............. 54 3 CLIMATE CHANGE IN THE NORTHERN NEOTROPICAL LOWLANDS DURING THE LAST DEGLACIATION ................................ ................................ ..... 59 Introduction ................................ ................................ ................................ .............. 59 Study Site ................................ ................................ ................................ ................ 61 Field and Laboratory Methods ................................ ................................ ................. 62 Proxy Interpretation: Oxygen Isotopes ................................ ................................ .... 63 Core Chronology ................................ ................................ ................................ ...... 64 Results ................................ ................................ ................................ ..................... 65 Climate Reconstruction for the Last Deglaciation ................................ .................... 66 Conclusions ................................ ................................ ................................ ............. 76

PAGE 7

7 4 ISOTOPE MEASUREMENTS OF SINGLE OSTRACOD VALVES AND GASTROPOD SHELLS FOR CLIMATE RECONSTRUCTION: EVALUATION OF WITHIN SAMPLE VARIABILITY AND DETERMINATION OF OPTIMUM SAMPLE SIZE ................................ ................................ ................................ ......... 85 Introduction ................................ ................................ ................................ .............. 85 Study S ites ................................ ................................ ................................ ............... 90 Laboratory Methods ................................ ................................ ................................ 90 Results ................................ ................................ ................................ ..................... 92 Discussion ................................ ................................ ................................ ............... 94 Within Horizon Variability and Calculation of Optimum Sample Size ................ 94 18 O Variability among Individuals within a Sample as a Proxy for High Frequency Climate Shifts ................................ ................................ ............... 99 Conclusions ................................ ................................ ................................ ........... 105 5 CONCLUSIONS ................................ ................................ ................................ .... 120 Late Pleistocene and Holocene Cli mate Change in the Maya Lowlands: An Overview ................................ ................................ ................................ ............. 120 LIST OF REFERENCES ................................ ................................ .............................. 123 BIOGRAPHICAL SKETCH ................................ ................................ ........................... 135

PAGE 8

8 LIST OF TABLES Table page 1 1 Location of drill sites, water depth, penetration, and core recovery at Lake PetŽn Itz‡. Modified from Hodell et al. (2006). ................................ .................... 31 1 2 Radiocarbon dates on terrestrial organic matter from Sites PI 2, PI 3, PI 6. ...... 32 1 3 Tie points used on the magnetic susceptibility records to project dated intervals from site PI 2 onto site PI 6. ................................ ................................ 34 1 4 Age depth relations for the Site PI 6 composite core used to derive the chronolog y discussed in Chapters 2 and 3. ................................ ........................ 35 3 1 Age depth points used to derive chronology shown in Figure 3 2. ..................... 78 4 1 Morphometric characteristics for the study lakes. ................................ ............. 107 4 2 Summary statistics for single valve ostracod and single valve gastropod 18 O datasets from Lake Punta Laguna. cmcd: centimeters core depth, SD: standard deviation, var: variance, *variance between P. coronatus and C. ilosvayi is significantly different ( "=0.05). ................................ .......................... 108 4 3 Summary statistics for single valve ostracod and single shell gastropod 13 C datasets from Lake Punta Laguna. cmcd: centimeters core depth, SD: standard deviation, var: variance, *variance between P. coronatu s and C. ilosvayi are significantly different ("=0.05). ................................ ....................... 109 4 4 Summary statistics for single valve ostracod and single shell gastropod 18 O datasets from Lake Chichancanab. cmcd: centimeters core depth, SD: standard deviation var: variance, *variance between P. coronatus and C. ilosvayi are significantly different ("=0.05). ................................ ....................... 110 4 5 Summary statistics for single valve ostracod and single shell gastropod 13 C datasets from Lake Chichancanab. cmcd: centimeters core depth, SD: standard deviation, var: variance, *variance between P. coronatus and C. ilosvayi are significantly different ("=0.05). ................................ ....................... 111 4 6 Summary statistics for single shell gastropod 18 O and 13 C datasets f rom Lake PetŽn Itz‡. cmcd: centimeters core depth, SD: standard deviation, var: variance. ................................ ................................ ................................ ............ 112

PAGE 9

9 LIST OF FIGURES Figure page 1 1 (A) Map of Central America showing the Location of the Yucatan Peninsula and the department of PetŽn, Guatemala. (b) Map of the PetŽn Lake District. (C). Bathymetric map of Lake PetŽn Itz‡ showing the location of primary and alternate coring sites. From Hodell et al. (2008). ................................ ................ 36 1 2 Calibrated radiocarbon dates versus meters composite depth (mcd) for sites PI 3 (blue) and PI 6 (red) used to construct the age model used by Hodell et al. (2008). ................................ ................................ ................................ ............ 37 1 3 Calibrated radiocarbon dates versus meters composite depth (mcd) for sites PI 3 (blue) and PI 6 (red) used to construct the age model used by Hodell et al. (2008) for the Last Glacial Maximum (LGM) and deglacial period. ................ 38 1 4 Spliced magnetic susceptibility records from Sites P I 6 and PI 2 for the Last Glacial Maximum (LGM) and deglacial period. Green triangles indicate tie points used to project dated intervals from site PI 2 onto site PI 6. Yellow dots indicate dated intervals on site PI 2. ................................ ............................ 39 1 5 Magnetic susceptibility (red) and 18 O (blue) in sediment core PI 6 from 50.0 to 10.0 meters composite depth (mcd) ................................ ............................... 40 1 6 Oxygen isotope data for Lake PetŽn Itz‡ for the last 45 cal ka. 18 O data indicat ed by blue circles ( Limnocythere sp.) are from PI 6. 18 O data indicated by red circles ( Limnocythere sp.) are from Lake PetŽn Itz‡ core 11A (Hillesheim et al. 2005). 18 O data indicated by purple circles (Cochliopina sp.), green circles ( Pyrgophorus sp.), brown circles ( Cytheridella ilosvayi ) and black circles ( Candona sp.) are from Lake PetŽn Itz‡ core 6 VII 93 (Curtis et al. 1998). Each record is plotted on its own, independent time scale. ................................ ................................ ...................... 41 2 1 (A) Map showing the location s of Lake PetŽn Itz‡, northern Guatemala, and paleoclimate reconstruction sites discussed in this paper. Cariaco (Lea et al. 2003 and Peterson et al. 2000), OCE 326 GGC5 (McManus et al. 2004), SU 8118 (Bard 2002), MD95 2043 (Cacho et al. 1999) (B) Map show ing the locations of Lake PetŽn Itz‡ and paleoglacier reconstructions discussed in this paper. (C) Map of Lake PetŽn Itz‡ showing the location of drilling sites PI 2, PI 3, and PI 6. ................................ ................................ ............................. 55 2 2 (A) Calibrated radiocarbon dates ver sus meters composite depth (mcd) for sites PI 3 (blue) and PI 6 (red) used to construct the age model used by Hodell et al. (2008). (B) Calibrated radiocarbon dates versus mcd for sites PI 3 (blue), PI 6 (red), and PI 2 (yellow) used to construct the age m odel used in this paper. Age model was derived by linear interpolation between selected age depth points. ................................ ................................ ................. 56

PAGE 10

10 2 3 Magnetic susceptibility (red) and 18 O (blue) in sediment core PI 6. Light gray band represents the Last Glacial Maximum (23.0 19.0 cal ka BP) as define by Mix et al. (2001). ................................ ................................ .................. 57 2 4 Summary map (modified from Bradbury 1997) of moisture conditions during the Last Glacial Maximum in the southwestern United States, Central Mexico and nort hern South America. Dashed line represents the wet dry longitudinal boundary proposed by Bradbury (1997). Black line represents the new wet dry longitudinal boundary proposed based on results from this study. ............... 58 3 1 (A) Map sh owing the locations of Lake PetŽn Itz‡, northern Guatemala, and paleoclimate reconstruction sites discussed in this paper: Cariaco (Lea et al. 2003 and Peterson et al. 2000), VM28 122 (Schmidt et al. 2004), OCE 326 GGC5 (McManus et al. 2004), SU 8118 (Bard 2002), NGRIP (North Greenland Ice Core Project members 2004), and Dome Concordia (Dome C) (Monnin et al. 2001). (B) Bathymetric map of Lake PetŽn Itz‡ showing the location of drill sites PI 2, PI 3, and PI 6. ................................ ............................ 79 3 2 (A) Calib rated radiocarbon dates versus mcd for sites PI 3 (blue) and PI 6 (red) used to construct the age model used by Hodell et al. (2008). (B) Calibrated radiocarbon dates versus meters composite depth (mcd) for sites PI 3 (blue), PI 6 (red), and PI 2 (yellow) used to construct the age model used in this paper. Age model was derived by linear interpolation between selected age depth points. ................................ ................................ ................. 80 3 3 Magnetic susceptibility (red) and 18 O (blue) in sediment core PI 6. 18 O data indicated by open circles are from Lake PetŽn Itz‡ core 11A (Hillesheim et al. 2005). Cores 11A and PI 6 were correlated using an ash layer in both cores at the Pleistocene Holo cene transition. Dated climate periods are based on Greenland Ice Core Chronology 2005 (GICC05) (Rasmussen et al. 2006). 2bk: ages relative to A.D. 2000. ................................ ................................ .......... 81 3 4 18 O in sediment core PI 6 and core 11A (blue). 18 O data indicat ed by open circles are from Lake PetŽn Itz‡ core 11A (Hillesheim et al. 2005) and are compared to the 18 O of the NGRIP (North Greenland Ice Core Project members 2004) (green) and sea surface temperatures derived from alkenones in core SU8118 from the subt ropical NE Atlantic (Bard 2002) (purple). Each record is plotted on its own, independent time scale. Dated climate periods are based on Greenland Ice Core Chronology 2005 (GICC05) (Rasmussen et al. 2006). 2bk: ages relative to A.D. 2000. ................. 82 3 5 18 O in sediment core PI 6 and core 11A (blue). 18 O data indicated by open circles are from Lake PetŽn Itz‡ core 11A (Hillesheim et al. 2005) and are compared to sea surface temperature reconstructions (Lea et al. 2003) (orange) and color reflectance (Pet erson et al. 2000) (brown) from the Cariaco Basin. Each record is plotted on its own, independent time scale.

PAGE 11

11 Dated climate periods are based on Greenland Ice Core Chronology 2005 (GICC05) (Rasmussen et al. 2006). 2bk: ages relative to A.D. 2000. ................. 83 3 6 18 O of the NGRIP (North Greenland Ice Core Project members 2004) (green). 18 O in sediment core PI 6 (blue). Color reflectance (Peterson et al. 2000) (brown) from the Cariaco Basin. Methane data (grey) from Dome C, Antarctica (Monnin et al. 2001). Each re cord is plotted on its own, independent time scale. Light gray band represents the mystery Interval (17.5 14.5 cal ka) as define by Denton et al. (2006). Dark grey bands represent ice armada events at ~17.5 and ~16.1 cal ka BP reported by Bard et al. (20 00). ................................ ................................ ................................ ........ 84 4 1 Map of study area showing the location of the three study lakes. ..................... 113 4 2 Optimum sample size (i.e. number of ostracod valves or gastropod shells) in Lake Punta Laguna (squares) and Lak e Chichancanab (circles) for P. coronatus (open symbol) and C. ilosvayi (closed symbol) for (a) 18 O and (b) 13 C measurements. ................................ ................................ ......................... 114 4 3 Intra sample variability as a proxy for high resolution climate reconstructions. (3a) 18 O values from composite samples, reported by Curtis et al. (1996) vs. average 18 O values calculated from measurements made on multiple, individual valves in this study. Dashed line indicates a 1 1 line. r = 0.95 (3b) 13 C values from composite samples, rep orted by Curtis et al. (1996) vs. average 13 C values calculated from measurements made on multiple, individual valves in this study. Dashed line indicates a 1 1 line. r = 0.94. ......... 115 4 4 (A) Oxygen isotopic composition of the o stracod C. ilosvayi in the sediment core from Lake Punta Laguna, Mexico (Curtis et al. 1996). The oxygen isotope data were not smoothed as in Curtis et al. (1996). Horizontal lines at several stratigraphic horizons represent the range of 18 # measurements on individual C. ilosvayi valves. Dates are based on the chronological model constructed on samples obtained from the sediment core (PL 22 VI 93) recovered in 1993 in 6.3 m of water column from the central basin of Punta Laguna (Curtis et al. 1996). (B) De tail of the Classic Maya collapse time period. Horizontal lines at several stratigraphic horizons represent the range of 18 O measurements on individual C. ilosvayi valves. (C) Detail of the Little Ice Age period. Horizontal lines at several stratigraphic horizons represent the range of 18 O measurements on individual C. ilosvayi valves. ................... 116 4 5 Intra sample variability as a proxy for high resolution climate reconstructions. (5a) mean 18 O values vs. standard deviation of 18 O values for C. ilosvayi in Lake Punta Laguna. (5b) mean 18 O values vs. standard deviation of 18 O values for C. ilosvayi in Lake Punta Laguna for the wet and dry periods during the time of the Classic Maya collapse. Numbers in the figure indicate years AD/ BC. Dates are based on the chronological model constructed on samples obtained from the sediment core (PL 22 VI 93) recovered in 1993 in

PAGE 12

12 6.3 m of water column from the central basin of Punta Laguna (Curtis et al. 1996). ................................ ................................ ................................ ................ 117 4 6 Number of individual ostracod valves and gastropod shells analyzed vs. 18 O (closed circle) and 13 C (open circle) standard deviation per stratigraphic level. (A) C. ilosvayi in Lake Punta Laguna. (B) P. coronatus in Lake Punta Laguna. (C) C. ilosvayi in L ake Chichancanab. (D) P. coronatus in Lake Chichancanab. ................................ ................................ ................................ .. 118 4 7 Sedimentation rates vs. 18 O (closed circles) and 13 C (open circles) standard deviation per stratigraphic level. (A) C. ilosvayi in Lake Punta Laguna. (B) P. co ronatus in Lake Punta Laguna. (C) C. ilosvayi in Lake Chichancanab. (D) P. coronatus in Lake Chichancanab. ................................ .. 119

PAGE 13

13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment o f the Requirements for the Degree of Doctor of Philosophy LATE PLEISTOCENE AND HOLOCENE CLIMATE CHANGE IN THE MAYA LOWLANDS By Jaime Escobar May 2010 Chair: Mark Brenner Major: Interdisciplinary Ecology I present the first continuous, well dated, high resolution (~decadal), continental isotope (! 18 O) record from the northern, lowland Neotropics that spans the Last Glacial Maximum and last deglaciation (25.0 10.0 cal ka BP). The record comes from measurements on ostracod shells ( Limnocythere sp.) in a lake sediment sequence from deep L ake PetŽn Itz‡, Guatemala. R esults show a transition from the Last Glacial Maximum (LGM) to the deglacial period similar to paleoclimate reconstructions from western North America, with a cold, wet LGM and a cold and/or dr y early deglacial. A wetter LGM in the lowlands of Central America might be explained by (1) greater summer precipitation due to a more northerly position of the Intertropical Converge Zone (ITCZ), or (2) an increase in winter precipitation due to the spli tting of the jet stream and/or movement of moist eastern tropical Pacific air to the northeast. 18 O results during the deglacial period show a link between North Atlantic and tropical Atlantic climate. Periods of reduced Atlantic Meridional Overturning Ci rculation (AMOC) and cool sea surface temperatures in the North Atlantic coincide with a more southerly position of the Intertropical Convergence Zone (ITCZ), and cold and/or dry climate in the lowlands of northern Central America. The deglacial climate of lowland Central America,

PAGE 14

14 reconstructed from the PetŽn Itz‡ sediment sequence, differs from climate reconstructions derived from ice and marine sediment cores, which show a "two step" deglaciation. The PetŽn Itz‡ oxygen isotope record displays three large, abrupt climate transitions, at ~16.5 16.3 cal ka BP, ~15.0 14.4 cal k a BP, and ~10.6 10.3 cal ka BP. Individual 18 O measurements on C. ilosvayi in sediments from Lake Punta Laguna show that samples from core depths that have high mean 18 O values, indica tive of low effective moisture, display lower variability, whereas samples with low mean 18 O values, reflecting times of higher effective moisture, display higher variability. Relatively dry periods were thus consistently dry, whereas relatively wet perio ds had both wet and dry years. This interpretation of dat a from the cores applies to an important period of the late Holocene, the Maya Terminal Class ic 18 O variability during the ancient Maya Terminal Classic Period (ca. 910 990 AD) indicates not only t he driest mean conditions in the last 3,000 years, but consistently dry climate. Variability of 13 C measurements in single stratigraphic layers displayed no relationship with climate conditions inferred from 18 O measurements.

PAGE 15

15 CHAPTER 1 INTRODUCTION The PetŽn Itz‡ Scientific Drilling Project: A Historical Perspective In the early 1970s, Professor Edward S. Deevey and colleagues at the University of Florida (UF) initiated the "Central PetŽn His torical Ecology Project" This pioneering project used archaeo logical and paleolimnological data from selected watersheds in the PetŽn region of northern Guatemala to investigate relationships between ancient Maya culture and the environment. Subsequently, the UF group extended paleoecological and paleoclimate invest igations into other areas of the Caribbean region. In 1985, a 7.5 m sediment core was collected from Lake Miragoane, southern Haiti. Analysis of stable oxygen isotopes and trace metals in this core yielded a high resolution Holocene climat e reconstruction (Hodell et al. 1991). In 1993, a regional paleolimnological project was undertaken to investigate high resolution Holocene climate variability in continental areas arou nd the Caribbean (Curtis et al. 2001; Hodell et al. 2000). One of the water bodies selec ted for coring was Lake PetŽn Itz‡, in the heart of the PetŽn Lake District (Fig ure 1 1 ). A 5.45 m sediment core was collected in 7 m of water in the southern basin of t he lake (Curtis et al. 1998). This sediment core provided a high resolution climate rec onstruction for the last ~9 14 C ka. In 1997, a short core was taken from the southern basin o f the lake to study recent (i.e. the last 100 years) eutrophication (Rosenm eier et al. 2004). Most shallow lakes in the Yucatan Peninsula were dry during the last glacial period due to aridity and l owered sea level (Hodell et al. 2000). Thus, glacial age lacustrine sediments are only found in a few deep lakes. In 1977 and 1978, Professor Edward S. Deevey and his multidisciplinary team collected sediment cores up to ~10 m long, in

PAGE 16

16 deep Lakes Macanche and Quexil. Radiocarbon dates and pollen stratigraphy showed that the Kullenberg corer used to collect the cores had failed to penetrate into Pleistocene deposits, and basal sediments in the cores possessed high forest po llen. Pleistocene age deposits were finally obtained in 1980 from deep Lakes Quexil and Salpeten with the help of a Guatemalan drilling company, Daho Pozos, S.A. Late Pleistocene sediments from both lakes contain abundant gypsum, shells, and sponge spicule s, which indicate deposition in moderately salin e, shallow water bodies (Deevey et al. 1983). Geochemical, isotopic, and pollen analysis on the core from Lake Quexil (Leyden et al. 1993), suggested that the last deglaciation occurred in two distinct phases i.e. a "two step" deglaciation, as seen in records from high latitude regions. Th e first phase, starting at 14 14 C ka BP, was characterized by expansion of temperate oaks and low 18 O values that indicated warmer, more humid conditi ons than during the LG M (Leyden 1995). A second phase (12 ka 10.3 14 C ka BP), with evidence of calcite precipitation in sediments, a reversal of forest expansion, but continued low 18 O values, indicated a return to a colder (~1.5C), but sustained humid climate (Leyden 1995) This interpretation, however, was based on poor chronological control for the Lake Quexil core, which was likely influenced by hard wate r error associated with a date on a single snail shell. Recently, an international team of scie ntists targeted Lake Pe tŽn Itz‡ for deep drilling (Fig ure 1 1 ). Lake PetŽn Itz‡ is the largest and deepest lake in the PetŽn Lake District and thus held water during the driest periods of the late Pleistocene. In 1999 and 2002, seismic surveys were conducted in Lake PetŽn Itz‡ w ith a 3.5 kHz pinger

PAGE 17

17 and a 1 i n3 a irgun, respectively, to construct a bathymetric map of the lake and identify potential sites for d eep drilling (Anselmetti et al. 2006). Thirteen sediment cores were recovered at ten different sites in 2002 using a Kullenb erg type gravity corer. As in the 1977 and 1978 coring expedition at Lakes Quexil, Macanche, and Salpeten, glacial age deposits were not recovered from PetŽn Itz‡ because a dense clay rich sediment layer deposited during the late Hol ocene limited core bar rel penetration to ~6 m. The bottom of the oldest sediment core retrieved was dated at 11,250 cal years BP, and therefore included only the terminal part of the last deglaciation (Hillesheim et al. 2005). Collection of late Pleistocene sediments from the b asin would have to await drilling carried out under the Lake PetŽn Itz‡ Scientific Drilling Project (PIS DP). The Lake PetŽn Itz‡ Scientific Drilling Project (PISDP) The main objective of th e PISDP was to obtain complete l ate Pleistocene lacustrine sediment sequences to study: (1) the paleoclimate history of the northern lowland Neotropics on decadal to millennial timescales, (2) the paleoecology and biogeography of the tropical lowland forest, and (3) the subsurface biogeochemistry, including integrated stu dies of microbiology, porewater geochemistry, and mineral authigenesi s and diagenesis (Hodell et al. 2006). Between 3 February and 11 March 2006, 1327 m of sediment were retrieved from seven sites in Lake PetŽn Itz‡ (Table 1 1 ). In most cases, multiple hol es were drilled at each site. Drilling was done with the Global Lake Drilling GLAD800 platform (R/V Kerry Kelts), operated by the Drilling, Observation and Sampling of the Earth's Continental Crust (DOSECC) team. A minimum of three cores was drilled at si x of the seven sites to ensure full recovery of stratigraphic sections. Most of the sediment cores were taken with a hydraulic piston corer. Sediment cores were taken in 3 m sections and were

PAGE 18

18 logged in the field for density, p wave velocity, and magnetic s usceptibility. Complete sediment recovery was verified with the help of Splicer, a computer program that allows comparison of core logging data (i.e. density, m agnetic susceptibility) among different holes (Hodel l et al. 2006). A complete sediment sequenc e that is 75.9 meters long and represents about 85 ka of continuos sediment accumulation was recovered at site PI 6. Part of this section became the subject of study for this dissertation. Sediment cores were deposited at the National Lacustrine Core Repos itory (LacCore), University of Minnesota, Minneapolis, MN. At LacCore, sediment density and magnetic susceptibility were measured again, this time at 0.5 cm resolution using another GEOTEK Multisensor core logger. The composite stratigraphic section for si te PI 6 was constructed using the magnetic susceptibility and density resul ts of the three cores drilled at PI 6. Sediment cores comprising the composite section were subsampled and sediment was transported t o the University of Florida for stable isotope a nalysis. Lake PetŽn Itz‡'s Modern Limnology and Regional C limate Lake PetŽn Itz‡ is located at ~16¡ 55' N, 89¡ 50' W in the Department of PetŽn, northern Guatemala (Figure 1 1). It is located ~110 m above sea level and has a maximum depth of >160 m (Ansel metti et al. 2006). Water is dominated by the cations calcium and magnesium, and the anions bicarbonate and sulfate. Its waters are now saturated with calcium carbonate, and pH is high (~8.0). Gypsum deposits (calcium sulfate) in sediment cores indicate th at the water column was saturated with CaSO 4 in the past (Hillesheim et al. 2005). Rainfall and runoff are the main sources of water input to Lake PetŽn Itz‡, whereas water is lost mainly to evaporation. There can be some subterranean drainage at high lake stage, via an opening in the rock on the north shore,

PAGE 19

19 but the lake is effectively a closed basin system. Thermal stratification is persistent for much of the year, with the thermocline located between ~20 and 35 m (Anselmetti et al. 2006). The PetŽn regio n has an average air temperature of ~25 ¡C. Annual rainfall varies from 900 to 2500 mm/year, with a regional average of ~1601 mm/year (Deevey et al. 1980). The rainy season occurs between May and October when the trade winds carry moisture from the Atlanti c Ocean to the Caribbean Sea (Amador 1998; Mestas Nunez et al. 2007). The rainy season usually ends in late October and the dry season persists from January to May. During the dry season light rains generated by polar air masses are brought to the Yucatan Peninsula by northern winds known locally as nortes Precipitation patterns in the Caribbean are closely related to variability in the position of the Intertropical Convergence Zone (ITCZ). Years with high precipitation in the Caribbean are linked to a nor therly position of the ITCZ, while years with low summer rainfall are linked to a more southerly position of the ITCZ. Paleoclimate P roxies in Lake PetŽn Itz‡ Sediments Lithology PetŽn Itz‡'s sediments consist of a mixture of inorganic and organic matter o f autochthonous and allochthonous origin. Sediment organic matter produced within the lake comes from phytoplankton, higher plants, zooplankton, benthic organisms, and their fecal matter. Autochthonous inorganic matter comes in the form of calcite, dolomit e, and gypsum, precipitated from the water column. External sources of organic matter include leaves, twigs, and pollen grains, as well as organic matter in eroded surface soils. Clay, principally montmorrillonite, constitutes the major detrital input, b ut carbonates from the watershed may also reach the lake in colluvium and alluvium.

PAGE 20

20 Gypsum crystals are probably precipitated in the lake when evaporation rates are high and rainfall is relatively low. The presence of gypsum in sediments therefore indicate s dry periods in the past, when the volume of Lake PetŽn Itz‡ was low. In contrast, deposition of clay indicates wet periods and higher lake levels, when runoff and the influx of detrital material were enhanced. Variations in sediment magnetic susceptibili ty and density reflect changes in the sediment lithology. Magnetic susceptibility in PetŽn Itz‡'s sediments is generally low (<60 x10 6 SI), except for a number of peaks associated with layers of volcanic ash. Some volcanic eruptions in the highlands of Gu atemala and Mexico contain magnetite, which has high magnetic susceptibility. Magnetic susceptibility is relatively higher in clay sediment layers and lower (<0 x10 6 SI) in sediments rich in gypsum or calcite. The density of the sediment varies between ab out 1 and 2 g/cm 3 wet. Higher density values are associated with gypsum sands, whereas lower values are associated with clay rich sediments. Alternating clay and gypsum deposition, which is recorded by variations in the density and magnetic susceptibility of the sediment, therefore represent wet and dry climate episodes, respectively. Gypsum deposited during low water levels, i.e. relatively dry conditions, has low magnetic susceptibility and high density, whereas clay rich sediments deposited during high w ater levels, i.e. relatively wet conditions have relatively high magnetic susceptibility and low dens ity. St able Isotopes of O xygen Atoms that have the same atomic number, but different masses, are called isotopes (Petrucci 1997). Oxygen has three stable isotopes (O 16 O 17 O 18 ). They are "stable," because unlike radioisotopes, they undergo no "decay" and are destined to be

PAGE 21

21 the same element, with the same mass, forever. The behavior of an atom or molecule, as it enters physical, chemical, or biological pr ocesses, depends on its atomic mass. Mass differences between molecules cause them to react differently when they enter into physical, chemical, or biological processes For instance, lighter isotopes diffuse faster than heavier isotopes and are incorporat ed first from the liquid into the vapor phase (evaporation kinetic effect) (Faure 1986). The differential response of isotopes during physical, chemical, and biological transformations accounts for most of the isotopic "signal" recorded in the environment (e.g. in water, plants, and animal shells or soft tissues) that is used to interpret environmental processes. The isotopic signature of environmental materials is measured by light isotope ratio mass spectrometry. The 18 O value of precipitation depends, in part, on the isotope composition of the moisture source (e.g. sea water) and the extent of moisture depletion in the air mass as it is transported from the source area to t he rainfall region (Leng et al. 2006). There is progressive depletion of H 2 18 O wi th increasing latitude, altitude, distance from the coast, and rainfall amount (Dansgaard 1964). The relative humidity and temperature of the atmosphere also contribute to isotopic fractionation in precipitation, from the time water evaporates from the oce an to the moment a raindrop hits the ground. Several factors can alter lakewater isotopic composition in closed basin lakes where precipitation is the main water source. Among these factors are: (1) changes in the isotopic composition of the source water, either through a change in the isotope signature of the original source, or a change in the source region from which moisture is derived, (2) changes in the routing of moist air masses, and (3) changes in the relative

PAGE 22

22 humidity and/or temperature at any tim e from moisture formation to raindrop precipitation (Darling et al. 2006; Leng et al. 2006). In closed basin Lake PetŽn Itz‡, the relative amount of evaporation to precipitation (E/P) is the principal control on the lakewater isotopic signature, as lakewat er is enriched with respect to regional rainfall and groundwa ter. Mean lakewater 18 O (2.62: Curtis et al. 1998) is greater than both the mean value for regional groundwater ( 4) and annual weighted mean rainwater values in the circum Caribbean lowlands ( 5.7 to 4.0) (Rozanski et al. 1993), indicating substanti al evaporative loss of water from the lake. Periods of high E/P display relatively lower lake level and higher water column 18 O values. Conversely, during episodes of low E/P, the lake shows higher water level and displays lower 18 O values. Lachniet and Patterson (2009) investigated the variables that control surface water 18 O values in the Yucatan Peninsula. They analyzed 18 O of surface waters from Guatemala and Belize as well as data from the Global Network for Isotopes in Precipitation (GNIP) databas e. Spatially distance from the coast ( the "continentality effect") an d mean catchment altitude ( the "altitude effect") explain 84% of the surface water 18 O variability. In a given water body, the most important control on precipitation 18 O values throug h time was rainfall amount ( the "amount effect"). Lakewater 18 O, together with water temperature, largely determines the 18 O in the calcium carbonate of shell forming aquatic organisms such as ostracods, gastropods, and foraminifera. Colder temperatures yield greater 18 O values in precipitated carbonate, whereas warmer temperatures produce lower 18 O values. Thus a switch from higher 18 O values to lower 18 O values in ostracod shells could

PAGE 23

23 reflect a change in precipitation source and/or amount, as well as a change from colder and/or drier, to warmer and/or wetter conditions. Conversely, a switch from lower 18 O values to higher 18 O could reflect a change in precipitation source and/or amount, as well as a change to colder and/or drier conditions. Oxyge n isotope records from lake sediment cores are usually developed using monospecific ostracod or gastropod shells. This is done to avoid so called "vital effects." M ultiple individual monospecific ostracod s or gastropod specimens from each stratigraphic le vel are composited to comprise a sample. The rationale for analyzing multiple individuals in a sample is that it reduces variance that might otherwise result from measuring a single, short lived individual, and thus provides a reliable record of low freque ncy climate trends (Curtis et al. 1996; Hodell et al. 2005). Few studies have attempted to determine how many shells or valves of individual organisms in each stratigraphic interval are necessary to identify low frequency climate trends, i.e. to yield a re liable mean climate signal for a stratigraphic layer Nor has there been an effort to evaluate the potenti al for using intra sample 18 O variability in ostracod and gastropod shells as a proxy measure of high frequency climate variabil ity. Chronology The initial age/depth model for the PI 6 core from Lake PetŽn Itz‡ was derived using 18 AMS 14 C dates (Figure 1 2), including 10 AMS 14 C dates from the LGM and the deglacial period (Figure 1 3), run on organic terrestrial remains from cores taken at the PI 6 and PI 3 sites. Dates from the PI 3 core were projected onto the PI 6 core by correlating their magnetic sus ceptibility records using the computer program Splicer. Dates selected for the age/depth model minimized apparent sedimentation rate

PAGE 24

24 changes (Hodell, personal communication), but may have missed fluctuations in sedimentation rate that would be expected to accompany abrupt changes in lithology. After publication of the initial age/depth model (Hodell et al. 2008), new dates from drill sites PI 2 and PI 6 enabled development of a refined age/depth model for the LGM and deglacial period. The new core chronolog y was developed with a total of 52 samples from sites PI 2, PI 3, and PI 6, including 22 samples that covered the LGM and deglacial period (Table 1 2). Newly obtained dates from the PI 2 core were projected onto the PI 6 core (Table 1 3, Figure 1 4) using the same method of graphic correlation used by Hodell et al. (2008). Sample selection and calibration for developing a new age/depth model for the Last Glacial Maximum (LGM) and deglacial period were accomplished using the computer based calibration progra m Oxcal ( Bronk Ramsey 2001, 2008 ). Initially, the "U sequence" command was used to choose a set of dated depths that does not include age reversals. This command assumes a linear sedimentation rate between boundary depths that are arbitrarily set. Boundari es were set for the beginning and end of each lithologic unit (i.e. clay and gypsum sand units). There were very few dated depths from the gypsum sand units and thus the program did not run satisfactorily. The standard "sequence" command was then used on s amples from all dated depths within the LGM and deglacial period. This command specifies that samples were deposited in the sediment in the order that they were found, but does not take into account the specific depth of each sample. The program ran satisf actorily and 17 samples were selected for the new age/depth model (Table 1 4). A linear sedimentation rate was assumed

PAGE 25

25 between the selected dated intervals and an age was calculated for every 1 cm interval between dated horizons. Radiocarbon measurements w ere done at the Center for Accelerator Mass Spectrometry (CAMS) at Lawrence Livermore National Laboratory. All radiocarbon dates were measured on samples of terrestrial organic matter, thereby avoiding potential dating errors associated with hard water lak es (Deevey and Stuiver 1964). A 45 ka Climate R ecord from Lake PetŽn Itz‡ The core obtained from site PI 6 contains an 85 ka record of terrestrial climate from lowland Central America. Here, I focus on the last 45 ka of that record. Sediments are composed of alternating clay and gypsum strata, reflecting relatively wet and dry climate conditions, respectively. During the last 45 ka, increasing 18 O values coincide with gypsum sand units, whereas decreasing 18 O values coincide w ith clay deposition During the latter part of MIS 3 (~46.0 to ~26.0 meters composite depth [mcd]), climate varied between wetter and/or warmer c onditions during interstadials, and drier and/or colder states during stadials (Figure 1 5) The pattern of alternating clay gypsum (wet dry) and changes in oxygen isotope values during the latter part of MIS 3 closely resembles the temperature records from Greenland i ce cores and precipitation proxies from the Cariaco Basin. The most arid periods (i.e. gypsum sand units with relatively greater 18 O values) coincide with Heinrich Events in the northern North Atlantic. A thick clay unit during the LGM (~26.0 to ~21.0 mcd) indicates wet conditions and sediment deposition under high lake level conditions. A decrease in 18 O values at the onset of t his clay unit indicates a transition to wet and/or warm conditions, whereas the increase in 18 O values at the end of the same clay unit indicates a return to dry and/or cold conditions (Figure 1 5). During the deglacial period (~21.0 to ~11.0 mcd) climate

PAGE 26

26 varied between drier and/or colder and wetter and/or warmer conditions. The pattern of clay gypsum alternation and changes in oxygen isotope values show two dry and/or cold periods (Oldest and Younger Dryas) preceding and following a wet and/or warm inter val, the Bolling Allerod Period (Figure 1 5) Stable oxygen isotope data from a near complete Holocene sediment record taken from Lake PetŽn Itz‡ (Curtis et al. 1998), together with isotope data from this study, indicate that the largest change in 18 O values over the entire 45 ka record occurred at the Pleistocene/Holocene transition. Fluctuations in 18 O values from ~45 to 10 cal ka BP were small compared with the large change at the Pleistocene/Holocene transition. A decrease i n 18 O values of ~5.0 between the Pleistocene/Holocene transition and the middle Holocene indicates a switch to wetter and/or warmer conditions (Figure 1 6). Although there is a continuous 45 ka oxygen isotope record according to the Hodell et al. (2008) age model, there is not yet a refined age model that includes new dates from sites PI 2 and PI 6 for the MIS 3 period ( ~46.0 to ~26.0 mcd) Thus, this dissertation focuses on the Last Glacial Maximum and deglacial periods. Statement of O bjectives The objec tives of this dissertation were to (1) produce a high resolution paleoclimate record from 25 10 cal ka BP using stable oxygen isotope measurements on ostracod remains, allowing comparisons with paleoclimate records from higher and lower latitudes, and (2) to evaluate the potential for using intra sample 18 O variability in ostracod and gastropod shells as a proxy measure of high frequency climate variability. To meet these objectives, this study addresses several questions: 1. Were shifts in the position of the ITCZ a response to high latitude North Atlantic climate (i.e. Atlantic Meridional Overturning Circulation [AMOC})?

PAGE 27

27 2. Did climate (evaporation/precipitation) changes in PetŽn, Guatemala coincide with climate excursions recorded in Greenland ice cores during the Last Glacial Maximum and deglacial period? 3. How do climate changes in PetŽn compare with those inferred from the Cariaco Basin during the Last Glacial Maximum and deglacial period? 4. What is the optimum number of ostracod valves or snail s hells required to infer low frequency climate changes? 5. Is the intra sample 18 O variability in ostracod and gastropod shells useful as a proxy measure for high frequency climate variability? Dissertation O rganization This dissertation is presented as discrete publishable papers. Chapter 1 is the Introduction. A modified version of the Introduction, together with results from Hodell et al. (08) is in press as a book chapter, written in Spanish. Chapter 2 is titled "Climate conditions in the lowlands of Central America during the Last Glacial Maxium (LGM)" and together with Chapter 3 ("Climate change in the northern Neotropical lowlands during the last deglaciation") will be submitted to Quaternary Science Reviews Chapter 2 contributes to the understandi ng of climate conditions during the LGM in the northern Neotropics. Results provide evidence for a wetter LGM, bracketed by two drier periods associated with Heinrich events. Wetter conditions recorded in the Lake PetŽn Itz‡ isotope record during the LGM c oincide with paleoglacier data from Mexico and Guatemala. Glacial advances down to ~3,400 meters above sea level (masl) in Mexico and Guatemala suggest that the region was cooler and moister than present. Retreat of glaciers in the region occurred some tim e after the increase in Lake PetŽn Itz‡ 18 O values, which indicates a transition to drier conditions. Wetter conditions during the LGM in the lowlands of Central America may be explained by (1) an increase in summer precipitation due to a more northerly p osition of the Intertropical Converge Zone (ITCZ),

PAGE 28

28 or (2) an increase in winter precipitation due to the splitting of the jet stream and/or movement of moist eastern tropical Pacific air to the northeast. Chapter 3, "Climate change in the northern Neotropi cal lowlands during the last deglaciation" contributes to the understanding of climate change in the lowlands of the northern Neotropics during the last deglaciation 18 10 cal ka BP. Results show a transition from the Last Glacial Maximum (LGM) to the degl acial period similar to paleoclimate reconstructions from western North America, with a cold, wet LGM and a cold and/or dry early deglacial. The results also show a link between North Atlantic and tropical Atlantic climate during the deglacial. Periods of reduced AMOC and cool sea surface temperatures in the North Atlantic (e.g. Oldest Dryas and Younger Dryas) coincide with a more southerly position of the ITCZ, and cold and/or dry climate in the lowlands of northern Central America. Evidence for a switch t o colder and/or drier conditions near Lake PetŽn Itz‡ is recorded in the late part of the Mystery Interval (16.3 15.0 cal ka BP). This shift to colder and/or drier conditions in the late Mystery Interval is also recorded in the Greenland ice core isotope r ecord and the marine Cariaco Basin sediment grey scale record. Melting of a second ice armada during the Mystery Interval could have brought even colder conditions to the North Atlantic and Greenland, as well as a more southerly mean position of the ITCZ i n the Neotropics. The deglacial climate of lowland Central America, reconstructed from the PetŽn Itz‡ sediment sequence, differs from climate reconstructions derived from ice and marine sediment cores, which show a "two step" deglaciation. The PetŽn Itz‡ o xygen isotope record displays three large, abrupt climate transitions, at ~16.5 16.3 cal ka BP, ~15.0 14.4 cal ka BP, and ~10.6 10.3 cal ka BP. The first shift, in the midst of the Mystery Interval, appears

PAGE 29

29 anomalous, whereas the other two shifts mark the start of the Bolling Allerod Period and the end of deglacial aridity, respectively. Decreases in 18 O also coincide with lithologic transitions from gypsum to clay. Concurrence of the lithologic and isotopic shifts indicates that the lakewater chemical com position changed quickly, from a state of gypsum saturation to under saturation. Synchrony between the isotopic and lithologic changes argues for a rapid climate transition and for the predominant role of change in evaporation/precipitation ratio (E/P), ra ther than change in temperature or moisture source, on the isotope record. The "three step" deglaciation record from northern Guatemala reflects combined hemispheric and regional climate forcings. Chapter 4 is titled "Isotope measurements of single ostrac od valves and gastropod shells for climate reconstruction: evaluation of within sample variability and determination of optimum sample size" and was published in the Journal of Paleolimnology Chapter 4 contributes to the understanding of "within horizon" stable isotope variab ility (! 18 O and 13 C) measured on multiple, single ostracod valves and gastropod shells, and the potential for using intra sample 18 O variability in ostracod and gastropod shells as a proxy measure for high frequency climate variability. Results show that calculated optimum sample numbers ("n") for 18 O and 13 C in the ostracod Cytheridella ilosvayi and the gastropod Pyrgophorus coronatus vary appreciably throughout the sediment core. Individual 18 O measurements on C. ilosvayi show that samples from core depths that have high mean 18 O values, indicative of low effective moisture, display relatively low variability, whereas samples with low mean 18 O values, reflecting times of higher effective moisture, display relatively high variability. Relatively dry periods were thus consistently dry, whereas relatively wet periods had both wet and

PAGE 30

30 dry years. This interpretation of data from the cores applies to two important periods of the late Holocene, the Maya Terminal Classic period and the Little Ice Age. The 1 8 O variability during the ancient Maya Terminal Classic Period (ca. 910 990 AD) indicates not only the driest mean conditions in the last 3,000 years, but consistently dry climate. Chapter 5 summarizes the main conclusions from the previous chapters.

PAGE 31

31 Tabl e 1 1 Location of drill sites, water depth, penetration, and core recovery at Lake PetŽn Itz‡. Modified from Hodell et al. (2006). Site Latitude Longitude Water depth Drilling depth Average North West (m) (m ) recovery A B C D E (%) PI 1 16¡ 59.9706' 89¡ 47.7396' 65 94.5 90.3 82.5 89.3 PI 2 16¡ 59.9712' 89¡ 44.6850' 54 66.5 41.2 82.4 42.0 68.5 86.3 PI 3 17¡ 00.2016' 89¡ 49.2400' 100 96.9 95.3 90.0 92.9 PI 4 17¡ 00.3342' 89¡ 50.7720' 150 67.4 46.1 25.4 86.7 PI 6 17¡ 0 0.0162' 89¡ 47.0868' 71 75.9 66.4 66.8 94.9 PI 7 16¡ 59.7234' 89¡ 47.6844' 46 133.2 122.8 63.8 92.1 PI 9 16¡ 59.4360' 89¡ 47.6460' 30 16.4 91.8

PAGE 32

32 Table 1 2. Radiocarbon dates on terrestrial organic matter from Site s PI 2, PI 3, PI 6. Sit e D epth in core at site depth in core at site 6 Age ( mcd ) ( mcd ) ( 14 C yr BP) (1s) 2 0.32 0.32 425 30 2 4.88 4.00 1715 35 3 4.22 4.18 2140 40 6 8.08 8.08 3905 35 6 8.08 8.08 3920 35 2 10.54 8.12 3705 30 2 10.54 8.12 3775 35 2 11.14 8.68 4280 30 3 9.37 8.90 4590 40 3 9.37 8.90 4550 35 2 12.16 9.63 7455 35 2 12.16 9.63 7480 45 2 12.76 10.19 9855 35 6 10.41 10.41 8630 60 6 10.41 10.41 8655 30 6 10.44 10.44 7985 40 2 13.06 10.47 7835 30 3 11.32 10.69 8625 35 3 11.32 10.69 8675 35 6 10.86 10.86 9040 35 3 14.17 13.06 11210 35 6 13.35 13.35 11290 60 2 16.71 13.39 11135 40 6 13.41 13.41 11380 140 6 13.87 13.87 11390 50 2 17.74 14.26 11880 35 6 14.49 14.49 12460 60 2 20.71 16.81 13095 40 6 17.12 17.12 12280 60 2 21.91 17.55 13480 45 2 22.13 17.68 13505 45 2 22.13 17.68 13560 45 2 23.93 19.76 14810 45 2 23.93 19.76 14850 45 6 20.03 20.03 14130 120 6 21.72 21.72 14630 50 2 25.74 21.20 15355 50 6 23.64 23.64 17650 240 2 28.62 24.63 19740 70 6 25.73 25.73 19990 180 2 31.75 27.8 5 21940 80 3 28.09 29.16 23210 90 3 28.09 29.16 23040 90

PAGE 33

33 Table 1 2. Continued. Site D epth in core at site depth in core at site 6 Age ( mcd ) ( mcd ) ( 14 C yr BP) (1s) 2 34.43 30.64 25540 160 3 30.09 31.77 26560 130 6 31.86 31.86 26660 240 6 31.8 9 31.89 26820 260 6 33.73 33.73 30700 3600 6 33.82 33.82 30050 490 6 33.83 33.83 29120 170 6 33.83 33.83 29010 170 2 38.66 34.01 32680 980 6 34.39 34.39 28040 470 6 35.14 35.14 31230 260 3 33.42 37.82 32820 260 2 44.83 40.12 34380 450 2 46.55 40. 77 39900 730 6 41.30 41.30 39000 700 2 47.32 41.46 38760 740 2 48.17 41.76 41350 1020 2 48.35 41.87 41600 900 6 42.76 42.76 35900 1200 2 43.66 38.98 32540 310 6 45.31 45.31 38100 1100

PAGE 34

34 Table 1 3. Tie points used on the magnetic susceptibility reco rds to project dated intervals from site PI 2 onto site PI 6 S ite 2 depth S ite 6 depth ( mcd ) ( mcd ) 4.12 3.52 8.73 6.42 13.39 10.78 16.45 13.17 18.45 14.86 20.43 16.64 22.65 18.00 24.83 20.99 30.11 26.06 33.96 30.26 38.86 34.17 44.78 40.10 46 .18 40.55 50.41 43.12

PAGE 35

35 Table 1 4. Age depth relations for the Site PI 6 composite core used t o derive the chronology discussed in Chapters 2 and 3. Site Depth in core at site Depth in core at site 6 Age (mcd) (mcd) ( 14 C yr BP) (1s) 2 13.06 10.4 7 7835 30 3 11.32 10.69 8675 35 6 10.86 10.86 9040 35 3 14.17 13.06 11210 35 6 13.35 13.35 11290 60 6 13.41 13.41 11380 140 6 13.87 13.87 11390 50 2 17.74 14.26 11880 35 6 14.49 14.49 12460 60 2 20.71 16.81 13095 40 2 21.91 17.55 13480 45 2 22.1 3 17.68 13560 45 2 25.74 21.20 15355 50 6 23.64 23.64 17650 240 2 28.62 24.63 19740 70 6 25.73 25.73 19990 180 2 31.75 27.85 21940 80

PAGE 36

36 Figure 1 1 (A) Map of Central America showing the Location of the Yucatan Peninsula and the department of PetŽn Guatemala. (b) Map of the PetŽn Lake District. (C). Bathymetric map of Lake PetŽn Itz‡ showin g the location of primary and alternate coring sites. From Hodell et al. ( 20 08).

PAGE 37

37 Figure 1 2 Calibrated radiocarbon dates versus meters composite depth (mcd) for sites PI 3 (blue) and PI 6 (red) used to construct the age model used by Hodell et al. (2008).

PAGE 38

38 Figure 1 3 Calibrated radiocarbon dates versus meters composite depth (mcd) for sites PI 3 (blue) and PI 6 (red) used to construct the age mod el used by Hodell et al. (2008) for the Last Glacial Maximum (LGM) and deglacial period.

PAGE 39

39 Figure 1 4. Spliced magnetic suscep tibility records from Sites PI 6 and PI 2 for the Last Glacial Maximum (LGM) and deglacial period. Green triangles indicate tie points used to project dated intervals from site PI 2 onto site PI 6. Yellow dots indicate dated intervals on site PI 2.

PAGE 40

40 Figure 1 5. Magnetic susceptibility (red) and 18 O (blue) in sediment core PI 6 from 50.0 to 10.0 meters composite depth (mcd)

PAGE 41

41 Figure 1 6. Oxygen isotope data for Lake PetŽn Itz‡ for the last 45 cal ka. 18 O data indicated by blue circles ( Limnocythere sp.) are from PI 6. 18 O data indicated by red circles ( Limnocythere sp.) are from Lake PetŽn Itz‡ core 11A (Hillesheim e t al. 2005) 18 O data indicated by purple circles (Cochliopina sp. ), green circles ( Pyrgophorus sp.) brown circles ( Cytheridella ilosvayi ) and black circles ( Candona sp.) are from Lake PetŽn Itz‡ core 6 VII 93 (Curtis et al. 1998). Each record is plotted on its own, independent time scale.

PAGE 42

42 CHAPTER 2 CLIMATE CONDITIONS I N THE LOWLANDS OF CE NTRAL AMERICA DURING THE LAST GLACIAL MAXIUM (LGM) Introduction Global climate conditions during the Last Glacial Maximum (LGM) chronozone (~23 19 cal ka BP: Mix et al 2001) differed significantly from those of the late Holocene. Although seasonal insolation was close to modern values, sea level and ice sheet conditions, as well as greenhouse gas (e.g. CO 2 ) concentrations, were not. Monin et al. (2001) reports a mean C O 2 value of 189 ppm between 18.1 and 17.0 cal ka BP, in contrast to the pre industrial value of 280 ppm (Friedli et al. 1986). The maximum height of the North American and Fennoscandian ice sheets was ~3,000 meters (Peltier 1994) and sea level was ~120 met ers lower than present (Fairbanks 1989; Yokoyama et al. 2000). These conditions influenced temperature changes and atmospheric circulation patterns and moisture balance around the globe. The earliest attempt to reconstruct climate variability during the LG M was made by members of the Climate: Long range Investigation, Mapping, and Prediction (CLIMAP) Project. CLIMAP results for tropical oceans showed that mean sea surface temperatures (SSTs) were ~1 2¡C cooler than today with some tropical regions possessin g warmer conditions than present (CLIMAP 1976, 1981). New continental and marine data show cooler conditions than CLIMAP estimates and strong spatial differences in tropical cooling (Farrera et al. 1999; Ballantyne et al. 2005; Kucera et al. 2005). Althoug h climate reconstruction records that span the LGM have increased in number since the first CLIMAP results, there are still some regions in the tropics with sparse data.

PAGE 43

43 Geochemical, isotopic, and pollen analysis from a core taken in Lake Quexil, lowland Guatemala (Leyden et al. 1993) indicated a transition from relatively cold, arid conditions during the LGM to warmer and moister conditions during the Holocene. This interpretation, however, was based on poor chronological control for the Lake Quexil core, which was compromised by hard water error on a single radiocarbon date from a snail shell. New 14 C dates on terrestrial plant fragments from Lake PetŽn Itz‡ sediment cores, combined with pollen (Bush et al. 2009) and high resolution lithologic (Hodell et al. 2008) analyses, show that the LGM was cold and relatively wet. Clay rich sediments deposited between ~24.0 and ~19.0 cal ka BP, displayed high accumulation rates (2.5 mm/yr) and indicate high lake water levels, reflecting relatively wet climate conditi ons (Hodell et al. 2008). Pollen analysis supports the lithological interpretation. The pollen assemblage during the LGM reflects a scrub oak or montane pine oak forest, consistent with about a 4 6 C cooling and moist conditions (Bush et al. 2009). Althou gh both the lithologic and pollen data provide evidence for a cool, wet LGM, pollen counts were done at broad sampling intervals (Bush et al. 2009) and climate inferences from sediment lithology fail to capture subtle shifts in the evaporation/precipitatio n ratio (E/P), as such changes in lithology are a binary ("on off") proxy, reflecting either saturation or under saturation with respect to gypsum (Hodell et al. 2008). Here we present the first continuous, high resolution (~decadal), continental isotope ( 18 O) record from the northern Neotropics that spans the transition from Heinrich Event 2 (H2), through the LGM chronozone (~23.0 to 19.0 cal ka BP, as defined by Mix et al. 2001), and into the deglacial period. The objectives of this paper are to 1) test the inference for relatively moist conditions and climate stability within the

PAGE 44

44 LGM chronozone determined from lithologic and pollen data, 2) examine the detailed structure of climate transitions into and out of the LGM chronozone, and 3) suggest potential atmospheric processes responsible for changes in temperature and precipitation at the onset and end of the LGM in lowland Central America. Study Site Lake PetŽn Itz‡ (~16¡55' N, 89¡50' W) is located in the PetŽn Lake District in the lowlands of northern Gu atemala (Figure 2 1). It has a surface area of ~100 km 2 and a maximum water depth of >160 m (Anselmetti et al. 2006). Lake PetŽn Itz‡ is fed hydrologically by direct rainfall, runoff, and subsurface groundwater. It lacks surface outlets and although some subs urface seepage may occur, the water body can be considered a closed basin lake (Hodell et al. 2008). The lake water 18 O averages 2.62 (Curtis et al. 1998), greater than both the mean value for regional groundwater ( 4) and annual weighted mean precipita tion values ( 5.7 to 4.0) (Rozanski et al. 1993), reflecting the impact of evaporation on the lake's water budget. Lake PetŽn Itz‡ is situated in a climatically sensitive region. The rainy season, associated with the northward migration of the Intertrop ical Convergence Zone (ITCZ), occurs between June and December when easterly trade winds transport moisture from the Atlantic into the Caribbean Sea and the Yucat‡n Peninsula. The rainy season is followed by a pronounced dry season during northern hemisphe re winter (January through May) as the ITCZ moves southward. During the dry season, light winter precipitation is brought to the Yucat‡n Peninsula by polar air masses carried by northerly winds (Hodell et al. 2008).

PAGE 45

45 Methods Between 3 February and 11 March 2006, we drilled Lake PetŽn Itz‡ at seven sites as part of the Lake PetŽn Itz‡ Scientific Drilling Project (Hodell et al. 2006, 2008). Multiple cores were collected at six of the seven sites to guarantee retrieval of complete sediment sequences. Three sed iment cores from Site 6 (PI 6) (Figure 2 1) provide a continuous stratigraphic sequence to ~75.9 m composite depth (mcd), which represents ~85 cal ka of sediment accumulation and a long term mean sedimentation rate of ~1m/ka. The first age/depth model for the LGM and deglacial period in the PI 6 core from Lake PetŽn Itz‡ was derived from ten AMS 14 C dates on cores from drill sites PI 6 and PI 3 (Figure 3b and Table 3 in Hodell et al. 08). Dates from the PI 3 core were projected onto the PI 6 core by correla ting the two using their magnetic susceptibility records. New dates from drill sites PI 2 and PI 6 enabled development of a refined age/depth model. The new chronology is based on 17 AMS 14 C dates from drill sites PI 2, PI 3 and PI 6 (Table 2 1 and Figure 2 2) that cover the LGM and deglacial period. Radiocarbon measurements were done at the Center for Accelerator Mass Spectrometry (CAMS) at Lawrence Livermore National Laboratory and dates were calibrated using IntCal04: Northern Hemisphere (Reimer et al. 2004) and Fairbanks et al. (2005). All radiocarbon dates were measured on samples of terrestrial organic matter, thereby avoiding potential dating errors associated with hard water lakes (Deevey and Stuiver 1964). One cm 3 sediment samples were disaggregate d using a 3% H 2 O 2 solution for 18 O analysis of ostracod ( Limnocythere sp.) valves. Samples were then washed through a 63 $m sieve and material was collected on filter paper and dried at 60¡C. When dry, the material was transferred to glass scintillation vials. Adult ostra cod specimens were

PAGE 46

46 picked from the sieved, >212 $m fraction in each 1 cm sample using a binocular microscope. Prior to isotopic analysis, ostracod specimens were cleaned using 15% H 2 O 2 to remove organic material, and then rinsed in methanol before drying. Ostracods were checked for impurities inside the valves, and cleaned again if necessary. Approximately 12 20 individual ostracod carapaces, weighing a total of ~20 60 $g, were combined to constitute a sample. Multiple ostracod specimens were measured from each 1 cm stratigraphic level to reduce the variance generated by running single short lived individuals (Heaton et al. 1995; Escobar et al. in press). Ostracod valves were loaded into glass vials and CO 2 was evolved from shells with a single aliquot acid digestion in a Kiel III carbonate preparation device attached to a Finnigan MAT 252 isotope ratio mass spectrometer at the University of Florida. Results There is an increase in 18 O (~2.6) from ~25.0 to 24.5 cal ka BP. This 18 O transition coincides with a shift to low magnetic susceptibility values and high concentrations of gypsum (Figure 2 3). A decrease in 18 O (~2.4) from ~24.5 to 24.0 cal ka BP coi ncides with a shift from gypsum to clay deposition, i.e. an increase in magnetic susceptibility values (Figure 2 3). There is a gap in the oxygen isotope record between ~23.5 and ~21.5 cal ka BP. Oxygen isotope values show no substantial changes between ~2 1.5 and ~19.0 cal ka BP, with 18 O values ranging only from 5.0 5.5 (Figure 2 3). Constant 18 O values coincide with steady clay deposition during this time period. There is an increase in 18 O (~1.3) from ~19.0 to 16.5 cal ka BP, marking the transition to the deglacial period (Escobar et al. in prep). This isotopic shift coincides with a switch to lower magnetic susceptibility values and the onset of gypsum deposition (Figure 2 3).

PAGE 47

47 Discussion The 18 O value of precipitation depends, in part, on the isoto pe composition of the moisture source (e.g. sea water) and the extent of moisture depletion in the air mass as it is transported from the source area to the rainfall region (Leng et al. 2006). There is progressive depletion of H 2 18 O with increasing latitud e, altitude, distance from the coast, and rainfall amount (Dansgaard 1964). The relative humidity and temperature of the atmosphere also contribute to isotopic fractionation in precipitation, from the time water evaporates from the ocean to the moment a ra indrop hits the ground. Several factors can alter lakewater isotopic composition in closed basin lakes where precipitation is the main water source. Among these factors are: (1) changes in the isotopic composition of the source water, either through a chan ge in the isotope signature of the original source, or a change in the source region from which moisture is derived, (2) changes in the routing of moist air masses, and (3) changes in the relative humidity and/or temperature at any time from moisture forma tion to raindrop precipitation (Darling et al. 2006; Leng et al. 2006). In closed basin Lake PetŽn Itz‡, the relative amount of precipitation to evaporation is the principal control on the lakewater isotopic signature, as lakewater is several per mil enric hed over regional rainfall and groundwater. Mean lakewater 18 O (2.62: Curtis et al. 1998) is greater than both the mean value for regional groundwater ( 4) and annual weighted mean rainwater values in the circum Caribbean lowlands ( 5.7 to 4.0) (Roza nski et al. 1993), indicating substantial evaporative loss of water from the lake. Periods of high E/P display relatively lower lake level and higher water column 18 O values. Conversely, during episodes of low E/P, the lake shows higher water level and di splays lower 18 O values.

PAGE 48

48 Lachniet and Patterson (2009) also investigated the variables that control surface water 18 O values in the Yucatan Peninsula. They analyzed 18 O of surface waters from Guatemala and Belize as well as data from the Global Network for Isotopes in Precipitation (GNIP) database. Spatially, distance from the coast (i.e. the "continentality effect") and mean catchment altitude (i.e. the "altitude effect") explain 84% of the surface water 18 O variability. In a given water body, the most important control on precipitation 18 O values through time was rainfall amount (i.e. the "amount effect"). Lakewater 18 O, together with water temperature, largely determines the 18 O in the calcium carbonate of shell forming aquatic organisms such as os tracods, gastropods, and foraminifera. Colder temperatures yield greater 18 O values in precipitated carbonate, whereas warmer temperatures produce lower 18 O values. Thus, the switch from higher 18 O values to lower 18 O values in ostracod shells during t he transition from H2 to the LGM could reflect changes in one or more of the following factors: 1) precipitation source, 2) isotope value of the water source, 3) atmospheric route of the moist air mass, and 4) a shift from colder and/or drier, to warmer a nd/or wetter conditions. Likewise, the switch from lower 18 O values to higher 18 O from the LGM to H1 could reflect changes in the first three factors listed above, as well as a shift to colder and/or drier conditions. Pollen counts on sediments from Lake PetŽn Itz‡ indicate dominance of a pine oak forest assemblage during the LGM chronozone, consistent with a cooling of ~4 6 ¡C (Bush et al. 2009). Sea surface temperature reconstructions from ocean sediment cores in the tropical Atlantic suggest a smaller drop in temperature during the LGM than the temperature decreases inferred for continental areas. Schmidt et al. (2004)

PAGE 49

49 suggested a ~2.5 ¡C cooling for the western Caribbean, whereas Lea et al. (2003) indicated a cooling of 2.6 ¡C 0.5 ¡C (Figure 2 1) rel ative to the late Holocene for the Cariaco Basin. There are no continental paleotemperature reconstructions for the transitions into and out of the LGM chronozone. Ocean sediment records from the northeast North Atlantic (Bard 2000) and the Mediterranean ( Cacho et al. 1999) show clear changes in SSTs, with colder temperatures during Heinrich events 2 and 1 than during the LGM (Figure 2 1). Cacho et al. (1999) report a ~4.1 ¡C decline in SST from ~25.0 to ~24.0 cal ka BP, and Bard (2000) reports a ~8.5 ¡C dr op in SST from ~18.0 to 17.0 cal ka BP. These two sites are, however, close to the Heinrich zone of influence (Hemming 2004). Sea surface temperature reconstructions from tropical marine regions do not show a significant change in temperature during these transitions. For example, SST reconstruction from a core taken in the Cariaco Basin, off northern Venezuela shows that temperatures during Heinrich event 1 are statistically indistinguishable from average temperatures during the LGM (Lea et al. 2003). If o xygen isotope changes in the Lake PetŽn Itz‡ record at the LGM chronozone boundaries were solely attributable to temperature, it would have required a temperature decrease of ~9.6 ¡C at the onset of the LGM, and a temperature increase of ~5.2 ¡C at the end of the LGM. Stability of SST reconstru ctions from tropical marine ocean records suggests that E/P changes, rather than temperature, account for the bulk of the 18 O change in Lake PetŽn Itz‡ record. Furthermore, changes in 18 O coincide with lithology changes. The lithologic shift is thought t o have been driven wholly by E/P, with gypsum precipitation under dry conditions when calcium sulfate concentrations

PAGE 50

50 reached saturation, or alternatively, clay deposition during moist periods, when the water column was undersaturated with respect to gypsum A temperature driven shift in isotopes would not have necessarily been accompanied by a concurrent shift in lithology. This synchrony between the isotope and lithologic changes argues for the predominant role of E/P, rather than temperature, in influenci ng the isotopic record. A shift to wetter conditions at the H2/LGM transition and a shift to drier conditions at the LGM/H1 transition, support previous interpretations from lithological (Hodell et al. 2008) and palynological (Bush et al. 2009) results. Pa leoglacier studies from several sites in Central America also support claims for a cool and wet LGM (White 1986; Anderson 1969; Hastenrath 1973). Glacial advances down to ~3,400 masl in Mexico and Guatemala suggest that the region was cooler and moister th an present. Glacial studies in the plateau of Altos de Cuchumatanes (3,837 masl), Guatemala, indicate that an ice cap ~60 km 2 in extent, existed above an elevation of 3,400 masl (Lachniet and Vasquez Selem 2005) (Fig 2 1). There are no radiocarbon dates fo r this glacial advance/retreat in the plateau of Altos de Cuchumatanes, but Lachniet (2004) and Lachniet and Vasquez Selem (2005) suggest a correspondence with Mexican glacier chronology. Tephrachronology, together with 32 Cl dating, point to an age between 20.0 and 17.5 36 Cl ka BP for Hueyatlaco 1 glacier deposits and between 17.0 and 14.0 36 Cl ka BP for Hueyatlaco 2 glacier deposits on the extinct Iztacc’huatl volcano, central Mexico (Lachniet and Vasquez Selem 2005) (Figure 2 1). Paleoglacier data from th e Tanc’taro volcano, central Mexico, also show glacier retreat between 17.5 and 15.0 36 Cl ka BP (Vazquez Selem unpublished data, cited in Lachniet and Vasquez Selem 2005) (Figure 2 1). Oxygen isotope data from Lake PetŽn

PAGE 51

51 Itz‡ are in agreement with glaciolo gy data in Mexico and Guatemala. Wetter conditions recorded in the Lake PetŽn Itz‡ isotope record during the LGM would have been necessary for the expansion of glaciers into high elevation terrain where no glaciers exist today. Furthermore, the retreat of glaciers occurred sometime after the increase in Lake PetŽn Itz‡ 18 O values, which indicates the onset of drier conditions (Escobar et al. in prep). An increase in summer or winter precipitation could explain the higher lake levels of Lake PetŽn Itz‡ betw een 24 and 18 ka cal yr BP. Bush et al. (2009) suggest a prolonged summer rainy season to explain wet conditions during the LGM. Ocean heat capacity, i.e. the slow uptake and release of heat, might have caused regional SST increases to lag changes in insol ation. This lag might have decreased summer rainfall, but increased fall precipitation (Bush et al. 2009). McManus (04) shows a slower Atlantic Meridional Overturning Circulation (AMOC) during Heinrich events 2 and 1 than during the LGM (Figure 2 1). A mor e southerly position of the Intertropical Converge Zone (ITCZ) during Heinrich events 1 and 2 could explain the drier conditions in PetŽn during these periods, whereas a more northerly position of the ITCZ during the LGM could account for a wetter LGM. McM anus (2004) shows AMOC strength for the LGM very similar to the strength seen during the Bolling Allerod period, a time for which there is evidence of a northerly ITCZ position (Peterson et al. 2000). The geographic position of the ITCZ during the LGM is, however, still controversial (Jaeschke et al. 2007; Koutavas and Lynch Stieglitz 2004). If indeed the ITCZ had a more southerly position during the LGM chronozone, then lower summer precipitation would be expected and

PAGE 52

52 increased winter precipitation might b est explain the wet LGM in the lowlands of Guatemala. During the LGM, the Laurentide Ice Sheet modified atmospheric circulation over the North America continent. Wetter climate in the American southwest and western Mexico during the LGM has been explained by the split of the jet stream into a dry north and wet south branch, due to the presence of the Laurentide Ice sheet (Kutzbach and Guetter 1986). Early climate models showed that a split westerly jet stream could bring precipitation to the western USA to at least 20¡ N (Kutzbach and Guetter 1986). More recent climate models suggest that the northern and southern branches were located farther south and that the splitting of the westerly jet occurred only during the winter months (Bromwich et al. 2004; Kim et al. 08). The model of Kim et al. (2008) shows, however, that although precipitation apparently increased in winter, it decreased during summer. In their model, annual average climate conditions during the LGM were drier than present. Their model results thus contradict the interpretations of the paleoclimate proxy data from Lake PetŽn Itz‡ sediments. The Yucatan Pensinsula however, lies at the southern boundary of many of these climate simulations (Bromwich et al. 2004) and results for Yucatan should the refore be considered carefully. Wet conditions during the LGM in the Yucatan Peninsula might have also been a consequence of the Mobile polar high (Leroux 1993). Polar outbreaks of air from the north Pacific could have reached lower latitudes and migrated to the west, and low pressure cells could have moved moister eastern tropical Pacific air to the northeast, providing winter precipitation for the Yucatan Peninsula (Leroux 1993).

PAGE 53

53 An increase in winter precipitation of Pacific origin would imply changes in the source region from which moisture was derived and in the routing of moist air masses. Such changes would be expected to cause a shift in PetŽn Itz‡ lake water isotope values. If increased winter precipitation of Pacific Ocean origin caused for a wet LGM, then the exact cause and timing of the shift to a modern climate regime (i.e. summer rainy season) in the PetŽn during the de glacial period remains unknown. Differences between 18 O results from the LGM and the deglacial period may have been a consequence of not only changes in E/P but also a consequence of changes in moisture source region and routing of moist air masses. A cha nge in the 18 O of the precipitation source and/or temperature, however, would probably not account for the large difference between LGM and deglacial 18 O results in the PetŽn Itz‡ record. Schrag et al. (2002) reports that the glacial interglacial change in 18 O of a deep North Atlantic site (ODP site 981) was 0.7 0.8, and change in the Southern Atlantic (ODP site 1093) was 1.1. These results corroborate other studies that show the global average difference between seawater 18 O values in the LGM and pre sent was 1.0 0.1. Modern values were probably reached by a change in temperature and ice/ocean volume during the deglacial period. Bradbury (1997) compiled results of LGM paleoclimate studies carried out in several regions of the southwest United States Central America and northern South America. He observed a boundary at 95 ¡W, with a cold and dry LGM east of 95¡W, and a cold and wet LGM west of 95¡W (Bradbury 97). Lake PetŽn Itz‡ isotope data, together with pollen and lithology data, contribute to the understanding of regional climate conditions during the LGM, with the wet/dry boundary now moved eastward to 89¡50' W

PAGE 54

54 (Figure 2 4). The causes of LGM climate change on the Yucatan Peninsula remain poorly understood. Central America and the Yucatan Peninsu la are challenges for climate modelers due to the complex topography within such a small area. Future work should focus on developing climate models for the Yucatan Peninsula during the LGM. Conclusions Oxygen isotopes (! 18 O), lithology and pollen data from a sediment core taken in Lake PetŽn Itz‡ provide evidence for a relatively wet LGM in lowland Central America, bracketed by drier conditions during Heinrich events 2 and 1. Wetter conditions during the LGM, inferred from the Lake PetŽn Itz‡ isotope record, are in agreement with paleoglacier data from Mexico and Guatemala. Glacial advances down to ~3,400 masl in Mexico and Guatemala suggest that the region was cooler and moister than present during the L GM. Retreat of glaciers in the region occurred sometime after the increase in Lake PetŽn Itz‡ 18 O values, which indicates a transition to drier conditions. Wetter conditions during the LGM in the lowlands of Central America may be explained by (1) an increase in summer precipitation due to a more northerly po sition of the Intertropical Converge Zone (ITCZ), or (2) an increase in winter precipitation due to the splitting of the jet stream and/or movement of moist eastern tropical Pacific air to the northeast. Factors that influenced climate change on the Yucata n Peninsula during the LGM are still poorly understood, but modeling efforts may provide insights into the climate forcing factors at that time.

PAGE 55

55 Figure 2 1 (A) Map showing the locations of Lake PetŽn Itz‡, northern Guatemala, and paleoclimate reconstru ction sites discussed in this paper. Cariaco (Lea et al. 2003 and Peterson et al. 2000), OCE 326 GGC5 (McManus et al. 2004), SU 8118 (Bard 2002), MD95 2043 (Cacho et al. 1999) (B) Map showing the locations of Lake PetŽn Itz‡ and paleoglacier reconstruction s discussed in this paper. (C) Map of Lake PetŽn Itz‡ showing the location of drilling sites PI 2, PI 3, and PI 6.

PAGE 56

56 Figure 2 2 (A) Calibrated radiocarbon dates versus meters composite depth (mcd) for sites PI 3 (blue) and PI 6 (red) used to construct the age model used by Hodell et al. (2008). (B) Calibrated radiocarbon dates versus mcd for sites PI 3 (blue), PI 6 (red), and PI 2 (yellow) used to construct the age model used in this paper. Age model was derived by linear interpolation between selected age depth points.

PAGE 57

57 Figure 2 3 Magnetic susceptibility (red) and 18 O (blue) in sediment core PI 6. Light gray band represents the Last Glacial Maximum (23.0 19.0 cal ka BP) as define by Mix et al. (2001).

PAGE 58

58 Figure 2 4. Summary map (modified from Bradbury 1997) of moisture conditio ns during the Last Glacial Maximum in the southwestern United States, Central Mexico and northern South America. Dashed line represents the wet dry longitudinal boundary proposed by Bradbury (1997). Black line represents the new wet dry longitudinal bounda ry proposed based on results from this study.

PAGE 59

59 CHAPTER 3 CLIMATE CHANGE IN TH E NORTHERN NEOTROPIC AL LOWLANDS DURING T HE LAST DEGLACIATION Introduction Central America and the Caribbean have been regions of intensive paleoenvironmental research in the last three decades (Brenner et al. 2002). Most research has focused on Holocene climate change (Hodell et al. 1991; Islebe et al. 1996) and environmental consequences of human land use (Brenner 1983; Binford et al. 1987), with few records extending beyond the Holocene (Leyden 1984, 1995; Leyden et al. 1993). Our knowledge of Pleistocene climatic and environmental changes in the region is limited by the scarcity of high resolution, pre Holocene sediment records with good temporal control (Leyden 1995). Most of w hat is known about late Pleistocene climate in the continental region comes from lacustrine records from La Chonta Bog, Costa Rica (Hooghiemstra et al. 1992; Islebe et al. 1995), Lake La Yeguada, Panama (Piperno et al. 1990; Bush et al. 1992), and Lake Que xil (Deevey et al. 1983; Leyden et al. 1993). Some of the best regional paleoclimate records for last deglacial period come from marine deposits in the Gulf of Mexico (Flower et al. 2004; Aharon 2005) and the Cariaco Basin, north of Venezuela (Peterson et al. 2000; Hughen et al. 1996; Lea et al. 2003). It was previously thought that the present interglacial represents a transition from relatively arid conditions during the Last Glacial Maximum (LGM), to moister conditions during the Holocene. Geochemical, i sotopic, and pollen analysis from Lake Quexil, lowland Guatemala (Leyden et al. 1993), suggested that the last deglaciation occurred in two distinct phases, i.e. a "two step" deglaciation, as seen in records from high latitude regions. The first phase, sta rting at 14 ka 14 C BP, was characterized by

PAGE 60

60 expansion of temperate oaks and low 18 O values that indicated warmer, more humid conditions than during the LGM (Leyden 1995). A second phase (12 ka 10.3 ka 14 C BP), with calcite precipitation in sediments, a reversal of forest expansion, but sustained low 18 O values, indicated a return to a colder (~1.5C), but sustained humid climate (Leyden 1995). This interpretation, however, was based on poor chronological control for the Lake Quexil core, which was likel y influenced by hard water error associated with a single date on a snail shell. New dates on terrestrial plant fragments from the PetŽn Itz‡ core, combined with high resolution lithologic analyses, indicate that, to the contrary, the LGM was relatively w et (Hodell et al. 2008). Pollen analysis of the PetŽn Itz‡ core shows that the driest period of the last glaciation was not the LGM, but instead, the deglacial period (Bush et al. 2009). Recently completed lithologic and pollen analyses do not, however, re veal a detailed picture of the timing, rate, or structure of the deglacial period following the end of the cold, moist LGM. Pollen counts were done at broad sampling intervals (Bush et al. 2009) and climate reconstructions inferred from lithology fail to c apture subtle shifts in E/P, as it reflects a binary ("on off") proxy for water column conditions i.e. saturation or under saturation with respect to gypsum (Hodell et al. 2008). Here we present the first continuous, high resolution (~decadal), continent al isotope (! 18 O) record from the northern Neotropics that spans the last deglaciation. This isotope record was generated using measurements on ostracod ( Limnocythere sp.) valves and allows detailed examination of critical climate events during the last de glaciation, such as the Mystery interval, the Bolling Allerod period, the Younger Dryas, and the Preboreal excursion in

PAGE 61

61 the lowlands of Central America. Furthermore, it helps elucidate climate teleconnections between the northern Neotropical lowlands and h igh latitude regions. Study Site Lake PetŽn Itz‡ (~16¡55' N, 89¡50' W) is located in the PetŽn Lake District in the lowlands of northern Guatemala (Figure 3 1). It has a surface area of ~100 km 2 and a maximum water depth of 160 m (Anselmetti et al. 2006). Lake PetŽn Itz‡ is fed hydrologically by direct rainfall, runoff, and subsurface groundwater. It lacks surface outlets and although some subsurface seepage may occur, it is effectively a closed basin lake (Hodell et al. 2008). Lake PetŽn Itz‡'s water is d ilute (11.22 meq/l) and it is dominated by calcium and magnesium cations and bicarbonate and sulfate anions (Hillesheim et al. 2005). Lakewater pH is high (~8.0) and at present the lake is saturated with calcium carbonate, which precipitates and accumulate s in shallow zones of the water body (Hodell et al. 2008). The lake is probably thermally stratified throughout most of the year, though temperature differences between the hypolimnion (~25.4 C) and epilimnion (~30.0 C) are relatively small, even in summ er. The lake water 18 O averages 2.62 (Curtis et al. 1998), greater than both the mean value for regional groundwater ( 4) and annual weighted mean precipitation values ( 5.7 to 4.0) (Rozanski et al. 1993), reflecting the importance of evaporation on the lake's water budget. Lake PetŽn Itz‡ is situated in a climatically sensitive region. The amount of rainfall is related to the seasonal migration of the Intertropical Convergence Zone (ITCZ) and the Azores Bermuda high pressure system (Hastenrath 1984). The rainy season, associated with the northward migration of the ITCZ, occurs between June and December as easterly trade winds transport moisture from the Atlantic into the

PAGE 62

62 Caribbean Sea and the Yucat‡n Peninsula. The rainy season is followed by a pronou nced dry season during northern hemisphere winter (January through May) as the ITCZ moves southward. During the dry season, light winter precipitation is brought to the Yucat‡n Peninsula by polar air masses carried by northerly winds (Hodell et al. 2008). Field and Laboratory Methods Between 3 February and 11 March 2006, we drilled Lake PetŽn Itz‡ at seven sites as part of the Lake PetŽn Itz‡ Scientific Drilling Project (Hodell et al. 2008). Sediment cores from Site PI 6 (Figure 3 1) provide a continuous st ratigraphic sequence to ~75.9 m composite depth (mcd), that represents ~85 cal ka of sediment accumulation, indicating a long term mean sedimentation rate of ~1m/ka. Here we focus on the deglacial record from Lake PetŽn Itz‡. To obtain ostracod valves ( Lim nocythere sp.) for 18 O analysis, 1 cm sediment samples were disaggregated using a 3% H 2 O 2 solution. Samples were then washed through a 63 $m sieve and material was collected on filter paper and dried at 60¡C. When dry, the material was transferred to glass scintil lation vials. Adult specimens of Limnocythere sp. were picked from the sieved, >212 $m fraction in each 1 cm sample using a binocular microscope. Prior to isotopic analysis, ostracod specimens were cleaned using 15% H 2 O 2 to remove organic material, and the n rinsed in methanol before drying. Ostracods were checked for impurities inside the valves, and cleaned again if necessary. Approximately 12 20 individual ostracod carapaces, weighing a total of ~20 60 $g, were used for all samples. Multiple ostracod spec imens were measured from each stratigraphic level to reduce the variance generated by running single short lived individuals (Heaton et al. 1995; Escobar et al. in press). Ostracod valves were

PAGE 63

63 loaded into glass vials and CO 2 was evolved from shells with a single aliquot acid digestion in a Kiel III carbonate preparation device attached to a Finnigan MAT 252 isotope ratio mass spectrometer at the University of Florida. Proxy Interpretation: Oxygen Isotopes The 18 O value of precipitation depends, in part, on the isotope composition of the moisture source (e.g. sea water) and the extent of moisture depletion in the air mass as it is transported from the source area to the rainfall region (Leng et al. 2006). T here is progressive depletion of H 2 18 O with increasing latitude, altitude, distance from the coast, and rainfall amount (Dansgaard 1964). The relative humidity and temperature of the atmosphere also contribute to isotopic fractionation in precipitation, fr om the time water evaporates from the ocean to the moment a raindrop hits the ground. Several factors can alter lakewater isotopic composition in closed basin lakes where precipitation is the main water source. Among these factors are: (1) changes in the i sotopic composition of the source water, either through a change in the isotope signature of the original source, or a change in the source region from which moisture is derived, (2) changes in the routing of moist air masses, and (3) changes in the relati ve humidity and/or temperature at any time from moisture formation to raindrop precipitation (Darling et al. 2006; Leng et al. 2006). In closed basin Lake PetŽn Itz‡, the relative amount of precipitation to evaporation is the principal control on the lakew ater isotopic signature, as lakewater is several per mil enriched over regional rainfall and groundwater. Mean lakewater 18 O (2.62: Curtis et al. 1998) is greater than both the mean value for regional groundwater ( 4) and annual weighted mean rainwater values in the circum Caribbean lowlands ( 5.7 to 4.0) (Rozanski et al. 1993), indicating substantial evaporative loss of water from

PAGE 64

64 the lake. Periods of high E/P display relatively lower lake level and higher water column 18 O values. Conversely, during episodes of low E/P, the lake shows higher water level and displays lower 18 O values. Lachniet and Patterson (2009) also investigated the variables that control surface water 18 O values in the Yucatan Peninsula. They analyzed 18 O of surface waters from Guatemala and Belize as well as data from the Global Network for Isotopes in Precipitation (GNIP) database. Spatially, distance from the coast (i.e. the "continentality effect") and mean catchment altitude (i.e. the "altitude effect") explain 84% of the s urface water 18 O variability. In a given water body, the most important control on precipitation 18 O values through time was rainfall amount (i.e. the "amount effect"). Lakewater 18 O, together with water temperature, largely determines the 18 O in the c alcium carbonate of shell forming aquatic organisms such as ostracods, gastropods, and foraminifera. Colder temperatures yield greater 18 O values in precipitated carbonate, whereas warmer temperatures produce lower 18 O values. Thus a switch from higher 18 O values to lower 18 O values in ostracod shells could reflect a change in precipitation source and/or amount, as well as a change from colder and/or drier, to warmer and/or wetter conditions. Conversely, a switch from lower 18 O values to higher 18 O co uld reflect a change in precipitation source and/or amount, as well as a change to colder and/or drier conditions. Core Chronology The first age/depth model for the deglacial period (~18 10 cal ka BP) in the PI 6 core from Lake PetŽn Itz‡ was derived from six AMS 14 C dates on cores from drill sites PI 6 and PI 3 (Figure 3b and Table 5, Hodell et al. 08). Dates from the PI 3 core were projected onto the PI 6 core by correlating the two using their magnetic susceptibility

PAGE 65

65 records. New dates from drill sites PI 2 and PI 6 enabled development of a refined age/depth model. The new chronology is based on 17 AMS 14 C dates from drill sites PI 2, PI 3 and PI 6 (Table 3 1 and Figure 3 2) that cover the LGM and deglacial period. Radiocarbon measurements were done at t he Center for Accelerator Mass Spectrometry (CAMS) at Lawrence Livermore National Laboratory and dates were calibrated using IntCal04: Northern Hemisphere (Reimer et al. 2004) and Fairbanks et al. (2005). All radiocarbon dates were measured on samples of t errestrial organic matter, thereby avoiding potential dating errors associated with hard water lakes (Deevey and Stuiver 1964). Results A shift from relatively moist to more arid conditions occurred at the end of the LGM, ~19 ka BP (Escobar et al. in prep) The last deglaciation was marked in Lake PetŽn Itz‡ by three large decreases in the oxygen isotope ratio of the water, beginning at ~16.5, ~15.0 and ~10.6 cal ka BP (Figure 3 3). These represent shifts to warmer and/or wetter conditions. Gradual increas es in 18 O started ~19.0, ~16.3 and ~13.1 cal ka BP, indicating drying and/or cooling climate conditions (Figure 3 3). A gradual increase in 18 O from ~19.0 16.5 cal ka BP coincides with low sediment magnetic susceptibility (Figure 3 3). Low magnetic susce ptibility is associated with high concentrations of gypsum, which precipitated from the water column as lake level declined, under relatively dry conditions (Hodell et al. 2008). The first large decline in the 18 O (2.2) from ~16.5 16.3 cal ka BP indicate s a shift to wetter and/or warmer conditions (Figure 3 3). The decline in 18 O coincides closely with the cessation of gypsum precipitation and the onset of clay deposition. From ~16.3 15.0 cal ka BP, generally increasing 18 O indicates a switch from relat ively warmer and/or more humid

PAGE 66

66 conditions to increasingly colder and/or drier conditions. Midway into the increasing 18 O trend, clay deposition ceased and gypsum precipitation began. The second decline in 18 O (2.6) was a two step process, with an abrupt drop from ~15.0 14.4 cal ka BP and a more gradual decline from ~14.4 13.1 cal ka BP. This downward turn in 18 O marks the onset of the Bolling Allerod period. Clays deposited during the Bolling Allerod period underlie and overlie gypsum layers that indica te relatively dry conditions. At the end of the Allerod period, clay deposition ceased and gypsum deposition resumed, indicating a return to drier conditions that mark the start of the Younger Dryas, ~13.1 cal ka BP. A coincident increase in 18 O indicates a switch to colder and/or drier conditions (Figure 3 3). The end of the Younger Dryas is marked by a gradual decline in 18 O (0.6) from ~11.5 10.6 cal ka BP, followed by an abrupt decrease (1.8) in 18 O from ~10.6 10.3 cal ka BP, i.e. the end of the Pre boreal period, indicating a change to warmer and/or wetter conditions. Unlike the isotope record, which displays a substantial decrease at the end of the Younger Dryas, lithology of the PetŽn Itz‡ core shows no change. Gypsum precipitation did not cease u ntil the end of the Preboreal period. Climate Reconstruction for the Last Deglaciation Strong inter hemispheric climate coupling is seen during the last glacial period in studies of ice cores from Antarctica (EPICA Community Members 2006) and Greenland (No rth Greenland Ice Core Project members 04). An opposite climate response between the two hemispheres, i.e. a "sea saw" response (Broecker et al. 1990; Broecker 1998), is evident in comparisons between large Antarctic warm events and stadial periods in the Northern Hemisphere (EPICA Community Members 2006). This "see saw" pattern is explained by changes in the heat and freshwater flux (Bond et al. 1997; Schmidt et al. 2004) connected to the Atlantic Meridional Overturning Circulation

PAGE 67

67 (AMOC) (EPICA Community Members 2006). In the Atlantic Ocean, warm, saline waters move northward from the tropics and subtropics to the Northern North Atlantic region. In this region, surface waters release heat into the atmosphere, become denser, sink and flow southward to the s outhern Hemisphere. Freshwater input in the northern North Atlantic or subtropical North Atlantic (e.g. Gulf of Mexico) would have influenced the AMOC with global climate consequences (Broecker 1994, 2003; Otto Bleisner and Brady et al. 09). A reduction in AMOC confines heat to the southern hemisphere with cooling of the Northern Hemisphere. A bipolar climate response, associated with AMOC changes, is also observed during the deglacial period (Broecker 1998; Carlson 2008), when changes in Earth's orbital pa rameters, specifically an increase in northern hemisphere summer insolation, initiated the last deglaciation (Imbrie et al. 1992, 1993; Raymo 1997). Although not synchronous and of different magnitudes (Wunsch 2003), warming events during the deglacial per iod in Antarctica are associated with cold periods in Greenland (Oldest and Younger Dryas), and the Antarctic Cold Reversal (ACR) is associated with the Bolling Allerod warm period in the Northern Hemisphere. Time series studies of instrumental climate dat a (Hastenrath and Greischar 1993; Nobre and Shukla 1996) and climate models (Chiang and Bitz 2005; Timmermann et al. 2005; Zhang and Delworth 2005) show that the Intertropical Convergence Zone (ITCZ) is located in the warmer of the two hemispheres. Consequ ently, when AMOC decreases, i.e. when North Atlantic sea surface temperatures cool, the ITCZ migrates southward (Schiller et al. 1997; Vellinga and Wood 2002; Peterson and Haug 2006). In contrast, strong AMOC formation and consequent warming of the North A tlantic favors a

PAGE 68

68 northward position of the ITCZ (Peterson and Haug 2006). Paleoclimate records show similarities between temperature reconstructions from high latitude North Atlantic sites and changes in temperature and the hydrological cycle in the northe rn Neotropics during the last deglaciation (Hughen et al. 1996; Peterson et al. 2000; Lea et al. 2003). Global sea level rose by ~120 m since the LGM (Fairbanks et al. 2005; Bard et al. 1996; Hanebuth et al. 2000). Coral records provide evidence for a grad ual rise punctuated by a rapid sea level rise of >20m, Meltwater pulse 1A (MWP 1a), at some point during the last deglaciation. The causes, timing, and climatic consequences of MWP 1a during the last deglaciation remain topics for debate. Sea level reconst ructions using fossil reefs from Barbados and Tahiti (Bard et al. 1996; Fairbanks et al. 2005; Peltier and Fairbanks 2006) place the onset of MWP 1A at ~14.2 14.0 cal ka BP. The Older Dryas climate event has been associated with MWP 1A (Stanford et al. 200 6). However, Sea level reconstructions from the Sunda Shelf (Hanebuth et al. 2000), long chain n alkane concentrations in cores from the northern South China Sea (Kienast et al. 2003), and a new record from Tahiti (Deschamps et al. 2009) place the onset of MWP 1A at ~14.7 14.6 cal ka BP. Clark et al. (2002) indicate that a considerable amount of water from MWP 1A was derived from the Antarctic Ice Sheet, with climate models (Weaver et al. 2003) suggesting this source as a triggering factor for the onset of the Bolling Allerod Period. Regardless of the timing, rate and source of meltwaters, isotope values of ocean water, i.e. the source waters for rainfall, changed during the deglacial period, with a general trend to more depleted mean values. Thus, the Lake Peten Itza 18 O record might have been influenced by the change in the 18 O of the precipitation source. Although there is a large difference between meltwaters and ocean

PAGE 69

69 isotope values, this effect is thought to have been small given the large differences in volume between the two reservoirs. The gradual increasing trend in oxygen isotope values from ~19.0 16.5 cal ka BP, and low magnetic susceptibility values associated with gypsum precipitation, indicate an increase in E/P and low lake levels during the early part of the "Mystery Period" (Denton et al. 2006). Isotope results are supported by pollen data and independent reconstructions of past water levels, which show that the early deglacial period was the time of maximum aridity (Bush et al. 2009) and lo west lake stage. Mueller et al. (in press) show a lake stage reduction of ~68 m, which implies a water depth at site PI 6 of only 3 m at that time. 231 Pa/ 230 Th data from the Bermuda Rise (McManus et al. 2004) and a magnetic susceptibility record from the E irik Drift, off the southern tip of Greenland (Stanford et al. 2006), show an almost complete shutdown of the AMOC during the early deglacial interval. Furthermore, a marine sediment core from the Iberian Margin (SU8188) indicates cold sea surface temperat ures (SST) and reduced salinity during Heinrich event 1 (H1) (Bard 2002). The timing of this increase in E/P in the lowlands of Central America thus coincides with evidence for cooling in Greenland (North Greenland Ice Core Project members 2004), a slowdow n in the Atlantic Meridional circulation (AMOC) (McManus et al. 2004), and cooler sea surface temperatures in the subtropical northeastern Atlantic (Bard 2002) (Figure 3 4). Tropical and subtropical climate reconstructions during H1 show saltier waters in the Caribbean (Schmidt et al. 2004), and drier conditions inferred from cores taken in the marine Cariaco Basin, north of Venezuela (Peterson et al. 2000) (Figure 3 5). Increasing 18 O values in the Lake PetŽn

PAGE 70

70 Itz‡ core, together with tropical and northern hemisphere climate reconstructions, support the idea of reduced AMOC and a southward position of the ITCZ during the early part of the Mystery Interval (Denton et al. 2006; Broecker et al. 2009) and Heinrich event 1 (Bond et al. 1992, 1993; Hemming 2004). The Lake PetŽn Itz‡ 18 O record shows decadal to centennial E/P fluctuations during the early part of the Mystery Period (Figure 3 3). Centennial climate variability during this time period is also seen in Greenland (North Greenland Ice Core Project member s 2004), the North Pacific (Kiefer and Kienast 2005), and in a 18 O record from a speleothem collected on the Yucatan Peninsula (Gentry et al. 2008). Hodell et al. (1991) suggested that long term E/P changes during the late Pleistocene and Holocene in the Caribbean region were a consequence of orbitally forced (Milankovitch) insolation variations. Changes in the intensity of the annual cycle, for example, could explain the broad pattern of a dry late Pleistocene, a moister early Holocene and the drying tren d in the late Holocene (Brenner et al. 2001). Orbital forcing, however, cannot explain decada l to centennial E/P fluctuations seen in the Lake PetŽn Itz‡ 18 O record during the early part of the Mystery Interval. Similar variability in the North Atlantic during H1 has been explained by the freshwater input from ice armadas coming into the Northern Atlantic Ocean. Grousset et al. (2001) found that during H1 and H2, four to six abrupt meltwater discharges happened on century time scales. Climate models indicate that freshwater addition to the North Atlantic during the last deglaciation would have inf luenced the strength of the AMOC. Otto Bleisner and Brady (2009) simulated freshwater addition and results show that the shift of the ITCZ to a more northerly position after the freshwater forcing was faster

PAGE 71

71 (decades) than the recovery of the AMOC (200 500 years), and related to changes in low latitude SST (Otto Bleisner and Brady et al. 2009). High latitude temperatures and land sea ice extent in the northern Neotropics might drive millennial variations in the position of the ITCZ (Chiang and Bitz 2005), w hereas centennial to decadal ITCZ variations, such as variations seen during the first part of the Mystery Interval, might be more related to changes in low latitude SST gradients (Otto Bleisner and Brady et al. 2009). The first large, abrupt decrease in 18 O from ~16.5 to ~16.3 cal ka BP occurred in the midst of the Mystery Interval. If the 2.2 decline recorded in the Lake PetŽn Itz‡ oxygen isotope record during this transition were solely attributable to temperature, it would have required a temperature increase of 8.8 ¡C. Although there are no paleotemperature reconstructions for this time interval in the lowlands Neotropics, Lea et al. (2003) report no change in sea surface temperature during this period for the mari ne Cariaco Basin (Figure 3 5). Furthermore, a speleothem oxygen isotope record from the Yucatan Peninsula also shows a pronounced and abrupt decrease in 18 O values at that time (Gentry et al. 2008). Synchrony between these two 18 O records, constructed with independent chronologies, provides stron g evidence for a change in moisture source and/or an abrupt shift to more humid conditions in the Peten region during this period. There is, however, no evidence for enhanced precipitation near the Cariaco Basin (Peterson et al. 2000) at this time (Figure 3 5), pointing to a more local climate change. Lithologic shifts in the sediment core from Lake Peten Itza are thought to have been driven largely by E/P, with gypsum precipitation occurring under dry conditions

PAGE 72

72 when ion concentrations exceeded saturation, and clay deposited during moist periods, when the water column was undersaturated with respect to gypsum. A switch to clay deposition occurred shortly after the abrupt decrease in 18 O. Evidence for fairly constant temperatures in the Cariaco Basin (Lea e t al. 2003), together with synchroneity between changes in oxygen isotopes in the speleothem (Gentry et al. 2008) and lithological changes in the Lake PetŽn Itz‡ core, suggest a predominant role of E/P, rather than change in temperature or moisture source, to account for decreased 18 O in the lake record. The cause for this fast, but short lived decrease in 18 O values remains elusive. From ~16.3 15.0 cal ka BP, generally increasing 18 O values indicate a switch from relatively warm and/or humid conditions t o increasingly colder and/or drier conditions. Pollen assemblages also indicate dry conditions around Lake PetŽn Itz‡ during this interval (Bush et al. 2009). Further evidence for changes in the hydrological cycle comes from the Cariaco Basin core grey sca le record (Figure 3 5), and lithologic changes in sediment core from Lake PetŽn Itz‡ (Figure 3 3). The onset of gypsum precipitation and cessation of clay deposition halfway into the increasing 18 O trend indicates a reduction in lake water level large enough to initiate gypsum precipitation. Furthermore, there is no evidence for significant changes in sea surface temperature during this time period in the Cariaco Basin (Lea et al. 2003). Thus, changes in E/P probably accounted for most of the shift in 18 O. Data compiled by Broecker et al. (2009) point to a two part Mystery Interval. Paleoclimate r econstructions at paleolake Estancia, New Mexico, for example, show a wet LGM and a two part Mystery Interval with a dry interval between 14.9 and 13.8 14 C

PAGE 73

73 ka BP (Allen and Anderson 2000). This two part Mystery Interval is also evident in the CH 4 records f rom Antarctica (Monnin et al. 2001), the 18 O record from Greenland (North Greenland Ice Core Project members 2004) and the Cariaco grey scale record (Peterson et al. 2000) (Figure 3 6). The lake PetŽn Itz‡ oxygen isotope record shows a two part Mystery Interval with two distinct dry periods. Broecker et al. (2009) state that melting of two different ice armadas and further disruption of the AMOC could have caused the onset of the Mystery Interval and its division into two episodes. They used an ice rafted debris (IRD) record off south ern Portugal where two distinct events were identified at ~17.5 and ~16.1 cal ka BP (Bard et al. 2000). Melting of a second ice armada should reinforce the early dry conditions and thus contradicts evidence for an early, dry Mystery Interval followed by a late, wet Mystery Interval in the Western USA (Broecker et al. 2009). The climate impacts of a second ice armada, however, offer an explanation for the two part Mystery Interval seen in Greenland and the Neotropics (Figure 3 6). The melting of a second ice armada would have brought even colder conditions to the North Atlantic and Greenland, and caused a more southerly position of the ITCZ in the Neotropics, with drier conditions in Cariaco and the Yucatan Peninsula (Figure 3 6). The Lake Peten Itza !18O record adds to the growing number of paleoclimate records that show a two part Mystery Interval. The second large, abrupt decreases in 18 O occur from ~15.0 14.4 and at ~14.4 13.1 cal ka BP, with the latter episode being more g radual. This decline in 18 O coincides with the onset of clay deposition, and marks the second transition to warmer and/or wetter climate conditions of the Bolling Allerod period. This interval coincides with the resumption of modern, deep ocean circulatio n (McManus et al. 2004). The

PAGE 74

74 transition in Lake PetŽn Itz‡ coincides with a switch to higher sea surface temperatures (Lea et al. 2003), and an increase in precipitation (Peterson et al. 2000) in the Neotropics (Figure 3 5), as well as higher temperatures in high latitude regions (Figure 3 4) (North Greenland Ice Core Project members 2004). If the 2.6 decline recorded in the Lake PetŽn Itz‡ oxygen isotope record during this transition were solely attributable to temperature, it would have required a temper ature increase of 10.4 ¡C. Although there are no paleotemperature reconstructions for this time interval in the lowland Neotropics, Lea et al. (2003) report an increase in sea surface temperature of ~3.5 ¡C during this period for the marine Cariaco Basin ( Figure 3 5). Although there might be a temperature component, with higher temperatures reached in continental regions than the ocean, to the abrupt decrease in 18 O in the Lake PetŽn Itz‡ record, changes in E/P probably account for the bulk of the 18 O change. Furt hermore, the decrease in 18 O coincides with a switch from gypsum to clay. This concurrent shift in lithology argues for the predominant role of E/P, rather than temperature, on the isotopic record during these wetter periods. A small return to higher 18 O values, i.e. a switch to colder and/or drier conditions and the onset of gypsum deposition at ~13.1 cal ka BP, marked the beginning of the Younger Dryas (YD) (Figure 3 3). During the YD, Cariaco also experienced a return to colder and drier conditions. Le a et al. (2003) report an abrupt decrease in temperature of 4.5 ¡C at ~12.9 cal ka BP and Peterson et al. (2000) show reduced precipitation around the same time. A decline in North Atlantic Deep Water (NADW) production (Hughen et al. 2000) and AMOC (McManu s et al. 2004) during the YD would have caused a more southerly position of the ITCZ, with reduced precipitation in the northern

PAGE 75

75 Neotropics (Peterson et al. 2000; Haug et al. 2001) and increased rainfall in the southern Neotropics (Baker et al. 2001). Cold er and/or drier conditions in the lowlands of Guatemala during the YD are consistent with paleoclimate reconstructions (Haug et al. 2001; Baker et al. 2001) and climate model results (Schiller et al. 1997; Vellinga and Wood 2002). The termination of the YD observed at ~11.5 cal ka BP in geographically distant regions (North Greenland Ice Core Project members 2004; Haug et al. 2001), is seen in the PetŽn Itz‡ 18 O data, but not in the magnetic susceptibility record (Figure 3 3). A gradual decline in 18 O fr om ~11.5 to ~10.6 cal ka BP, followed by the third abrupt decrease in 18 O from ~10.6 to ~10.3 cal ka BP, indicates a change to warmer and/or wetter conditions at the end of the predominantly dry deglacial. Although the pollen data suggest an increase in r ainfall at that time (Bush et al. 2009), gypsum precipitation, which indicates low water levels, continued until the end of the Preboreal period at 10.3 cal ka BP (Hodell et al. 2008). The discrepancy between the isotopic and lithologic data can be reconci led if we accept the notion that water levels began to increase around 11.5 cal ka BP, inferred from the decrease in 18 O, but only attained a depth sufficient to reduce ionic concentration and terminate gypsum precipitation, at 10.3 cal ka BP. Isotope res ults from the PetŽn Itz‡ sediment core are in agreement with hydrological proxies Ti and Fe from the Cariaco Basin that show wetter conditions during the period 11.5 10.5 cal ka BP (Haug et al. 2001), presumably as a consequence of a more northerly positi on of the ITCZ. Temperature in the Cariaco Basin (Lea et al. 2003) increased ~3.5 ¡C from ~11.7 11.5 cal ka BP and remained nearly constant thereafter (Figure 3 5). The 18 O decrease in PetŽn Itz‡ of 2.4 from

PAGE 76

76 ~11.5 to ~10.3 cal ka BP would have required a temperature increase of 9.6 ¡C had there been no change in E/P. Temperature changes inferred from nearby marine sediment records, together with the calculated t emperature increase that would have had to occur to account for the change in the Lake PetŽn Itz‡ isotope record, point to the principal role of E/P on the isotope signal during the last, abrupt 18 O transition. Climate records from Greenland (North Greenl and Ice Core Project members 2004), the North Atlantic (Jansen and Veum 1990) and Antarctica (Jouzel et al. 1995) all exhibit a "two step" behavior during the last deglaciation. Deglacial climate in the lowlands of Central America evidently differed from c limate elsewhere. The PetŽn Itz‡ isotope record displays three large, abrupt climate transitions. This continental, "three step" deglacial record was a consequence of combined hemispheric and regional climate forcings. Conclusions Our results illustrate t he connection between North Atlantic and tropical Atlantic climate during the deglacial. Periods of reduced AMOC and cool SST temperatures in the North Atlantic coincide with a more southerly position of the ITCZ and thus colder and/or drier climate in the lowlands of northern Central America (The Oldest Dryas and the Younger Dryas). Colder and/or drier conditions around PetŽn Itz‡ are recorded during the late part of the Mystery Interval. Similar trends are also seen in the isotope record from Greenland an d the gray scale record from the Cariaco Basin. Deglacial climate inferred for the lowlands of Central America differed from paleoceanographic climate reconstructions, which show a "two step" deglaciation. In the PetŽn Itz‡ record, the cause of the first s hift in the midst of the Mystery Interval remains elusive, whereas the other two shifts mark the start of the Bolling Allerod Period, and the end of deglacial

PAGE 77

77 aridity. Future research will focus on trying to tease apart the relative impacts of temperature and E/P on the Lake PetŽn Itz‡ oxygen isotope record.

PAGE 78

78 Table 3 1. Age depth points used to derive chronology shown in Fig ure 3 2. Depth Depth Fairbanks et al. 05 Intcal 04 Site in site in site 6 Age 14 C Age cal Age cal mcd mcd yr B.P. (1s ) yr B.P. (1 s) yr B.P. (1 s) 2 13.06 10.47 7835 30 8603 27 8614 38 3 11.32 10.69 8675 35 9602 46 9619 48 6 10.86 10.86 9040 35 10213 16 10213 22 3 14.17 13.06 11210 35 13048 43 13118 47 6 13.35 13.35 11290 60 13125 72 13177 54 6 13.41 13.41 11380 14 0 13229 154 13258 134 6 13.87 13.87 11390 50 13236 74 13255 48 2 17.74 14.26 11880 35 13719 48 13747 44 6 14.49 14.49 12460 60 14396 150 14535 196 2 20.71 16.81 13095 40 15252 109 15489 159 2 21.91 17.55 13480 45 15693 119 16035 188 2 22.13 17.68 135 60 45 15784 119 16143 190 2 25.74 21.20 15355 50 18595 68 18723 54 6 23.64 23.64 17650 240 20896 321 20896 343 2 28.62 24.63 19740 70 23589 115 23641 145 6 25.73 25.73 19990 180 23880 217 23939 243 2 31.75 27.85 21940 80 26399 131

PAGE 79

79 Fig ure 3 1. (A) Map showing the locations of Lake PetŽn Itz‡, northern Guatemala, and paleoclimate reconstruction sites discussed in this paper: Cariaco (Lea et al. 2003 and Peterson et al. 2000), VM28 122 (Schmidt et al. 2004 ), OCE 326 GGC5 (McManus et al. 2004), SU 8118 (Bard 2002), NGRIP ( North Greenland Ice Core Project members 2004), and Dome C oncordia (Dome C) (Monnin et al. 2001). (B) Bathymetric map of Lake PetŽn Itz‡ showing the location of drill sites PI 2, PI 3, and PI 6.

PAGE 80

80 Figure 3 2. (A) Calibrated rad iocarbon dates versus mcd for sites PI 3 (blue) and PI 6 (red) used to construct the age model used by Hodell et al. (2008). (B) Calibrated radiocarbon dates versus meters composite depth (mcd) for sites PI 3 (blue), PI 6 (red), and PI 2 (yellow) used to c onstruct the age model used in this paper. Age model was derived by linear interpolation between selected age depth points.

PAGE 81

81 Figure 3 3 Magnetic susceptibility (red) and 18 O (blue) in sediment core PI 6. 18 O data indicated by open circles are from Lake PetŽn I tz‡ core 11A (Hillesheim et al. 2005). Cores 11A and PI 6 were correlated using an ash layer in both cores at the Pleistocene Ho locene transition. Dated climate periods are based on Greenland Ice Core Chronology 2005 (GICC05) (Rasmussen et al. 2006). 2bk: ages relative to A.D. 2000.

PAGE 82

82 Fig ure 3 4. 18 O in sediment core PI 6 and core 11A (blue). 18 O data indicated by open circles a re from Lake PetŽn I tz‡ core 11A (Hillesheim et al. 2005) and are compared to the 18 O of the NGRIP (North Gre enland Ice Core Project members 2004) (green) and sea surface temperatures derived from alkenones in core SU8118 from the subtropical NE Atlantic (Bard 2002) (purple). Each record is plotted on its own, independent time scale. Dated climate periods are based on Greenland Ice Core Chronology 2005 (GICC05) (Rasmussen et al. 2006). 2bk: ages relative to A.D. 2000.

PAGE 83

83 Figure 3 5. 18 O in sediment core P I 6 and core 11A (blue). 18 O data indicated by open circles are from Lake PetŽn I tz‡ core 11A (Hillesheim et al. 2005) and are compared to sea surface temperat ure reconstructions (Lea et al. 2003) (orange) and col or reflectance (Peterson et al. 2000) (bro wn) from the Cariaco Basin. Each record is plotted on its own, independent time scale. Dated climate periods are based on Greenland Ice Core Chronology 2005 (GICC05) (Rasmussen et al. 2006). 2bk: ages relative to A.D. 2000.

PAGE 84

84 Figure 3 6. 18 O of the NGRIP (North Gre enland Ice Core Project members 2004) (green). 18 O in sediment core PI 6 (blue). Col or reflectance (Peterson et al. 2000) (brown) from the Cariaco Basin. Methane data (grey) from Dome C, Antarctica (Monnin et al 2001). Each record is plotted on its own, independent time scale. Light gray band represents the mystery Interval (17.5 14.5 cal ka) as define by Denton et al. (2006). Dark grey bands represent ice armada events at ~17.5 and ~16.1 cal ka BP reported by Bard et al. (2000).

PAGE 85

85 CHAPTER 4 I SOTOPE MEASUREMENTS OF SINGLE OSTRACOD V ALVES AND GASTROPOD SHELLS FOR CLIMATE R ECONSTRUCTION: EVALU ATION OF WITHIN SAMPLE VARIABILITY AND DETE RMINATION OF OPTIMUM SAMPLE SIZE Introduction Lake sediment cores from the Yucatan Peninsula have the potential t o contain high resolution paleoclimate records (Leyden et al. 1993; Brenner et al. 2002). The peninsula is composed of low lying Tertiary limestones (Wilson 1984). Thus, most regional water bodies are rich in calcium carbonate and often contain abundant re mai ns of calcite and aragonite shell bearing aquatic invertebrates (i.e. ostracods and gastropods). Oxygen isotope (! 18 O) analyses from gastropods and ostracods in Yucatan lake sediments have been used to infer past changes in the evaporation/precipitation (E /P) ratio, i.e. effective moisture availability (Curtis et al. 1996, 1998; Hodell et al. 1995, 2005). Assuming oxygen isotope equilibrium is achieved during calcite precipitation, two principal factors control the 18 O of calcium carbonate shells in freshw ater environments: (1) the lake water !18O at the time of shell formation, and 2) the temperature at which the shell formed. We assume temperatures on the Yucatan Peninsula did not change appreciably during the late Holocene and, therefore, changes in lake water 18 O were the major control on 18 O of shell carbonate over the last ~3.5 ka (Hodell et al. 2007). Assuming no major change in source waters entering a lake, 18 O of water in tropical, closed basin water bodies is determined by the relationship betw een hydrologic input (precipitation) and output (evaporation) (Covich and Stuiver 1974; Fontes and Gonfiantini 1967). During dry periods, when E/P is relatively high, lake water and shell carbonate 18 O are also high because enhanced evaporation of

PAGE 86

86 lighter H 2 16 O leaves the water body enriched in H 2 18 O. Conversely, during wetter times, the lake water displays relatively low 18 O. The 13 C of dissolved inorganic carbon (DIC) in lake water is affected by multiple processes (Leng and Marshall 2004), including: (1) the carbon isotope composition of inflow waters (Leng et al. 1999), (2) CO 2 exchange between the lake and the atmosphere (Talbot and Kelts 1986; Ricketts and Johnson 1996), and (3) photosynthesis respiration of aquatic plants and calcium carbonate prec ipitation (McKenzie and Hollander 1993). Furthermore, processes that occur near the lake bed (methanogenesis, sulfate reduction, and sediment diagenesis) affect DIC values in profundal waters. Thus carbon isotopes undergo fractionation several times before being incorporated into the carbonate shells of aquatic organisms. In lakes, 13 C of organic matter and carbonate in sediments can be used to track changes in carbon cycling and lacustrine productivity. In previous paleoclimate studies on the Yucatan Peni nsula, oxygen isotope records were developed using multiple ostracod or gastropod specimens from each stratigraphic level in lake sediment cores. The number of individual shells analyzed in these studies, >15 per stratigraphic level when possible, was chos en based on a study by Xia et al. (1997) that explored seasonal variability in the geochemistry of ostracod valves. Two of the taxa used in those studies, and also used here, display large differences in habitat preference and ecological requirements. The ostracod Cytheridella ilosvayi Daday is a benthic organism and its carapace is composed of calcite, whereas the gastropod Pyrgophorus coronatus Pfeiffer typically lives in the littoral zone and its shell is made of aragonite. Furthermore, the two taxa prod uce their

PAGE 87

87 shells by different processes. Ostracods molt and secrete new calcareous carapaces within very short time periods (Peypouquet et al. 1988), whereas gastropods continuously accrete new shell material (Curtis et al. 1996). Assemblages of ostracods and gastropods within a single stratigraphic layer are thus often composed of multiple generations that secreted their shells at different times. The rationale for analyzing multiple individuals from a sample is that it reduces variance that might otherwis e result from measuring single, short lived individuals, and thus provides a reliable record of low frequency climate trends (Curtis et al. 1996; Hodell et al. 2005). Few studies have attempted to determine how many shells or valves of individual organisms in each stratigraphic interval are necessary to identify low frequency climate trends, i.e. to yield a reliable mean climate signal for a stratigraphic layer, or what factors influence the required sample size. Previous investigations along these lines us ed ostracods and foraminifera, but focused on trace element variability (Chivas et al. 1985; De Deckker et al. 1988; Palacios Fest and Dettman 2001; Holmes 2008). A few such studies used stable isotopes on ostracods (Heaton et al. 1995; Xia et al. 1997). To our knowledge, only one study using the isotope signature of gastropods shells tried to address this fundamental question (Jones et al. 2002). No study has specifically considered the use of stable isotope variability among individuals to determine the required sample size. Heaton et al. (1995) found large variability of 18 O and 13 C in valves from modern monospecific ostracods taken in single sampling campaigns from lakes in Jamaica (Wallywash Great Pond) and Mexico (Lake P‡tzcuaro). Furthermore, stratigraphic 18 O measurem ents on individual ostracod valves from a sediment core taken in Wallywash

PAGE 88

88 Great Pond showed that variations in E/P over a period of ~100 yrs were equal in magnitude to changes recorded throughout the late Quaternary (Heaton et al. 1995). Xia et al. (1997) explored seasonal variability of ostracod valve geochemistry (! 18 O, 13 C, Mg/Ca, and Sr/Ca) in the species Candona rawsoni Tressler from lakes of the American Midwest. They found high intra annual variability in live populations. Xia et al. (1997) suggest ed that 10 15 valves might be needed from each stratigraphic interval to remove the influence of high frequency variations (seasonal changes) on the low frequency trend (long term climate change). Nevertheless, they did not explain explicitly how they arri ved at these values. Jones et al. (2002) looked at intra depth variability of 18 O in the gastropods Gyraulus piscinarum Bourguignat and Valvata cristata MŸller from a small lake in southwest Turkey. They plotted the number of shells analyzed individually vs. the range of 18 O and concluded that, on average, six individuals represent 87% of the true 18 O range. Holmes (2008) looked at sample variability in several ostracod species from non marine, marginal marine, and shallow marine environments. He measure d trace element chemistry (Mg and Sr) to assess how many individuals should constitute a sample. The statistical approach he used is also valid for determining the correct sample size for stable isotope analyses. Similar to earlier studies, Holmes (2008) f ound high variability among the single valves measured and concluded that variable numbers of ostracod valves, sometimes >10, although sometimes fewer, are necessary to distinguish low frequency changes from high frequency trends. Climate reconstructions i nferred from stable isotope measurements on sub fossil valves or shells commonly use multiple remains to constitute a sample. This approach

PAGE 89

89 helps to distinguish low frequency trends from high frequency variability. Furthermore, many ostracod species are to o small to make a reliable measurement on the mass spectrometer, thus requiring multiple individuals to constitute a sample. For example, the Kiel III carbonate preparation device attached to the Finnigan MAT 252 isotope ratio mass spectrometer at the Univ ersity of Florida requires a minimum of 20 $g per sample. Only one valve of the species C. ilosvayi is needed to achieve the required mass. As many as 12 valves, however, are needed when working with some species of the genus Lymnocythere Intra sample iso tope variability, which could potentially reveal information about past climate, is generally neglected. Ostracods and small gastropods have life spans that range from months to years (Meisch 2000). The range and variance of isotope values measured on a su ite of short lived individuals from a single stratigraphic level can be used to assess climate variability within the short period of sediment deposition. Koutavas et al. (2006) employed this approach to quantify past El Ni–o Southern Oscillation (ENSO) va riability by looking at the variance of 18 O among individual foraminifera from a marine core collected near the Galapagos Islands. They concluded that reduced variance of within sample 18 O measures reflected attenuated ENSO amplitude. To our knowledge, t his approach has not been used in a non marine context to evaluate high frequency continental climate variability. The objectives of this study were to (1) investigate "within horizon" variability in stable isotope signatures of ostracod valves and gastrop od shells from lake sediment cores collected on the Yucatan Peninsula, (2) determine the optimum sample size, i.e. number of individuals required to make low frequency climate reconstructions, and (3)

PAGE 90

90 evaluate the potential for using intra sample 18 O variability in ostracod and gastropod shells as a proxy measure of high frequency climate variability. Study Sites For this study, we used sediment cores from Lakes Punta Laguna, Chichancanab, and PetŽn Itz‡ (Figure 4 1). The lakes are all located on the Yucatan Peninsula, but display different morphometries (Table 4 1). Sediments from these lakes were chosen for several reasons. First, existing, well dated (AMS 14 C dates) oxygen isotope records from Lakes Punta Laguna and Chichancanab showed substant ial low frequency evaporation/precipitation (E/P) variability during the late Holocene. Second, sediments are replete with well preserved ostracod and gastropod remains. Third, all lakes are effectively closed basin lakes so that changes in the oxygen isot opic signature of ostracods and gastropods faithfully record changes in lake level fluctuations and thus past moisture availability in the region. Lastly, the lakes are located in the region where lowland Maya civilization developed, flourished and eventua lly collapsed (Hodell et al. 1995, 2007; Curtis et al. 1996). This offers the possibility of relating climate reconstructions to ancient Maya cultural changes detected in the archaeological record. Laboratory Methods We used sediment samples from cores co llected from Lakes Punta Laguna (Curtis et al. 1996) and Chichancanab (Hodell et al. 1995) in 1993, and from Lake PetŽn Itz‡ in 1999. Individual ostracod valves and gastropod shells were selected from 1 cm stratigraphic sediment sections that were shown to have b een deposited during both wet and dry climate periods, represented by low and high 18 O, respectively. Individual valves of adult specimens of the ostracod C. ilosvayi were picked from the sieved 425 500 $m fraction in each 1 cm sample, using a fine paint brush under a low power

PAGE 91

91 binocular microscope. Prior to isotope analysis, ostracod valves were cleaned using 15% H 2 O 2 to remove organic material, and rinsed in methanol before drying. Valves were then checked for impurities using a binocular microscope. Ind ividual gastropods, when present, were also picked from the same stratigraphic horizons used for stable isotopic analysis of ostracods. Single specimens of P. coronatus were picked from the sieved sediment and cleaned in the same manner used for ostracods. Gastropod shells were individually crushed to look for impurities inside the shell, further cleaned when necessary, and ground to a fine powder using a mortar and pestle. Individual ostracod valves were weighed and loaded into glass vials. Likewise, groun d carbonate from each gastropod was weighed and placed into a glass vial. Carbonate samples were measured with a single aliquot acid digestion in a Kiel III carbonate preparation device attached to a Finnigan MAT 252 isotope ratio mass spectrometer at the University of Florida. Ground carbonate from a single gastropod in each stratigraphic horizon was run five times to evaluate homogenization of the carbonate powder. Analytical variability was measured by replicate analysis of the standard material NBS 19 ( 18 O = 2.20, 13 C = 1.95). The mean and standard deviation of NBS 19 were calculated for all single day runs as well as for the entire study. Optimum sample size, defined as the minimum number of individuals that achieves acceptable error for each stratigr aphic level, was calculated using the formula: n = (t x s/E) 2 (4 1) where t is the student's t value for a given level of significance, s is the standard deviation, and E is the acceptable error (Holmes 2008). The level of significance for all stratigraphic layers was set at = 0.01. The acceptable error was chosen using the

PAGE 92

92 variability seen in the stratigraphic sequence in its entirety and set at 10% of t he total variability. A smaller acceptable error (e.g. 5%) generates a larger optimum number of individuals for a sample, whereas a larger acceptable error (e.g. 15%) generates a smaller optimum number of individuals for a sample. The method used in this s tudy to estimate sample size was retrospective in that each sediment sequence had been analyzed previously. Comparison of stable isotope variability for the two taxa in e ach stratigraphic level (i.e. 18 O in C. ilosvayi vs. 18 O in P. coronatus ) was achieved using the F test. The statistical approach for calculating optimum sample size and applying the F test requires that the statistical population is distributed normally This requirement was verified using standard "goodness of fit" tests. All reported variability is 1 standard deviation (1 sd). Results Eight hundred and thirty eight individual stable isotope (! 18 O, 13 C) measurements from 51 stratigraphic horizons were analyzed in this study. The NBS 19 standard deviation on single day runs (n = 8 for all runs) ranged from 0.02 to 0.12 for 18 O and from 0.01 to 0.11 for 13 C. Summary statistics on NBS 19 for the length of the study (n = 253) yielded a mean and stand ard deviation = 2.190.08 for 18 O, and 1.960.04 for 13 C. In the Lake Punta Laguna core, 21 stratigraphic layers were picked for the ostracod C. ilosvayi and 10 stratigraphic layers for the gastropod P. coronatus The number of individual measurements (n) from each stratigraphic layer ranged from 7 to 30 for C. ilosvayi and from 8 to 30 for P. coronatus (Table 4 2). The average weight for an individual C. ilosvayi valve was 28 g (n=370), and there was little correlation between valve weight and 18 O ( r=0.16, p<0.01) and 13 C (r=0.28, p<0.01)

PAGE 93

93 measurements. Standard deviation values within stratigraphic levels in Lake Punta Laguna ranged from 0.46 to 1.89 for 18 O and from 0.47 to 1.74 for 13 C measurements on C. ilosvayi (Tables 4 2 and 4 3). Standa rd deviation values on P. coronatus ranged from 0.19 to 1.47 for 18 O and from 0.33 to 1.41 for 13 C (Tables 4 2 and 4 3). Standard deviation on individual P. coronatus which were run five times to check for sample homogenization, ranged from 0.02 to 0.10 for both 18 O and 13 C. Optimum sample size values in Lake Punta Laguna ranged from 9 to 138 for 18 O and from 6 to 60 for 13 C measurements on C. ilosvayi (Tables 4 2 and 4 3). Optimum sample sizes for P. coronatus ranged from 1 to 43 for 18 O and from 2 to 34 for 13 C (Tables 4 2 and 4 3). In the Lake Chichancanab core, four stratigraphic layers were picked for the ostracod C. ilosvayi and 11 stratigraphic layers for the gastropod P. coronatus The number of individual measurements (n) from each st ratigraphic layer ranged from 9 to 19 for C. ilosvayi and from 7 to 25 for P. coronatus (Table 4 4). The average weight for an individual C. ilosvayi valve was 32 g (n=60), and there was no correlation between valve weight and 18 O (r=0.15) or 13 C (r=0.0 4) measurements. Standard deviation values in Lake Chichancanab ranged from 0.59 to 0.83 for 18 O and from 0.50 to 0.99 for 13 C measurements on C. ilosvayi (Tables 4 4 and 4 5). Standard deviation values on P. coronatus ranged from 0.24 to 1.14 for 18 O and from 0.31 to 1.12 for 13 C (Tables 4 4 and 4 5). Standard deviation on the individual P. coronatus which were run five times to check for sample homogenization, ranged from 0.04 to 0.13 for 18 O and from 0.02 to 0.08 for 13 C. Optimum sampl e size values in Lake Chichancanab ranged from 25 to 55 for 18 O and from 2 to 11 for 13 C measurements

PAGE 94

94 on C. ilosvayi (Tables 4 4 and 4 5). Optimum sample size values on P. coronatus ranged from 1 to 45 for 18 O and from 2 to 32 for 13 C (Tables 4 4 and 4 5). In the Lake PetŽn Itz‡ core, five stratigraphic layers were picked for the gastropod species P. coronatus The number of individual measurements (n) from each stratigraphic layer ranged from 10 to 27 for P. coronatus (Table 4 6). Standard deviation va lues in Lake PetŽn Itz‡ ranged from 0. 23 to 0.40 for 18 O and from 0.41 to 0.73 for 13 C measurements on P. coronatus (Table 4 6). Standard deviation on the individual P. coronatus which were run five times to check for sample homogenization, ranged from 0.03 to 0.07 for 18 O and from 0 .05 to 0.06 for 13 C. Optimum sample size values in Lake PetŽn Itz‡ ranged from 6 to 17 for 18 O and from 6 to 15 for 13 C measurements on P. coronatus (Table 4 6). Discussion Standard deviation on ostracod/gastropod measurements was one order of magnitu de greater than the standard deviation measured on NBS 19 during single day runs or over the entire study. Variability among individual P. coronatus measurements within a stratigraphic level was also one order of magnitude greater than the variability with in a single individual P. coronatus shell used to check for sample homogenization. There was little mass dependency of 18 O and 13 C measurements on C. ilosvayi from Lake Punta Laguna and no mass dependency from Lake Chichancanab. Thus, variability among individuals within a stratigraphic s ample is not explained by instrument precision, laboratory procedures (i.e. poor homogenization), or sample mass. Within Horizon Variability and Calculation of Optimum Sample Size Calculated optimum sample sizes for both isotopes in C. ilosvayi and P. coro natus indicate that "optimum" numbers vary appreciably throughout the core (Tables 4 2, 4 3,

PAGE 95

95 4 4, 4 5, 4 6). In many cases, the calculated optimum sample size value was greater than the number of measurements used to calculate the optimum value, creating s ome uncertainty with regard to the calculated optimum n value. Optimum sample size numbers sometimes differ greatly in layers that are only centimeters apart. In Lake Punta Laguna, optimum sample size numbers for 13 C analysis of C. ilosvayi ranged from 1 5 to 46 valves in layers only 1 cm apart (i.e. on average, ~6 years apart). Similarly, optimum "n" values for 18 O analysis of P. coronatus in Lake Chichancanab ranged from 1 to 22 individuals in contiguous layers (i.e. on average, ~20 years apart). This v ariability may be explained by differences in the ecology and life histories of the two organisms, about which virtually nothing is known, as well as the oxygen and carbon isotope behavior within the water body, changes in lake morphometry (surface area/vo lume), and sedimentation rate. For example, during water level high stands, Lake Punta Laguna possesses three connected basins with broad littoral areas. Biogenic calcium carbonate production occurs primarily in this extensive littoral zone. A reduction in water level of only 2 m isolates the three basins and reduces the littoral area (Hodell et al. 2007). Heterogeneous littoral areas show variable water isotopic compositions and may thus lead to large variability in the 18 O of individual ostracods and gas tropods shells. Optimum sample size numbers are large in some cases. In Lake Punta Laguna, at 130 cm core depth (cmcd) (~1109 AD), calculations indicate that 138 individual valves should be used to characterize the 18 O of the population and achieve the sp ecified level of significance and acceptable error. When calculating statistical variables for populations, sample size is important. Large samples yield greater statistical reliability.

PAGE 96

96 Optimum sample size, however, must achieve a balance among the number of remains available, statistical considerations, and practical limitations on the number of samples that can be analyzed. Cytheridella ilosvayi is a benthic ostracod, whereas the snail P. coronatus typically occupies the littoral zone. Mueller et al. (20 08) found that sediments in Lake PetŽn Itz‡ are characterized by gastropod rich carbonate silt to a water depth of ~23 m, whereas sediments in waters deeper than 23 m had no gastropods. Lake water variables may display greater variability in the littoral z one than in deeper a reas, thus one might expect gastropod species to display higher variability in isotopic values than ostracod species. Our results, however, show contrary results. For stratigraphic layers in which 18 O data were collected for both P. coronatus and C. ilosv ayi isotopic variance between taxa was significantly different (F test, = 0.05) in about half the sediment layers analyzed (Tables 4 2 and 4 4). In layers that displayed a significant difference, variance was generally greater for C. ilosvayi than for P coronatus Consequently, optimum sample size numbers were also larger for C. ilosvayi than for P. coronatus (Tables 4 2 and 4 4, Figure 4 2). These differences in variance are probably a consequence of the different processes of calcification. Ostracods moult 9 times before reaching adulthood, with growth of a new carapace (two valves) at each moult (Turpen and Angell 1971; Okada 1982) over very short time periods (2 3 days) (Peypouquet et al. 1988). After the last molt, ostracods can live up to 3 years ( Simstich et al. 2004) without any further addition of calcite to the shell. Gastropods, on the other hand, accrete new shell layers continuously and can live for more than a year. This difference in shell production enables ostracods to record short durati on isotopic values in lake water (i.e. days),

PAGE 97

97 whereas gastropods record isotope values in the water continuously, over longer periods of time (seasonal/annual). This may account for the larger 18 O variance among individuals of C. ilosvayi than P. coronatus In the littoral zone, macrophytes and phytoplankton discriminate against 13 CO 2 during carbon fixation, thus the DIC pool in the epilimnion becomes enriched in 13 C. This 13 C enriched DIC is incorporated into CaCO 3 shells of littoral organisms. We might therefore e xpect higher 13 C values for P. coronatus than for C. ilosvayi as P. coronatus is primarily confined to the littoral zone. Variance comparison (F test, = 0.05) between 13 C of P. coronatus and C. ilosvayi however, shows significant differences only in a few stratigraphic layers, with no clear pattern as to which organism has higher or lower variability (Tables 4 3 and 4 5, Figure 4 2). This lack of pattern in 13 C variability might be explained by the complexity of the carbon isotope system within a lak e, together with different food sources and physiological pathways for shell construction, as well as micro scale variations in the carbon isotope composition of the DIC in the littoral zone and sediment porewaters. Metabolic carbon comes from the animal's food s ource, which usually has a much lower 13 C value than DIC. Although physiological isotopic studies have shown that the metabolic source is <10% of the total carbon used in shell construction (McConnaughey et al. 1997; Shanahan et al. 2005), it could be an important factor in shell construction when 13 C values are lower than equilibrium values (McConnaughey et al. 1997; Dettman et al. 1999; Shanahan et al. 2005). High variability in 13 C values from ostracods and gastropods might therefore be a partial cons equence of different food sources used by these two taxa.

PAGE 98

98 Gradients of pore water 13 C could be another factor that influences the variability seen in the 13 C of ostracod valves. The 13 C of pore water is controlled by aerobic respiration and anaerobic de gradation of organic matter (von Grafenstein et al. 1999). Organisms thriving in the upper, ventilated zone should display more negative values, as oxidation of 13 C depleted organic matter will be dominant at these depths. On the other hand, production of 13 C depleted methane (and the accompanying 13 C enriched, co genetic CO 2 ) under anaerobic conditions in deeper sediments would produce more 13 C enriched 13 C values in ostracods living at these depths. The depth and thickness of these micro environments in the sediment column, and thus the 13 C gradients in DIC of porewaters, are controlled by the amount of organic matter, the oxygen concentratio n, and the porosity of the sediments (von Grafenstein et al. 1999). The degree to which these variables influence the 13 C gradient in porewater DIC varies from lake to lake and within lakes, as a function of water depth, oxygen distribution, and sediment t ype (von Grafenstein et al. 1999). Hence, there are many causes for variations in th e 13 C of benthic ostracods among and within lakes. Similar numbers of layers and similar numbers of individuals per layer were analyzed for P. coronatus in Lake Punta Laguna and Lake Chichancanab. Lake Chichancanab is larger than Lake Punta Laguna and the 18 O of water is less easily changed by shifts in E/P in the larger volume system. Dependency of water 18 O on the residence time of water may be reflected in the different 18 O standard deviations measured in gastropods from both lakes. Lake Punta Laguna has a larger range of 18 O standard deviation values (0.19 1.47) on snails than does Lake Chichancanab (0.24 1.14). The same trend is not seen in P. coronatus 13 C variability. Lake Punta Laguna

PAGE 99

99 displays a range of 13 C standard deviation (0.33 1.41), similar to that in Lake Chichancanab (0.31 1.12). 18 O Variability among Individuals within a Sample as a Proxy for High Frequency Climate Shifts Paleoclimate reconstructions that rely on 18 O measurements typically use multiple, composited ostracods or gastropods from each stratigraphic level in a lake sediment core. Multiple, mono specific individuals in a sample are used to reduce variability that would result from measuring single, short lived individuals. The rationale for measuring multiple individu als within a stratigraphic level is that the approach provides a reliable record of low frequency climate trends or that multiple individuals are required to attain sufficient sample mass (Holmes et al. 1997). Curtis et al. (1996) obtained oxygen and carbo n isotope data for the Punta Laguna core by measuring 18 O and 13 C on samples composed of multiple (>15) C. ilosvayi whereas in this study we report the mean oxygen and carbon isotope value for each stratigraphic level based on measurements of multiple, single ostracod valves from each level. Results from the two studies indicate that regardless of the approach used, average oxygen and carbon isotope values are very similar. Either approach captures the low frequency climate trends satisfactorily (Figure 4 3 ). We also compared 18 O values from composited gastropod samples (Curtis et al. 1996) to the mean 18 O values from multiple, individual gastropods removed from the same 10 depths in the Punta Laguna core. The correlation between data sets is strong (r=0. 71). This indicates that, like the findings with ostracods, both approaches yield similar inferences with respect to low frequency climate reconstructions.

PAGE 100

100 Oxygen isotope records from Lakes Punta Laguna (Curtis et al. 1996) and Chichancanab (Hodell et al 1995) provide strong evidence for climate drying coincident with the Classic Maya collapse. The Punta Laguna 18 O record shows two increases in 18 O during the Maya Terminal Classic Period, from ca. 750 to 850 AD and ca. 910 to 990 AD, and a relatively m oist period from ca. 850 910 AD (Curtis et al. 1996). These two increases in 18 O are greater than 2.0 and 1.5, respectively, for the C. ilosvayi 18 O record. Calculation of the optimum sample size for these two dry intervals, with acceptable error of 1. 5 and a 99% level of significance, yielded sample sizes of <10 individuals per stratigraphic level. In their high resolution (each cm, on average = ~9 yr sample spacing for this part of the record), low frequency climate reconstruction, Curtis et al. (199 6) used >15 ostracod valves per stratigraphic level. We thus conclude, with statistical certainty, that these two increases in 18 O during the Maya Terminal Classic Period reflect significant (p=0.01) climate shifts. High frequency climate information can also be obtained by looking at 18 O variability among single shells/valves from a stratigraphic horizon. This approach provides information on climate variability within the time of sediment deposition. In other words, mean 18 O, along with measures of var iability, can provide information on both low and high frequency climate trends. For closed basin lakes on the Yucatan Peninsula, low 18 O variability about the mean, based on multiple measurements of individual valves, indicates low variability in the ev aporation/precipitation ratio. On the other hand, high variability about the mean indicates relatively greater climate fluctuations during the period of sediment accumulation.

PAGE 101

101 Results from our study show that there are three stratigraphic layers in which t he 18 O variability of C. ilosvayi is larger than the variability of the mean values seen in the entire Punta Laguna record (Figure 4 4). If we assume that the variability of 18 O measurements on individual valves within a stratigraphic layer reflects clim ate variability within the time of sediment deposition, we can argue that climate variability during that very short time period was larger than the climate variability seen in the entire sediment sequence. Heaton et al. (1995) also report larger variabili ty within stratigraphic layers than variability throughout the entire sequence. Stratigraphic 18 O measurements on individual ostracod valves from a sediment core taken in Wallywash Great Pond, Jamaica showed that variations in E/P over a period of ~100 yrs w ere equal in magnitude to changes recorded throughout the late Quaternary (Heaton et al. 1995). Data from the Punta Laguna core show that high mean 18 O values are associated with low variability, whereas low mean 18 O values are associated with high varia bility (Figures 4 4 and 4 5). These results indicate that sub decadal (on average, 1 cm = ~6 yr for the entire core), relatively dry periods were constantly dry, whereas relatively wet periods consisted of wet and dry times. This interpretation of high var iability climate is in agreement with historical precipitation records from the Yucatan Peninsula (Giddings and Soto 2003). Periods of protracted precipitation deficit are relatively common, whereas extended periods of excessive rainfall are not as evident (Giddings and Soto 2003). Furthermore, one might expect relatively larger fluctuations in 18 O during dry times when lake level is low. During such episodes, when lake stage is low, small variations in rainfall can generate rather substantial fluctuations in lake volume and the

PAGE 102

102 18 O of water. The low variability during the dry episodes is thus indicative of persistent dry conditions, uninterrupted by periods of higher precipitation. The mean and variability of 18 O data for the time period of the Classic M aya collapse (Figures 4 4 and 4 5) yield the same inference. That is, high mean 18 O values, indicating dry conditions, are mostly associated with low 18 O variability, whereas low mean 18 O values, indicating moister conditions, are associated with high 18 O variability. Thus, the Terminal Classic period from ca. 910 to 990 AD was not only the driest period in the last 3,000 years (Curtis et al. 1996), but also a persistently dry period. The 18 O record from Aguada X'caamal (Hodell et al. 2005), in the nor thwestern sector of the Yucatan Peninsula, revealed that climate became drier at the start of the Little Ice Age (LIA). High frequency climate reconstruction during that period for the Lake Punta Laguna sediment record (Figure 4 4) shows, once more, that h igh mean 18 O values are typically associated with low 18 O variability, whereas low mean 18 O values are associated with high 18 O variability. This suggests that the beginning of the LIA was persistently dry and probably uninterrupted by periods of highe r precipitation. Regional drying during the beginning of LIA is also recorded in pre Conquest documents from Yucatan (Gill 2000). The Book of Chilam Balam, a Maya chronicle, describes cold climate conditions, starvation and the destruction of the city of M ayapan during Katun 8 Ahau (AD 1441 1460) (Craine and Reindorp 1979). Although this method for high frequency climate reconstruction has been used on ocean sediment records (Simstich et al. 2004; Koutavas et al. 2006), its validity in continental records n eeds further verification. Our high frequency climate interpretation

PAGE 103

103 based on 18 O measurements of single ostracod valves could be influenced by several variables, including number of valves analyzed per stratigraphic horizon, changes in sedimentation rate and spatial variability of lake water 18 O. The standard deviation of a population is a measure of the variability or dispersion within that population. Thus, the standard deviation is influenced by the number of data points. A high mean 18 O, with a low standard deviation, might not reflect a consistently dry period. Instead, it may simply reflect the fact that insufficient valves were analyzed to estimate the true distribution of values in the population. Ideally, the same number of valves should be ana lyzed individually in all stratigraphic horizons. This, however, is seldom possible, as the number of valves available varies throughout the core. Results from this study show that there is no relation between the number of valves analyzed per stratigraphi c level and the standard deviation of isotopic values (Figure 4 6). For example, in Lake Punta Laguna, several stratigraphic layers had 19 individual valves of C. ilosvayi analyzed, but show a broad range of standard deviations (Figure 4 6). Sedimentation rate may influenc e 18 O variability within a single stratigraphic level. If the sedimentation rate within a given sample interval, e.g. 1 cm thick, was slow, climate may have changed during the relatively long time period. Thus, during periods of slow sediment accumulation one might expect to find greater isotope variability. Results from this study, however, suggest there is no relation between sedimentation rate and variability of isotopic measures (Figure 4 7). Indeed, the core section encompassing the Classic Maya coll apse has the lowest sedimentation rate among all layers (Table 4 2) and the lowest oxygen isotope variability. These results support the claim for a persistent dry period. Nevertheless, the data must be evaluated critically. For

PAGE 104

104 instance, age/depth relatio n for the Lake Punta Laguna core, and hence sedimentation rates, were derived by linear interpolation between AMS 14 C dates, but sedimentation rate may have fluctuated between dated intervals (Curtis el al. 1996). Spatial variability in lake water 18 O cou ld override any high frequency climate signal inferred from individual 18 O measurements. Ostracods and gastropods live and produce their shells in different areas of the lake that have different water 18 O values. Changes in E/P influence lake water level s and spatial variability of lake water 18 O. A shallower lake is more prone to mixing and thus to have more homogenous 18 O water values throughout the water column. On the other hand, a deep, thermally stratified lake could display marked isotopic differ ences between the epilimnion and hypolimnion, with ostracods thriving at all water depths and in the littoral zone. Thus, variability in the 18 O measures on single ostracod valves from a horizon could reflect high frequency climate changes, but may also b e attributable to spatial variability in lake water 18 O over short time spans. Few calibration studies have looked at 18 O variability in modern ostracods from different areas within a lake (von Grafenstein et al. 1999), and there are no such studies from the Yucatan Peninsula. A modern calibration study in Yucatan lakes is needed to validate and test our high frequency climate interpretation. These data contribute to our understanding of paleoclimate on the Yucatan Peninsula and its relation to ancient Ma ya culture. Our findings confirm the interpretations of Hodell et al. (1995, 2007) and Curtis et al. (1996), that there were persistent dry climate episodes associated with the Terminal Classic Maya Period. Nevertheless, an important question that remains is whether reductions in rainfall came at critical times in the annual agricultural cycle (Brenner et al. 2002). Further information

PAGE 105

105 on the life history, especially the seasonality of the ostracod species used in these studies might help answer this questi on. Unfortunately, few studies have addressed the seasonal behavior of ostracods (Ferguson 1944; McGregor 1969; Mclay 1977; Xia et al. 1997) and nothing is known about the life history of C. ilosvayi Conclusions We calculated the optimum number of indivi duals that should be used to constitute a sample for measurement of 18 O and 13 C in C. ilosvayi and P. coronatus in paleoclimate studies. Our results indicate that the optimum value varied throughout the cores we studied. The number of shells/valves neede d to reconstruct low frequency climate trends is probably influenced by (1) the ecology and life history of the organisms measured, (2) lake morphometry, in conjunction with inter annual climate variability in the study region, (3) sedimentation rate, biot urbation, and sampling interval, and (4) the statistical confidence level set by the researcher. Ideally, pilot studies should be undertaken prior to using multiple ostracods or gastropods in paleoclimate reconstructions. These studies should evaluate mode rn isotope variability by measuring multiple individual shells or valves, as well as past variability, using several stratigraphic layers. If information is available from other paleoclimate proxy variables, the stratigraphic layers selected for initial ev aluation should be chosen to reflect contrasting E/P conditions. Otherwise, preliminary measurements should be made on stratigraphic layers distributed evenly over the length of the sediment record. Results from the pilot study allow calculation of mean is otope values and thus provide a rough estimate of the low frequency variability over the entire sediment sequence. The range of variability is necessary to establish the acceptable error needed to calculate the optimum sample size. Calculated optimum

PAGE 106

106 sampl e sizes will vary throughout the core and can only be determined retrospectively. We therefore recommend the use of the largest optimum sample size calculated for the entire sequence. Intra sample isotope variability might reveal important high resolution climate information. The range and variance of isotope measures on ostracods from a single stratigraphic level can be used to assess climate variability within the short period of sediment deposition. Results from this study indicate that relatively dry pe riods were persistently dry, whereas relatively wet periods were composed of wet and dry times. Paleoclimate investigations in regions with marked summer and winter precipitation, such as the Yucatan Peninsula, will benefit from detailed knowledge of the e cology and life histories of the ostracod and gastropod species used for climate reconstructions.

PAGE 107

107 Table 4 1. Morphometric characteristics for the study lakes. Punta Laguna Chichancanab PetŽn Itz‡ Latitude 20 38' N 19 52' N 16 55' N Longitude 87 37 W 88 46' W 89 50' W Area (km 2 ) 0.9 5.3 100 max depth (m) >20 15 160

PAGE 108

108 Table 4 2 Summary statistics for single valve ostracod and single valve gastropod 18 O datasets from Lake Punta Laguna. cmcd: centimeters core depth, SD: standard deviation, var: variance, *variance between P. coronatus and C. ilosvayi is significantly different (" =0.05) Punta Laguna Pyrgophorus coronatus Cytheridella ilosvayi Var Depth Age sed rate 18 O n n 18 O n n cmcd (yr) mean SD Var optimum mean SD Var optimum 75 1414 AD 8 1.48 0.46 0.21 10 9 77 1399 AD 8 0.85 1.32 1.75 2 4 55 81 1368 AD 5 1.50 0.67 0.45 7 25 88 1331 AD 5 0.30 0.96 0.92 18 17 0.84 1.69 2.84 17 97 130 1109 AD 5 2.94 1.89 3.56 12 138 140 1056 AD 5 1.01 0.19 0.04 8 1 0.61 0.52 0.27 9 12 141 1051 AD 5 0.70 0.97 0.94 9 42 148 1003 AD 9 0.85 0.89 0.79 19 26 150 986 AD 9 0.95 0.56 0.31 19 10 152 968 AD 9 1.14 0.61 0.38 14 14 158 915 AD 9 0.50 0.84 0.70 21 13 0.43 1.38 1.91 18 64 159 906 AD 9 0.60 1.10 1.21 24 38 160 897 AD 9 0.51 1.28 1.63 22 53 161 888 AD 9 0.01 1.44 2.09 12 80 180 720 AD 9 1.23 0.64 0.42 10 10 0.83 1.31 1.72 19 57 195 587 AD 9 0.76 0.90 0.81 10 34 277 122 AD 6 1.54 1.05 1.11 15 22 0.78 1.26 1.58 14 57 318 108 BC 6 1.37 1.20 1.44 15 28 1.13 0.94 0.88 19 29 342 243 BC 6 1.96 0.64 0.41 14 8 387 487 BC 5 1.48 1.14 1.30 15 26 0.32 0.90 0.81 19 27 574 1342 BC 5 0.83 0.59 0.35 30 6 0.76 0.68 0.46 19 15 592 1424 BC 5 1.14 1.47 2.15 15 43 0.94 0.98 0.96 3 0 29

PAGE 109

109 Table 4 3 Summary statistics for single valve ostracod and single shell gastropod 13 C datasets from Lake Punta Laguna. cmcd: centimeters core depth, SD: standard deviation, var: variance, *variance between P. coronatus and C. ilosvayi are sig nificantly different ("=0.05) Punta Laguna Pyrgophorus coronatus Cytheridella ilosvayi Var Depth Age sed rate 13 C n n 13 C n n cmcd (yr) mean SD Var optimum mean SD Var optimum 75 1414 AD 8 3.17 0.48 0.23 10 6 77 1399 AD 8 5.19 1.74 3.01 24 56 81 1368 AD 5 4.05 0.47 0.22 7 7 88 1331 AD 5 2.82 0.88 0.77 18 13 4.91 1.41 1.98 17 40 130 1109 AD 5 5.42 1.62 2.63 12 60 140 1056 AD 5 1.98 0.47 0.22 8 5 3.14 0.75 0.56 9 15 141 1051 AD 5 3.17 1.32 1.73 9 46 148 1003 AD 9 2.53 0.96 0.92 19 18 150 986 AD 9 2.56 1.13 1.28 19 25 152 968 AD 9 2.62 0.83 0.69 14 15 158 915 AD 9 2.44 0.47 0.22 21 3 3.17 0.80 0.64 18 13 159 906 AD 9 2.81 1.04 1.08 24 20 160 897 AD 9 3.30 1.19 1.41 22 27 161 888 AD 9 2.87 1.11 1.22 12 28 180 720 AD 9 3.23 0.90 0.80 10 16 3.50 1.02 1.04 19 20 195 587 AD 9 1.47 1.35 1.82 10 45 277 122 AD 6 3.06 0.80 0.64 15 11 3.25 1.51 2.27 1 4 49 318 108 BC 6 2.30 0.84 0.70 15 12 2.84 0.71 0.50 19 10 342 243 BC 6 3.34 0.33 0.11 14 2 387 487 BC 5 2.55 1.20 1.44 15 25 2.84 0.94 0.89 19 17 574 1342 BC 5 2.53 0.64 0.42 30 6 3.77 0.61 0.37 19 7 592 1424 BC 5 3.59 1. 41 1.98 15 34 3.98 1.18 1.39 30 25

PAGE 110

110 Table 4 4 Summary statistics for single valve ostracod and single shell gastropod 18 O datasets from Lake Chichancanab. cmcd: centimeters core depth, SD: standard deviation, var: variance, *variance between P. coronatus and C. ilosvayi are significantly different ("=0.05) Chichancanab Pyrgophorus coronatus Cytheridella ilosvayi Var Depth Age sed rate 18 O n n 18 O n n cmcd (yr) mean SD Var optimum mean SD Var optimum 62 962 AD 17 2.55 1.0 1 1.02 16 22 63 945 AD 16 2.85 0.24 0.06 13 1 69 830 AD 21 2.73 0.33 0.11 16 2 72 768 AD 20 2.59 0.36 0.13 10 3 132 465 BC 20 1.40 1.06 1.13 21 23 2.31 0.75 0.56 9 55 142 670 BC 21 1.57 0.58 0.33 22 7 2.11 0.59 0.35 19 25 149 814 BC 21 1.22 0.73 0.53 25 10 1.71 0.83 0.69 16 52 151 855 BC 21 1.56 0.52 0.27 24 5 1.72 0.83 0.68 16 52 371 5376 BC 21 1.25 0.74 0.55 18 12 381 5582 BC 20 1.61 1.14 1.29 7 45 410 6178 BC 20 3.47 0.58 0.33 21 7

PAGE 111

111 Table 4 5 Summary statistics for single valve ostracod and single shell gastropod 13 C datasets from Lake Chichancanab. cmcd: centimeters core depth, SD: standard deviation, var: variance, *variance between P. coronatus and C. ilosvayi ar e significantly different ("=0.05) Chichancanab Pyrgophorus coronatus Cytheridella ilosvayi Var Depth Age sed rate 13 C n n 13 C n n cmcd (yr) mean SD Var optimum mean SD Var optimum 62 962 AD 17 0.87 1.12 1.27 16 32 63 945 AD 16 0.28 0.53 0.29 13 8 69 830 AD 21 0.89 0.31 0.09 16 2 72 768 AD 20 0.86 0.33 0.11 10 3 132 465 BC 20 0.63 0.90 0.81 21 19 2.48 0.99 0.98 9 11 142 670 BC 21 0.57 0.64 0.41 22 9 1.72 0.66 0.44 19 4 149 814 BC 21 0.49 1.05 1.11 25 25 1.91 0.50 0.25 16 2 151 855 BC 21 0.06 0.59 0.35 24 8 1.84 0.55 0.31 16 3 371 5376 BC 21 1.90 0.91 0.82 18 20 381 5582 BC 20 0.87 0.65 0.43 7 17 410 6178 BC 20 3.01 0.77 0.60 21 14

PAGE 112

112 T able 4 6 Summary statistics for single shell gastropod 18 O and 13 C datasets from Lake PetŽn Itz‡. cmcd: centimeters core depth, SD: standard deviation, var: variance PetŽn Itz‡ Pyrgophorus coronatus Depth Age 18 O n n 13 C n n cmcd mean SD Var optimum mean SD Var optimum 95 6354 BC 0.71 0.40 0.16 15 17 2.00 0.66 0.44 15 12 215 9061 BC 2.75 0.23 0.06 10 7 0.54 0.41 0.17 10 6 241 9809 BC 2.80 0.26 0.07 27 6 0.20 0.57 0.32 27 8 244 10106 BC 2.98 0.31 0.10 14 11 0.44 0.73 0.53 14 15 2 50 10699 BC 2.67 0.38 0.15 12 17 0.10 0.44 0.20 12 6

PAGE 113

113 Fig ure 4 1 Map of study area showing the location of the three study lakes

PAGE 114

114 Fig ure 4 2 Optimum sample size (i.e. number of ostracod valves or gastropod shells) in Lake Punta Laguna (squares) a nd Lake Chichancanab (circles) for P. coronatus (open symbol) and C. ilosvayi (closed symbol) for (a) 18 O and (b) 13 C measurements

PAGE 115

115 Figure 4 3. Intra sample variability as a proxy for high resolution climate reconstructions. (3a) 18 O values from comp osite samples, reported by Curtis et al. (1996) vs. average 18 O values calculated from measurements made on multiple, individual valves in this study. Dashed line indicates a 1 1 line. r = 0.95 (3b) 13 C values from composite samples, reported by Curtis e t al. (1996) vs. average 13 C values calculated from measurements made on multiple, individual valves in this study. Dashed line indicates a 1 1 line. r = 0.94.

PAGE 116

116 Figure 4 4. (A) Oxygen isotopic composition of the ostracod C. ilosvayi in the sediment core from Lake Punta Laguna, Mexico (Curtis et al. 1996). The oxygen isotope data were not smoothed as in Curtis et al. (1996). Horizontal lines at several stratigraphic horizons represent the range of 18 # measurements on individual C. ilosvayi valves. Dates are based on the chronological model constructed on samples obtained from the sediment core (PL 22 VI 93) recovered in 1993 in 6.3 m of water column from the central basin of Punta Laguna (Curtis et al. 1996). (B) Detail of the Classic Maya collapse time p eriod. Horizontal lines at several stratigraphic horizons represent the range of 18 O measurements on individual C. ilosvayi valves. (C) Detail of the Little Ice Age period. Horizontal lines at several stratigraphic horizons represent the range of 18 O mea surements on individual C. ilosvayi valves.

PAGE 117

117 Figure 4 5. Intra sample variability as a proxy for high resolution climate reconstructions. (5a) mean 18 O values vs. standard deviation of 18 O values for C. ilosvayi in Lake Punta Laguna. (5b) mean 18 O val ues vs. standard deviation of 18 O values for C. ilosvayi in Lake Punta Laguna for the wet and dry periods during the time of the Classic Maya collapse. Numbers in the figure indicate years AD/BC. Dates are based on the chronological model constructed on s amples obtained from the sediment core (PL 22 VI 93) recovered in 1993 in 6.3 m of water column from the central basin of Punta Laguna (Curtis et al. 1996).

PAGE 118

118 Figure 4 6. Number of individual ostracod valves and gastropod shells analyzed vs. 18 O (closed circle) and 13 C (open circle) standard deviation per stratigraphic level. (A) C. ilosvayi in Lake Punta Laguna. (B) P. coronatus in Lake Punta Laguna. (C) C. ilosvayi in Lake Chichancanab. (D) P. coronatus in Lake Chichancanab.

PAGE 119

119 Figure 4 7. Sedimentatio n rates vs. 18 O (closed circles) and 13 C (open circles) standard deviation per stratigraphic level. (A) C. ilosvayi in Lake Punta Laguna. (B) P. coronatus in Lake Punta Laguna. (C) C. ilosvayi in Lake Chichancanab. (D) P. coronatus in Lake Chichancanab.

PAGE 120

120 CHAPTER 5 CONCLUSIONS Late Pleistocene and Holocene Climate Change in the Maya Lowlands: An Overview Between 3 February and 11 March, 2006, we retrieved 1327 m of sediment from seven sites in Lake PetŽn Itz‡. A complete sediment sequence that is 75.9 met ers long was recovered at site PI 6, and was the subject of study in this dissertation. The composite stratigraphic section for site PI 6 was constructed using the magnetic susceptibility and density results of the three cores drilled in PI 6. Sediment cor es comprising the composite section were subsampled and sediment was transported to the University of Florida (UF) for stable isotope analysis. The age/depth model for the LGM and deglacial period in the PI 6 core from Lake PetŽn Itz‡ was derived from 17 A MS 14 C dates from drill sites PI 2, PI 3 and PI 6. Dates from the PI 2 and PI 3 core were projected onto the PI 6 core by correlating their magnetic susceptibility records. Radiocarbon measurements were done at the Center for Accelerator Mass Spectrometry (CAMS) at Lawrence Livermore National Laboratory and dates were calibrated using IntCal04: Northern Hemisphere (Reimer et al. 2004) and Fairbanks et al. (2005). All radiocarbon dates were measured on samples of terrestrial organic matter, thereby avoiding potential dating errors associated with hard water lakes. Oxygen isotopes (! 18 O), lithology and pollen data from a sediment core taken in Lake PetŽn Itz‡ provide evidence for a relatively wet LGM in lowland Central America, bracketed by drier conditions during Heinrich events 2 and 1. Wetter conditions during the LGM, inferred from the Lake PetŽn Itz‡ isotope record, are in agreement with paleoglacier data from Mexico and Guatemala. Wetter conditions during the LGM in the

PAGE 121

121 lowlands of Central America may be explained by (1) an increase in summer precipitation due t o a more northerly position of the Intertropical Converge Zone (ITCZ), or (2) an increase in winter precipitation due to the splitting of the jet stream and/or movement of moist eastern tropical Pacific air to the northeast. Factors that influenced climate change on the Yucatan Peninsula during the LGM are still poorly understood, but modeling efforts may provide insights into the climate forcing factors at that time. Oxygen isotopes (! 18 O), lithology and pollen data from a sediment core taken in Lake PetŽn Itz‡ illustrate the connection between North Atlantic and tropical Atlantic climate during the deglacial. Periods of reduced AMOC and cool SST temperatures in the North Atlantic coincide with a more southerly position of the ITCZ and thus colder and/or dr ier climate in the lowlands of northern Central America (The Oldest Dryas and the Younger Dryas). Colder and/or drier conditions around PetŽn Itz‡ are recorded during the late part of the Mystery Interval. Similar trends are also seen in the isotope record from Greenland and the gray scale record from the Cariaco Basin. Deglacial climate inferred for the lowlands of Central America differed from paleoceanographic climate reconstructions, which show a "two step" deglaciation. In the PetŽn Itz‡ record, the ca use of the first shift in the midst of the Mystery Interval remains elusive, whereas the other two shifts mark the start of the Bolling Allerod Period, and the end of deglacial aridity. Future research will focus on trying to tease apart the relative impac ts of temperature and E/P on the Lake PetŽn Itz‡ oxygen isotope record. Oxygen isotopes (! 18 O) from sediment cores taken in Lakes PetŽn Itz‡, Punta Laguna, and Chichancanab indicate that the optimum value needed to reconstruct low frequency climate trends varied throughout the cores. The number of shells/valves

PAGE 122

122 needed is probably influenced by (1) the ecology and life history of the organisms measured, (2) lake morphometry, in conjunction with inter annual climate variability in the study region, (3) sedim entation rate, bioturbation, and sampling interval, and (4) the statistical confidence level set by the researcher. Intra sample isotope variability might reveal important high resolution climate information. Results from this study indicate that relativel y dry periods were persistently dry, whereas relatively wet periods were composed of wet and dry times. Paleoclimate investigations in regions with marked summer and winter precipitation, such as the Yucatan Peninsula, will benefit from detailed knowledge of the ecology and life histories of the ostracod and gastropod species used for climate reconstructions.

PAGE 123

123 LIST OF REFERENCES Allen BD, Anderson RY (2000) A continuous high resolution record of late Pleistocene climate variability from the Estancia basin, New Mexico. Geol Soc Am Bull 112:1444 1458 Anderson TH (1969) First evidence for glaciation in Sierra Los Cuchumatanes Range, northwestern Guatemala. Geological Society o f America Special Paper 121:387 Anselmetti FS, Ariztegui D, Hodell DA, Hillesheim MB, Brenner M, Gilli A, McKenzie JA, Mueller AD (2006) Late Quaternary Climate induced Lake Level variations in Lake PetŽn Itz‡ Guatemala, inferred from seismic strat igraphic analysis. Palaeogeogr Palaeocl P alaeoecol 230:52 69 Baker PA, Seltzer GO, Fritz SC Dunbar RB, Grove MJ, Tapia PM, Cross SL, Rowe HD, Broda JP (2001) The history of South American tropical precipitation for the past 25 ,000 years. Science 291:640 643 Ballantyne AP, Lavine M, Crowley TJ, Liu J, Baker PB (2005) Metaanalysis of tropical sur face temperatures during the Last Glacial Maxim um. Geophys Res Lett doi:10.1029/2004GL021217 Bard E (2002) Abrupt climate changes over millennial time scales: climat e shock. Physics Today 55:32 38 Bard E, Rostek F, Turon JL, Gendreau S (2000) Hydrological impact of Heinrich events in the subtropical Northeast Atlantic. Science 289:1321 1324 Bard E, Hamelin B, Arnold M, Montaggioni L, Cabioch G, Faure G, Rougerie F (1996) Sea level record from Tahiti corals and the timing of deglacial meltwate r discharge. Na ture 382:241 244 Binford MW, Brenner M, Whitmore TJ, Higuera Gundy A, Deevey ES, Leyden B (1987) Ecosystems, paleoecology and human disturbance in subtropical and tropical America Quarternary Sci Rev 6:115 128 Bond G, Showers W, Cheseby M, Lotti R, Almasi P, deMenocal P, Priore P, Cullen H, Hajdas I, Bonani G (1997) A pervasive millennial scale cycle in North Atlantic Holocene and glacial climates. Science 278:1257 1266 Bond GC, Hartmut H, Broecker WS, Labeyrie L, McManus J, Andrews J, Huon S, Jantschik R, Clasen S, Simet C, Tedesco K, Klas M, Bonani G, Ivy S (1992) Evidence for massive discharges of ice bergs into the North Atlantic Ocean during the last gla cial period. Nature 360:245 249 Bond GC, Broecker WS, Johnsen S, McManus JF, Labeyrie L, Jouzel J, Bonani G (1993) Correlation between climate records from North Atlantic sediments and Greenland ice. Nature 365:143 147

PAGE 124

124 Bradbury JP (1997) Sources of glacial moisture in Mesoamerica. Quate rn Int 43 44:97 110 Brenner M (1983) Paleolimnology of the PetŽn lak e district, Guatemala : II. Mayan population density and sediment and nutrient loading of Lake Quexil. Hydrobiologia 103:205 210 Brenner M, Rosenmeier MF, Hodell DA, Curt is JH (2002) Paleolimnology of the Maya Lowlands: Long term perspectives on interact ion s among climate, environment. and hu mans. Ancient Mesoam 13:141 157 Brenner M, Hodell DA, Curtis JH, Rosenmeier MF, Binford MW, Abbott M (2001) Abrupt Climate Change and Pre Columbian Cultural Collapse. In: Markgraf V (ed) Interhemispheric climate linkages pp 87 103. Broecker WS (2003) Does the trigger for abrupt climate change reside in the ocean or in the at mosphere? Science 300:1519 1522 Broecker WS (1998) Paleocean circulation during the last deglaciation: A bipolar seesaw? Paleoceanography 13:119 121 Broecker WS (1994) Massive iceberg discharges as triggers for global cli mate change. Nature 372:421 424 Broecker WS, Bond G, Klas M (1990) A salt oscillator in the glacial Atlantic? 1. The concept: Paleoceanography 5:469 477 Broecker WS, McGee D, Adams KD Cheng H, Lawrence ER, Oviatt CG, Quade J (2009) A Great Basin wide dry episode during the first half of the Mistery Interval? Quaternary Sci Rev doi:10.1016/j.quascirev.2009.07 .007 Bronk Ramsey C (2008) Deposition models for chronological records. Quater nary Sci Rev 27:42 60 Bronk Ramsey C (2001) Development of the Radiocarbon calibration program OxCal. Radiocarbon 43:355 363 Bromwich DH, Toracinta ER, Wei H, Oglesby RJ, Fastook JL, Hughes TJ (2004) Polar MM5 simulations of the winter climate of the laure ntide ice sheet at the LGM. J Climate 17:3415 3433 Bush MB, Correa Metrio A, Hodell DA, Brenner M, Anselmetti FS, Ariztegui D, Mueller AD, Curtis JH, Grzesik D, Burton C, Gilli A (2009) Re evaluation of climate change in Lowland Central America during the Last Glacial Maximum using new sediment cores from Lake PetŽn Itz‡, Guatemala. In: Vimeux F, Sylvestre F, Khodry M (eds) Past Climate Variability from the Last Glacial Maximum to the Holocene in South America and Surrounding Regions: Developments in Paleoe nvironmental Resear ch, Springer Verlag, pp 113 128

PAGE 125

125 Bush MB, Piperno DR, Colinvaux PA, Krissek L, De Oliveira PE, Miller MC, Rowe W (1992) A 14,300 year paleoecological profile of a lowland tropical lake in Panama. Ecol Monogr 62:251 276 Cacho I, Grimalt JO Pelejero C, Canals M, Sierro FJ, Flores JA, Shackleton N (1999) Dansgaard Oeschger and heinrich event imprints in Alboran Sea paleotemperatur es. Paleoceanography 14:698 705 Carlson AE, Oppo DW, Came RE, LeGrande AN, Keigwin LD, Curry WB (2008) Subtropica l Atlantic salinity variability and Atlantic meridional circulation during the last d eglaciation. Geology 36:991 994 Chiang JCH, Bitz M (2005) Influence of high latitude ice cover on the marine Intertropical Converge nce Zone. Clim Dynam 25:477 496 Chivas A R, De Deckker P, Shelley JMG (1985) Strontium content of ostracods indicates lacustrine pal aeosalinity. Nature 316:251 253 Clark PU, Mitrovica JX, Milne GA, Tamisiea ME (2002) Sea level fingerprinting as a direct test for the source of global meltwater p ul se 1a. Science 295:2438 2441 CLIMAP (1976) The surface of the ice a ge earth. Science 191:1131 1136 CLIMAP (1981) Seasonal reconstructions of the Earth's surface at the last glacial maximum. Geo l Soc Amer Map Chart Ser, MC 36 Covich AP, Stuiver M (1974) Ch anges in Oxygen 18 as a Measure of Long Term Fluctuations in Tropical Lake Levels and Molluscan Populations. Limnol Oceanogr 19:682 691 Craine ER, Reindorp C (1979) The Codex Perez and the Book of Chilam Balam of Mani University of Oklahoma Press. Curtis JH, Brenner M, Hodell DA, Balser RA, Islebe GA, Hooghiemstra H (1998) A multi proxy study of Holocene environmental change in the Maya lowlands of PetŽn, Guatemala. J Paleolimnol 19:139 159 Curtis JH, Hodell DA, Brenner M (1996) Climate Variability on the Yucatan Peninsula (Mexico) During the Past 3,500 Years, and Implications for Maya Cultural Evo lution. Quaternary Res 46:37 47 Dansgaard W (1964) Stable isotopes in precipitation. Tellus 16:436 468 Darling WG, Bath AH, Gibson JJ, Rozanski K (2006) Isotopes in water. Isotopes in lake sediments. In: Leng MJ (ed) Isotopes in Paleo environmental Research, pp 1 67

PAGE 126

126 De Deckker P, Chivas AR, Shelley JMG, Torgersen T (1988) Ostracod shell chemistry a new palaeoenvironmental indicator applied to a regressive transgres sive record from the Gulf of Carpentaria, Australia. Palaeogeogr Palaeoclimatol Palaeoecol 66:231 241 Deevey ES, Brenner M, Binford MW (1983) Paleolimnology of the PetŽn lake district, Guatemala : III. Late Pleistocene and Gamblian environments of the Maya area. Hydrobiologia 103:211 216 Deevey ES, Stuiver M (1964) Distribution of natural isotopes of carbon in Linsley Pond and other New England lakes. Limnol Oceanogr 9:1 11 Denton GH, Broecker WS, Alley RB (2006) The mistery Interval 17.5 to 14. 5 kyrs ago. P AGES News 14:14 16 Deschamps P, Durand N, Bard E, Hamelin B, Camoin G, Thomas AL, Henderson GM, Yokoyama Y (2009) The New Tahiti Sea Level Record (IODP Expedition 310) sets the Age of Melt Water Pulse 1A as Synchronous with the Bolling Warming. Abstract. G eological Associati on of Canada Joint Assembly Dettman DL, Reische AK, Lohmann KC (1999) Controls on the stable isotope composition of seasonal growth bands in freshwater bivalves (unionidae). Geochim Cosmochim Acta 63:1049 1057 EPICA Community Members (20 06) One to one coupling of glacial climate variability in Greenland and Antarctica. Nature 444:195 199 Escobar J, Hodell DA, Brenner M, Curtis JH (in prep) Climate conditions in the lowlands of Central America during the Last Glacial Maxium (LGM). Escobar J, Curtis J, Brenner M, Hodell DA, Holmes J (in press) Isotope measurement of ostracod and gastropod shells for climate reconstruction: evaluation of within sample variability and determination of optimum sample size. J Paleolimno l DOI 10.1007/s10933 009 9 377 9 Fairbanks RG, Mortlock RA, Chiu TC, Cao L, Kaplan A, Guilderson TP, Fairbanks TW, Bloom AL (2005) Marine radiocarbon calibration curve spanning 10,000 to 50,000 years B.P. based on paired 230Th/234U/238U and 14C dates on Pristine Corals. Quaternary S ci Rev 24:1781 1796 Fairbanks RG (1989) A 17000 year glacio eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep ocean circulation. Nature 342:637 642

PAGE 127

127 Farrera I, Harrison SP, Prentice IC, Ramstein G, Guiot J, Ba rtlein PJ, Bonnefille R, Bush M, Cramer W, von Grafenstein U, Holmgren K, Hooghiemstra H, Hope G, Jolly D, Lauritzen SE, Ono Y, Pinot S, Stute M, Yu G (1999) Tropical climates at the last glacial maximum: a new synthesis of terrestrial paleoclimate data, v egetation, lake levels and geochemistry. Clim Dynam 15:823 856 Ferguson E (1944) Studies on the Seasonal Life History of Three Species of Freshwater Ost racoda. Am Midl Nat 32:713 727. Flower BP, Hastings DW, Hill HW, Quinn TM (2004) Phasing of deglacial wa rming and Laurentide Ice sheet meltwater in the Gulf of Mexico. Geology 32:597 600 Fontes JC, Gonfiantini R (1967) Compor tement isotopique au cours de l'evaporation de deux bassins sahariens. Earth Planet Sc Lett 3:258 266 Friedli H, Lotscher H, Oeschger H, Siegenthaler U, Stauffer B (1986) Ice core record of the 13C/12C ratio of atmospheric CO2 in the past 2 centuries. Nature 324:237 238 Gentry C, Hodell DA, Curtis JH, Brenner M, Gallup C (2008) Centenial Scale Climate Variability During the Early Deglaci ation (19 to 16 ka) Expressed in Speleothem Oxygen Isotopes From the Yucatan Peninsula, Mexico. Abstracts, A merican Geophysical Union (AGU) Giddings L, Soto M (2003) Rhythms of precipitation in the Yucatan Peninsula. In: Gomez Pompa A, Allen MF, Fedick SL, Jimenez Osornio JJ (eds) The Lowland Maya: Three Millennia at the Human Wildland Interface. Haworth Press, Binghamton, pp 77 89 Gill RB (2000) The great Maya droughts: water, life, and death. University o f New Mexico Press, Albuquerque Grousset FE, Cortij o E, Huon S, Herve L, Richter T, Burdloff D, Duprat J, Weber O (2001) Zooming in on Heinrich Layers. Paleoceanography 16:240 259 Hanebuth T, Stattegger K, Grootes PM (2000) Rapid flooding of the Sunda shelf: A late glacial sea leve l record. Science 288:103 3 1035 Hastenrath S, Greischar L (1993) Further work on the prediction of northeast Brazil rainfall anomalies. J Climate 6:743 758 Hastenrath S (1984) Interannual variability and annual cycle: mechanisms of circulation and climate in the tropical Atlantic sector Mon Weather Rev 112:1097 1107 Hastenrath S (1973) On the Pleistocene glaciation of the Cordillera de Talamanca, Costa Rica. Zeitschrift fŸr Gletscherkunde und Glazialgeologie 9:105 121

PAGE 128

128 Haug GH, Hughen KA, Sigman DM, Peterson LC, Ršhl U (2001) South ward migration of the Intertropical Convergence Zone through the Holocene. Science 293:1304 1308 Heaton THE, Holmes JA, Bridgwater ND (1995) Carbon and oxygen isotope variations among lacustrine ostracods: Implications for paleoclimatic studies. Holocene 5 :428 434 Hemming SR (2004) Heinrich events: massive late Pleistocene detritus layers of the North Atlantic and their global climate imprint. Review of Geophysics, 42:RG1005 Hillesheim MB, Hodell DA, Leyden BW, Brenner M, Curtis JH, Anselmetti FS, Ariztegui D, Buck DG, Guilderson TP, Rosenmeier MF, Schnurrenberger DW (2005) Lowland neotropical climate change during the late deglacial and early Holocene. J Quaternary Sci 20:363 376 Hodell DA, Anselmetti FS, Ariztegui D, Brenner M, Curtis JH, Gilli A, Grzesik DA, Guilderson TJ, Muller AD, Bush MB, Correa Metrio YA, Escobar J, Kutterolf S (2008) An 85 ka Record of Climate Change in lowland Central America. Quaternary Sci Rev 27:1152 1165 Hodell DA, Brenner M, Curtis JH (2007) Climate and cultural history of the Northeastern Yucatan Peninsula, Quintana Roo, Mexico. Climatic Change 83:215 240 Hodell DA, Anselmetti F, Brenner M, Ariztegui D, PISDP Scientific Party (2006) The Lake PetŽn Itz‡ scientific drilling proje ct. Scientific Drilling 3:25 29 Hodell DA, Brenner M, Curtis JH, Medina Gonzalez R, Ildefonso Chan Can E, Albornaz Pat A, Guilderson TP (2005) Climate change on the Yucatan Peninsula during the Little Ice Age. Quaternary Res 63:109 121 Hodell DA, Curtis JH, Brenner M (1995) Possible Role of Climate in the Collapse of Classic Maya C ivilization. Nature 375:391 394 Hodell DA, Curtis JH, Jones GA, Higuera Gundy A, Brenner M, Binford MW, Dorsey KT (1991) Reconstruction of Caribbean climate change over the past 10,500 years. Nature 352:790 793 Holmes JA (2008) Sa mple size implications of the trace element variability of ostracod shells. Geoc him Cosmochim Acta 72:2934 2945 Holmes JA, Street Perrott FA, Allen MJ, Fothergill PA, Harkness DD, Kroon D, Perrott RA (1997). Holocene palaeolimnology of Kajemarum Oasis, Nor thern Nigeria: An isotopic study of ostracodes, bulk carbonate and organic carbon. J Geol Soc 154:311 319

PAGE 129

129 Hooghiemstra H, Cleef AM, Noldus GW, Kappelle M (1992) Upper Quaternary vegetation dynamics and palaeoclimatology of the La Chonta bog area, Costa R ic a. J Quaternary Sci 7:205 225 Hughen KA, Southon JR, Lehman SJ, Overpeck JT (2000) Synchronous radiocarbon and climate shifts during the last degl aciation. Science 290:1951 1954 Hughen KA, Overpeck JT, Peterson LC, Trumbore SE (1996) Rapid climate changes in the tropical Atlantic during the last deglaciation Nature 380:51 54 Imbrie J, Berger A, Boyle EA, Clemens SC, Duffy A, Howard WR, Kukla G, Kutzbach J, Martinson DG, McIntyre A, Mix AC, Molfino B, Morley JJ, Peterson LC, Pisias NG, Prell WL, Raymo ME, S hackleton NJ, Toggweiler JR (1993) On the structure and origin of major glaciation cycles: 2. The 100,000 year cy cle. Paleoceanography 8:699 735 Imbrie J, Boyle EA, Clemens SC, Duffy A, Howard WR, Kukla G, Kutzbach J, Martinson DG, McIntyre A, Mix AC, Molf ino B, Morley JJ, Peterson LC, Pisias NG, Prell WL, Raymo ME, Shackleton NJ, Toggweiler JR (1992) On the structure and origin of major glaciation cycles 1. Linear responses to Milankovitch forc ing. Paleoceanography 7:701 738 Islebe GA, Hooghiemstra H, Bren ner M, Curtis JH, Hodell DA (1996) A Holocene vegetation history from lowlan d Guatemala. Holocene 6:265 271 Islebe GA, Hooghiemstra H, van der Borg K (1995) A cooling event during the Younger Dryas chron in Costa Rica. Palaeogeogr Palaeocl Palaeoecol 117:7 3 80 Koutavas A, de Menocal PB, Olive GC, Lynch Stieglitz J (2006) Mid Holocene El Nino Southern Oscillation (ENSO) attenuation revealed by individual foraminifera in eastern tropical Pacifi c sediments. Geology 34:993 996 Jaeschke A, Ruehlemann C, Arz H, H eil G, Lohmann G (2007) Coupling of millennial scale changes in sea surface temperature and precipitation off northeastern Brazil with high latitude climate shifts during the last glacial period. Paleoceanography 22:PA4206 doi:10.1029/2006PA001391 Jansen E Veum T (1990) Evidence for two step deglaciation and its impact on North Atlantic deep water circulation. Nature 343:612 616 Jones MD, Leng MJ, Eastwood WJ, Keen DH, Turney CSM (2002) Interpreting stable isotope records from freshwater snail shell carbon ate: a Holocene case study from Lake G.olhisar, Turkey. The Holocene 12:629 634 Jouzel J, Vaikmae R, Petit JR, Martin M, Duclos Y, Stievenard M, Lorius C, Toots Melieres MA, Burckle LH, Barkov NI, Kotlyakov VM (1995) The two step shape and timing of the la st deglaciation in An tarctica. Clim Dynam 11:151 161

PAGE 130

130 Kiefer T, Kienast M (2005) Patterns of deglacal warming in the Pacific Ocean: a review with emphasis on the time interval of Heinrich event 1. Quaternary Sci Rev 24:1063 1081 Kienast M, Hanebuth TJJ, Pel ejero C, Steinke S (2003) Synchroneity of meltwater pulse 1a and the Blling warming: New evidence from the South China Sea. Geology 31:67 70 Kim SJ, Crowley TJ, Erickson DJ, Govindasamy B, Duffy PB, Lee BY (2008) High resolution climate simulation of the last glacial maximum. Clim Dynam 31:1 16 Koutavas A, Lynch Stieglitz J (2004) Variability of the marine ITCZ over the eastern Pacific during the past 30,000 years: regional perspective and global context. In: Diaz HF, Bradley RS (eds) The Hadley Circulatio n: Present, Past, and Future. Kluwer, Dordrecht, Germany, pp 347 369 Kucera M, Rosell Mele A, Schneider R, Waelbroeck C, Weinelt M (2005) Multiproxy approach for the reconstruction of the glacial ocean surface (MARGO). Quaternary Sci Rev 24:813 819 Kutzbac h JE, Guetter P (1986) The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18000 years. J Atmos Sci 43:1726 1738 Lachniet MS (2004) Late Quaternary glaciation of Costa Rica and Guatemala. In: Ehl ers J, Gibbard PL (eds) Quaternary Glaciations Extent and Chronology, Part III: South America, Asia, Africa, Australia, Antarctica. Elsevier, Amsterdam. Lachniet MS, Patterson WP (2009) Oxygen isotope values of precipitation and surface waters in northern Central America (Belize and Guatemala) are dominated by temperature and amount effects. Earth Planet Sc Lett doi:10.10 16/j.epsl.2009.05.010 Lachniet MS, Vasquez Selem L (2005) Last Glacial Maximum equilibrium line altitudes in the circum Caribbean (Mexico, Guatemala, Costa Rica, Colombia, and Venezuela). Quatern Int 138 139:129 144 Lea DW, Pak DK, Peterson LC, Hughen KA (2003) Synchroneity of tropical and high latitude Atlantic temperatures over the last glacial ter mination. Science 301:1361 1364 Leng MJ, L amb AL, Heaton THE, Marshall JD, Wolfe BB, Jones MD, Holmes JA, Arrowsmith C (2006) In: Leng MJ (ed) Isotopes in lake sediments. Isotopes in Paleoenv ironmental Research, pp 147 185 Leng MJ, Marshall, JD (2004) Paleoclimate interpretation of stable isotope date from lake sediment archives Quaternary Sci Rev 23:811 831

PAGE 131

131 Leng MJ, Roberts N, Reed JM, Sloane HJ (1999) Late quaternary climatic and limnological variations based on carbon and oxygen isotope data from authigenic and ostracods carbonate in the Konya basin, Turkey. Jo urnal of Paleolimnol 22:187 204 Leroux M (1993) The mobile polar high: a new concept explaining present mechanisms of meridional air mass and energy exchanges and global propagation of paleoclimatic changes. Global Planetary Change 7:69 93 Leyden BW (1995) Evidence of the Younger Dryas in Central America Quaternary Sci Rev 14:833 839 Leyden BW, Brenner M, Hodell DA, Curtis JH (1993) Late Pleistocene climate in the Central American lowlands. In: Swart PK, Lohmann KC, McKenzie J, Savin S (Ed s) Climate Change in Continental Isotopic Records. Geophysical Monograph 78. American Geophysical Un ion, Washington, DC, pp 165 178 Leyden BW (1984) Guatemalan forest synthesis after Pleistocene aridity. P Na tl Acad of Sci USA 81:4856 4859 McConnaughey TA Burdett J, Whelan JF, Paull CK (1997) Carbon isotopes in biological carbonates: Respiration and photosynthesis. Geoc him. Cosmochim. Acta 61:611 622 McGregor DL (1969) The reproductive potential, life history and parasitism of the freshwater ostracod Darw inula stevenson; (Brady and Robertson). In: Neale JW (ed) The taxonomy, morphology and ecology of recent ostracoda. Oliver and Boyd, Edinburgh, pp 194 221 McKenzie JA, Hollander DJ (1993) Oxygen isotope record in recent carbonate sediments from lake Greife n, Switzerland (1750 1986): application of continental isotopic indicator for evaluation of changes in climate and atmospheric circulation patterns. In: Swart PK, Lohmann KC, McKenzie J, Savin S (eds) Climate Change in Continental Isotopic Records, Geophy sical Monograph 78, American Geophysical Un ion, pp 101 111 McLay CL (1978) Comparative observations on the ecology of four species of ostracods living in a temporary freshwater puddle. Can J Zoolog 56:663 675 McManus JF, Francois R, Gherardi JM, Keigwin LD Brown Leger S (2004) Collapse and rapid resumption of Atlantic meridional circulation linked to deglacial clim ate changes. Nature 428:834 837 Meisch C (2000) Freshwater Ostracoda of western and central Europe. S٤wasserfauna von Mitteleuropa 8/3. Gustav Fischer, Stuttgart, 522 p Mix A, Bard E, Schneider R (2001) Environmental processes of the ice age: land, oceans, glaciers (EPILOG) Quaternary Sci Rev 20:627 657

PAGE 132

132 Monnin E, Indermuhle A, Dallenbach A, Fluckiger J, Stauffer B, Stocker TF, Raynaud D, Barnola JM (2001) Atmospheric CO2 concentrations over the last glacial termination. Science 291:112 114 Mueller AD, Anselmetti FS, Ariztegui D, Brenner M, Curtis JH, Escobar J, Gilli A, Grzesik DA, Hodell DA, Guilderson TJ, Kutterolf S, Ploetze ML (Accepted). Lat e Quaternary Palaeoenvironment of Lake PetŽn Itz‡, Northern Guatemala: Evidence from Deep Drill Holes and Seis mic Stratigraphy. Sedimentology Mueller AD, Islebe GA, Hillesheim MB, Grzeisk DA, Anselmetti, FS, Ariztegui D, Brenner M, Curtis JH, Hodell DA, Ve nz KA (2009). Climate drying and associated forest decline in the lowlands of northern Guatemala during the late Holocene. Quaternary Res 71:133 141 Nobre P, Shukla J (1996) Variations of sea surface temperature, wind stress and rainfall over the tropical Atlantic and South America. J Climate 9:2464 2479 North Greenland Ice Core Project members (2004) High resolution record of Northern Hemisphere climate extending into the last interglacial period. Nature 431:147 151 Okada Y (1982) Structure and cuticle for mation of the reticulated carapace of the ostracode Bycornucyther e bisanensis. Lethaia 15:85 101 Otto Bliesner BL, Brady EC (2009) The sensitivity of the climate response to the magnitude and location of freshwater forcing: last glacial maximum experiments Quaternary Sci Rev doi: 10.1016/j.quascirev.2009.07.004 Palacios Fest MR, Dettman DL (2001) Temperature controls monthly variation in Ostracode valve Mg/Ca: Cypridopsis vidua from a small lake in Sonora, Mexico. Geoc him Cosmochim Acta 65:2499 2507 Peltie r WR, Fairbanks RG (2006) Global glacial ice volume and Last Glacial Maximum duration from an extended Barbados sea level record. Quaternary Sci Rev 25:3322 3337 Peltier WR (1994) Ice age paleotopography. Science 265:195 201 Peterson LC, Haug GH (2006) Var iability in the mean latitude of the Atlantic Intertropical Convergence Zone as recorded by riverine input of sediments to the Cariaco Basin (Venezuela). Palaeogeogr Palae oclimatol Palaeoecol 234:97 113 Peterson LC, Haug GH, Hughen KA, Ršhl U (2000) Rapid changes in the hydrologic cycle of the tropical Atlantic during the Last Glacial. Science 290:1947 1951 Peypouquet JP, Carbonel P, Ducasse O, Tolderer Farmer M, Lete C (1988) Environmentally cued polymorphism of ostracods. In: Hanai T, Ikeya T, Ishizaki K (eds) Evolutionary Biology of Ostracoda. Its Fundamentals and App lications. Elsevier, Amsterdam.

PAGE 133

133 Piperno DR, Bush MB, Colinvaux PA (1990) Paleoenvironments and human occupation in late Glacial Pa nama. Quaternary Res 33:108 116 Reimer PJ, Baillie MGL, Bard E, Bayliss A, Beck JW, Bertrand CJH, Blackwell PG, Buck CE, Burr GS, Cutler KB, Damon PE, Edwards RL, Fairbanks RG, Friedrich M, Guilderson TP, Hogg AG, Hughen KA, Kromer B, McCormac G, Manning S, Ramsey CB, Reimer RW, Remmele S, Southon JR, Stuiver M, Tal amo S, Taylor FW, Van der Plicht J, Weyhenmeyer CE (2004) INTCAL04 terrestrial radiocarbon age calibration, 0 26 cal kyr BP. Radiocarbon 46:1029 1058 Ricketts RD, Johnson TC (1996) Early Holocene changes in lake level and productivity in lake Malawi as int erpreted from oxygen and carbon isotopic measurements from authigenic carbonates. In: Johnson TC, Odada EO (eds) The Limnology, Climatology and Palaeoclimatology of the East Africa lakes Gordon and Breach, Amsterdam. Rozanski K, Araguas Araguas L, Gonfian tini R (1993) Isotopic patterns in Modern Global Precipitation. Climate Change in Continental Isotopic Records, Geophysical Monograph 78. Ed. American Geophysical Union. Schiller A, Mikolajewicz U, Voss R (1997) The stability of the North Atlantic thermoha line circulation in a coupled ocean atmosphere general circulati on model. Clim Dynam 13:325 347 Schmidt MW, Spero HJ, Lea DW (2004) Links between salinity variation in the Caribbean and North Atlantic thermohaline circulation. Nature 428:160 163 Schrag DP, Adkins JF, McIntyre K, Alexander JL, Hodell DA, Charles CD, McManus JF (2002) The oxygen isotopic composition of seawater during the Last Glacial Maximum. Quaternary Sci Rev 21:331 342 Shanahan TM,Pigati JS, Dettman D, Quade J (2005) Isotopic variability in the aragonite shells of freshwater gastropods living in springs with nearly constant temperature and isotopic composition. Geochim Cosmochim Acta 69:3949 3966 Simstich J, Stanovoyb V, Baucha D, Erlenkeuserc H, Spielhagena RF (2004) Holocene variability of bottom water hydrography on the Kara Sea shelf (Siberia) depicted in multiple single valve analyses of stable isotopes in ostracods. Mar Geol 206:147 164 Stanford JD, Rohling EJ, Hunter SE, Roberts AP, Rasmussen SO, Bard E, McManus J, Fairbanks RG (200 6) Timing of meltwater pulse 1a and climate responses to meltwater injections. Paleoceanography 21:PA4103 doi:10.1029/2006PA001340. Talbot MR, Kelts K (1986) Primary and diagenetic carbonates in the anoxic sediments of lake Bosu mtwi, Ghana. Geology 14:912 916

PAGE 134

134 Timmermann A, Krebs U, Justino F, Goosse H, Ivanochko T (2005) Mechanisms for millennial scale global synchronization during the last glacial period. Paleoceanography 20:PA4008 doi:10.1029/2004PA001090 Turpen JB, Angell RW (1971) Aspects of molting and calcification in the ostracod Hete rocypris. Biol Bull 140:331 338 Vellinga M, Wood RA (2002) Global climatic impacts of a collapse of the Atlantic thermohaline circulation. Climatic Change 54:251 267 von Grafenstein U, Erlernkeuser H, Trimborn P (1999) Ox ygen and carbon isotopes in modern fresh water ostracod valves: assessing vital offsets and autecological effects of interest for palaeoclimate studies. Palaeogeogr Palaeo climatol Palaeoecol 148:133 152 Weaver AJ, Saenko OA, Clark PU, Mitrovica JX (2003) M eltwater pulse 1A from Antarctica as a trigger of the Blling Allerd warm interval. Science 299:1709 1713 Wilson EM, (1984) Physical geography of the Yucatan Peninsula. In: Moseley EH, Terry ED (Eds) Yucatan: A World Apart, second ed. University of Ala bam a Press, Tuscaloosa, pp 5 40 White SE (1986) Quaternary glacial stratigraphy and chronology of Mexico. Quaternary Sci Rev 5:201 205 Wunsch C (2003) Greenland Antarctic phase relations and millennial time scale climate fluctuations in the Greenland ice core s. Quaternary Sci Rev 22:1631 1646 Xia J, Engstrom DR Ito E (1997) Geochemistry of ostracode calcite: Part 2. The effect of water chemistry and seasonal temperature variation on Candona rawsoni. Geochim Cosmochim Acta 61:383 391 Yokoyama Y, Lambeck K, De Deckker P, Johnston P, Fifield LK (2000) Timing of the last glacial maximum from observed sea l evel minima. Nature 406:713 716 Zhang R, Delworth TH (2005) Simulated tropical response to a substantial weakening of the Atlantic thermohaline circulation. J C limate 18:1853 186

PAGE 135

135 BIOGRAPHICAL SKETCH Jaime Escobar was born in Medellin, Colombia (1977), where he spent his childhood and college years. He graduated from Colombo Britanico High School with an honor award in chemistry sciences. Jaime graduated in 2002 from La Escuela de Ingenieria de Antioquia where he received a BS in environmental engineering. As an environmental engineering student Jaime had the opportunity to work in the Institute for Coastal and Marine Research, INVEMAR, located in Santa Marta, C olombia. In this institute he was part of the Environmental Marine Quality Assessment Group. For two years Jaime was also part of the university news paper editing committee In August 2002 Jaime initiated his master's studies in the Department of Geologica l Sciences at the University of Florida thanks to a teaching assistantship offered by the Department. He obtained his master's degree in May 2005 and right away started his PhD studies at the School of Natural Resources and Environment (SNRE) at the Univer sity of Florida. Jaime obtained his PhD degree in interdisciplinary Ecology in May 2010. Jaime now lives in Panama with his wife Natalia Hoyos and two year old daughter, Antonia Escobar, where they work at t he Center for Tropical Paleoecology and Archaeolo gy (CTPA) at the Smithsonian Tropical Research Institute (STRI)