A Multi-Proxy Reconstruction of Paleolimnology and Paleoclimatic Variability Using Authigenic Lake Carbonates from Cleland Lake, British Columbia

Home , Mahoney Lake, Paleoecology, Paleolimnology

A MULTI-PROXY RECONSTRUCTION OF PALEOLIMNOLOGY AND PALEOCLIMATIC VARIABILITY USING AUTHIGENIC LAKE CARBONATES FROM CLELAND LAKE, BRITISH COLUMBIA

A dissertation submitted

to Kent State University in partial

fulfillment of the requirements for the

degree of Doctor of Philosophy

Lorita Nivanthi Mihindukulasooriya

December 2014

Dissertation written by

Lorita Nivanthi Mihindukulasooriya

B.Sc., University of Peradeniya, Sri Lanka, 2006

M.Sc., Ohio University, Athens, Ohio, 2009

Ph.D., Kent State University, 2014

Approved by

Dr. Joseph Ortiz, PhD, Department of Geology, Doctoral Advisor

Dr. Alison J. Smith, PhD, Department of Geology

Dr. Neil Wells, PhD, Department of Geology

Dr. Darren Bade, PhD, Department of Biology

Dr. Laura Leff, PhD, Graduate Representator, Department of Biology

Accepted by

Dr. Daniel Holm, PhD, Chair, Department of Geology

James L. Blank, PhD, Dean, College of Arts and Sciences

TABLE OF CONTENTS

LIST OF FIGURES ...... vii LIST OF TABLES ...... xi PREFACE...... xii ACKNOWLEDGEMENTS ...... xiii ABSTRACT ...... xvi CHAPTER 1 ...... 1 INTRODUCTION ...... 1 Research Questions and Objectives ...... 1 Study Area ...... 2 Geology of the Study Area...... 3 Lake Hydrologic Properties ...... 7 Research Hypothesis ...... 8 Holocene Climate of Southern British Columbia: A literature Review ...... 10 Late Glacial Climate of British Columbia ...... 12 Early Holocene Climate of British Columbia ...... 14 Mid Holocene Climate of British Columbia ...... 18 Late Holocene Climate of British Columbia ...... 20 CHAPTER 2 ...... 24 Reconstruction of Late Quaternary paleohydrologic conditions in southeastern British Columbia using visible derivative spectroscopy of Cleland Lake sediment ...... 24 (Manuscript in review in Quaternary Research) ...... 24 Abstract ...... 24 Introduction ...... 25 Study Area ...... 27 Methods ...... 28 Sediment Core Collection ...... 28 Geochronology ...... 29 Visible Derivative Spectroscopy ...... 30 X-ray Fluorescence ...... 31 Results ...... 32

iii

Lake Limnological Properties ...... 32 Age Model ...... 32 Interpretation of the Principal Components ...... 34 Discussion ...... 38 Evolution of the Lake Phytoplankton Community over the Holocene ...... 38 Paleohydrological Conditions Inferred from Phytoplankton Abundance ...... 41 Regional Comparison ...... 48 Possible Mechanism for the Late Quaternary Drought ...... 52 Conclusions ...... 54 CHAPTER 3 ...... 56 MODERN HYDROLOGY IN THE PACIFIC NORTWEST DURING THE MID-HOLOCENE ...... 56 Abstract ...... 56 Introduction ...... 57 Methods ...... 60 Geochronology ...... 60 Visible Derivative Spectroscopy ...... 61 Principal Component Analysis...... 61 Carbon and Oxygen Isotope Analysis ...... 62 X-ray Fluorescence ...... 63 Results ...... 63 Present Isotopic Properties of the Lake Water ...... 63 Age model ...... 64 Visible Derivative Spectroscopy and Principal Component Analysis ...... 66 Elemental Concentrations ...... 69 Discussion ...... 73 Conclusions ...... 79 CHAPTER 4 ...... 81 EFFECT OF MAZAMA TEPHRA ON LAKE SEDIMENTARY COMPOSITION AND PRIMARY PRODUCTIVITY ...... 81 Abstract ...... 81 Introduction ...... 82 Methods ...... 84

Preparation of a Common Age Model ...... 84 Results and Discussion ...... 90 Interpretation of the Principal Components ...... 90 Reconstruction of the Sedimentary Composition and the Lake Phytoplankton Composition ...... 92 Conclusions ...... 100 CHAPTER 5 ...... 102 HOLOCENE PALEOPRODUCTIVITY AND LAKE LEVEL CHANGES AT CLELAND LAKE ..... 102 Abstract ...... 102 Introduction ...... 103 Lake Sedimentary Structures ...... 106 Methods ...... 108 Sedimentological Analysis ...... 108 Loss on Ignition (LOI) ...... 109 Ostracode analysis ...... 110 Geochemical Analysis...... 111 Age Models ...... 112 Interpretation of Paleohydrologic Conditions ...... 112 Results ...... 114 Deep Basin Core (B-09) ...... 114 Intermediate-Depth Core (F-09) ...... 117 Shallow Core (E-09) ...... 123 Variation of Organic Matter and Calcium Carbonate with Water Depth ...... 125 Elemental Concentrations ...... 131 Discussion ...... 139 Reconstruction of Paleohydrological Conditions ...... 139 Wavelet Analysis ...... 144 Regional Comparison ...... 145 Conclusions ...... 147 CHAPTER 6 ...... 149 SYNTHESIS OF RESULTS AND FUTURE DIRECTIONS ...... 149 Introduction ...... 149 Reconstruction of Paleolake Lake Phytoplankton Abundance ...... 150

Paleohdrologic Conditions and Lake Level variability ...... 152 Forcing Mechanisms of Pluvial and Drought Variability and the Cyclicity (Chapters 2,3 and 5) ...... 154 REFERENCES ...... 157 APPENDICES ...... 172 APPENDIX 1 ...... 173 Reflectance data and results from principal component analysis of visible derivative spectroscopic data ...... 173 APPENDIX 2 ...... 177 Core Logs ...... 177 APPENDIX 3 ...... 185 Elemental concentrations from scanning XRF ...... 185

LIST OF FIGURES

Chapter 1

Figure 1.1 Relief map of Western Canada, showing major mountain systems and ranges...... 4

Figure 1.2 A. Map of British Columbia showing the location of Cleland Lake, Lake Bathymetry map of Cleland Lake, showing the location of core sites and topographical map of watershed in the Cleland Lake locale...... 5

Figure 1.3. Variation of temperature and precipitation in Brisco 10 km from the Cleland Lake. .. 6

Figure 1.4. Hydrological properties of Cleland Lake in July 2009 ...... 8

Figure 1.5. Map of BC showing the locations of previous palynological reconstructions (After Galloway, 2011)...... 11

Figure 1.6. Locations of midge paleotemperature reconstructions in British Columbia...... 12

Figure 1.7. Extent of the Cordilleran ice sheet during the last glacial period (Booth et al., 2004)...... 14

Figure 1.8. Midge inferred July summer temperature anomaly from four sites in Southern British Columbia...... 17

Figure 1.9. Mean annual precipitation (A) and mean July temperature (B) reconstructions based on pollen abundance from Marion Lake (Mathews and Heusser, 1981)...... 17

Chapter 2

Figure 2.1. Cleland lake location map A. Map of British Columbia showing the location of Cleland Lake and selected sites referenced in the text and limnological measurements of the Cleland Lake water column in July 2011 ...... 28

Figure 2.2. Age vs. depth model of Cleland Core F-09 (0 to 361.5cm) based on four radiocarbon dates and two tephra layers. Red dotted lines represent the 95 % confidence interval...... 33

Table 2.2. Total variance explained by the first four components from the Principal component analysis...... 53

Figure 2.3. Scatter plot of Dinoflagellate algae (DFA, PC 2) versus Diatoms and Cyanobacteria (PC 4) for Cores B-09, E-09 and F-09...... 47

Figure 2.4. Down core variation of PC 2, PC 4, Si:Ti ratio from core F-09 compared against summer insolation at 60°N and three independent paleoclimatic reconstructions from southern British Columbia...... 50

vii

Chapter 3

Figure 3.1. Oxygen and Hydrogen isotope ratios of surface water and precipitation samples from the closed basin lakes of Southern Yukon Territory (after Anderson et al., 2005) ...... 59

Figure 3.2. Oxygen and Hydrogen isotope ratios of surface water and precipitation samples from Western Canada...... 64

Figure 3.3. Age model of the Cleland C-09 and B-09 cores...... 66

Figure 3.4. Scatter plot between PC 1 versus PC 4...... 68

Figure 3.5.Plot of 18O verseus 13C from the Cleland Lake sediments ...... 69

Figure 3.6. Down core variation of reflectance PC 1 (illite+sphalerite), reflectance PC 4:diatom+ cyanobacteria pigments (cyanobacteria), δ18O of carbonates and Holocene glacial advances in Western Canada (after Menouns et al., 2009) ...... 71

Figure 3.7. Variation of major and trace metal concentrations between 7500 to 400 cal yr BP, with the generalized stratigraphic column of the deep core (B-09)...... 72

Figure 3.8. Cleland Lake watershed showing the flow lines ...... 76

Chapter 4

Figure 4.1 Magnetic susceptibility of the shallow (E-09) and the intermediate (F-09) cores...... 86

Figure 4.2. Reflectance PC 2 of the shallow (E-09) and the intermediate (F-09) cores. Top - plotted on their original depth scales, bottom PC-2 of the E-09 core plotted on the depth scale of the intermediate core...... 87

Figure 4.3. Reflectance PC 2 of the deep (B-09) and the intermediate (F-09) cores. plotted on their original depth scales and plotted on the depth scale of the intermediate core...... 104

Figure 4.4. First five principal components of reflectance data from the three cores of Cleland Lake plotted on a common depth scale...... 90

Figure 4.5. All available age data from Cleland Lake plotted on the depth scale of the intermediate core ...... 109

Figure 4.6. Elemental concentrations of the Mazama tephra layer (209 cm to 213 cm) in the intermediate core...... 94

Figure 4.7. First five principal components of reflectance data on the age scale ...... 96

viii

Chapter 5

Figure 5.1. Percent organic carbon, percent carbonate, percent mineral matter, and dry bulk density in gcm-3 compared with the generalized stratigraphic column of the Cleland Lake deep core...... 115

Figure 5.2. High resolution image of the deep basin core B-09...... 132

Figure 5.3. Limnocythere sp. cf friabilis female right valve...... 133

Figure 5.4. Percent organic carbon, percent carbonate, percent mineral matter (% Mineral), and dry bulk density, compared with the generalized stratigraphic column of the Cleland Lake intermediate core...... 120

Figure 5.5. High resolution images of the core F-09...... 137

Figure 5.6 (a) Detrended percent organic matter content in the intermediate core during the past 8000 years (b) The wavelet power spectrum for organic matter percentages interpolated to 100 year intervals...... 122

Figure .5.7 (a) Detrended percent calcium carbonate content in the intermediate core during the past 8000 years. (b) The wavelet power spectrum for calcium carbonate percentages interpolated to 100 year intervals...... 123

Figure 5.8. Percent organic carbon, percent carbonate, percent mineral matter, and dry bulk density compared with the generalized stratigraphic column of the Cleland Lake shallow core, E-09...... 126

Figure 5.9. High Resolution Images of the shallow water core E-09...... 143

Figure 5.10. Down core variation in the percentage of organic matter and the percentage of CaCO3, in the intermediate (F-09), shallow core (E-09), and the deep core (B-09)...... 144

Figure 5.11. Scatter plot of LOI derived carbonate and organic matter percentage for cores B-09, E-09 and F-09...... 129

Figure 5.12. The generalized stratigraphic sections of the shallow, intermediate and deep cores of the Cleland lake...... 130

Figure 5.13. Down core variation of percent mineral matter, concentrations of and Si, K and Ti, Fe and Ni (cps), as well as XRF-PC1 in core F-09...... 135

Figure 5.14. Down core variation of reflectance PC 2, XRF-PC 3, as well as Sr and Ca concentrations (cps), percent CaCO3 and ostracode counts in the core F-09...... 136

Figure 5.15. Down core variation of Si, K and Ti and Fe concentrations (cps), XRF-PC 1, magnetic susceptibly (MS), and percent mineral matter (mineral m) percentage in core E- 09...... 137

Figure 5.16. Down core variation of percent CaCO3, Sr and Ca concentrations (cps), XRF-PC 3, reflectance PC 2 (Ref PC 2), and ostracode counts in the core E-09...... 138

Chapter 6

Figure 6.1. Schematic diagram showing the locations of the North Pacific high pressure system and the low pressure system and the direction of moist air masses ...... 156

LIST OF TABLES

Table 1.1. Summary of previous paleoclimatic reconstructions from British Columbia (BC).....29

Table 2.1. AMS Radiocarbon and tephra dates for Cleland Lake Core F-09 ...... 50

Table 2.2. Total variance explained by the first four components from the Principal component analysis...... 52 Table 2.3 Average values of PC 2 and PC 4 for all three sediment cores...... 64 Table 3.1. AMS Radiocarbon and tephra dates for the Cleland lake B-09 core…………………67 Table 4.1 Age control points used for calibration of the depths of the shallow core to the depth scale of the intermediate core...... 88

Table 5.1. Average values of percent carbonate and organic matter for all three sediment cores...... 145 Table 5.2. Initial eigen values and variance explained by the first five principal components of the XRF data...... 147

Table 5.3. Rotated Component Matrix for the combined XRF data matrix...... 148

PREFACE

This research is a collaboration between University of Pittsburgh and Kent State University.

Contributions from co-authors:

David Pompiani- University of Pittsburgh - involved in colleting lake sediment samples, performed laboratory analysis for Loss of Ignition, C and Oxygen stable isotope analysis and age dating, assisted in preparing diagrams and proof reading

Byron Steinman- University of Minnersota Dulth- involved in colleting lake sediment samples, assisted in preparing some of the text and proof reading

Mark Abbott - University of Pittsburgh, Pittsburgh, Co Principal investigator of the project, involved in colleting lake sediment samples, assisted in preparing some of the text and proof reading

Ortiz Joseph- Kent State University - Dissertation Director, Co Principal investigator of the project, involved in colleting lake sediment samples

xii

ACKNOWLEDGEMENTS

I wish to acknowledge my advisor Dr. Ortiz Joseph for his constant guidance throughout this project. I am indebted to him for developing my passion for paleoceanography with his never changing patience and dedication throughout the course of my project. Secondly, I wish to thank Dr. Alison Smith, Dr. Neil Wells, Dr. Darren Bade and Dr. Laura Leff for kindly serving in my dissertation advising committee and the guidance and valuable ideas that made this dissertation possible. Many thanks to Dr. Bade and Dr. Smith for allowing me to audit their classes, which helped prepare me for this dissertation. I am especially thankful to Dr.Smith and

Dr. Palmer for their invaluable guidance and helping me to understand that lakes are not just puddles of water. I am also thankful for my wonderful team of collaborators from University of

Pittsburgh; Dr. Mark Abbott, Dr. Byron Steinman and David Pompiani for their guidance and continuous support to make this research project a success. David Pompiani helped me in every stage of this project by providing data, samples as well as preparing some of the figures and proof reading a manuscript. Thank you Dave for all that support.

I also wish to acknowledge the faculty and staff of the Department of Geology for providing me with necessary facilities and guidance to successfully complete my doctoral program. I am grateful to my colleagues Nick Bonini and Samantha Yost for reading my dissertation and providing feedback to improve its quality. Of course I am eternally grateful to my parents for guiding me through the journey of my life. Their intellectual and financial guidance in every aspect of my life is respectfully appreciated. Lastly I would like to thank my husband and best friend Nilantha, for his encouragement and support throughout my work.

Without him this work would never have become a reality. Thank you for all the dedication you

xiii made to pursue my passion. Also a big thanks goes to Tharusha for tolerating your mom's long busy graduate student life.

This research was funded by the National Science Foundation and small grants from the

Geological Society of America and Graduate Students Senate of the Kent State University.

Thank you all.

xiv

DEDICATION

I dedicate this dissertation to my family, especially… to my loving parents for instilling the importance of hard work

and higher education

Nilantha; my best friend and the strength in life

and Tharusha for his patience

MIHINDUKULASOORIYA LORITA N., PhD., DECEMBER 2014 APPLIED GEOLOGY

A MULTI-PROXY RECONSTRUCTION OF PALEOLAKE PRODUCTIVITY AND PALEOCLIMATIC VARIABILITY USING AUTHIGENIC LAKE CARBONATES FROM CLELAND LAKE BRITISH COLUMBIA

DIRECTOR OF DISSERTATION: Dr. JOSEPH D. ORTIZ

ABSTRACT

In small closed-basin lakes in semi-arid regions, variations in precipitation/evaporation

(P/E) balance affect the physical, biological, and chemical composition of the lake water and sediment. This study presents visible derivative spectroscopy (VDS), XRF-derived elemental

18 18 concentrations, sedimentological compositions, and δ O values of carbonates (δ Ocarb) in sediment cores from Cleland Lake, British Columbia. The data provides insight into paleolimnological variations during the past 14000 years, based on three sediment cores from shallow, intermediate and deep parts of the lake, and can be used as an analogue to understand future climatic changes and associated changes to aquatic ecosystems. VDS-based principal components (PCs) derived two PCs that correlate with illite/sphalerite and smectite/chlorite that indicate paleohydrology and lake sedimentary composition, respectively. Four of the PCs correlate with the standard reflectance curves of dinoflagellate algae, diatoms and cyanobacteria, and were used as paleoproductivity proxy. Lake productivity is limited to early successional cyanobacteria during deglaciation (14000 to 11500 cal yr BP). Rapid rise in dinoflagellates and a second cyanobacteria community suggest lower lake levels from 11600 to 8600 cal yr BP.

Laminated to massive carbonate mud, rich in Ca, Sr, and calcium carbonate (LOI derived) provide additional evidence for a shallow lake. Conditions promoting cyanobacteria shifted

xvi toward those favoring diatoms around 9400 and lasted until 170 cal yr BP. A climate transition to wet conditions is suggested by rapid drops in percent CaCO3 and Ca levels and facies changes to well-laminated, organic-matter-rich sediment between 8700 and 8600 cal yr BP. A rapid decline in diatoms, cyanobacteria and dinoflagellate algae is recorded from 8500 to 7700 cal yr

BP, simultaneous with the global-cooling event centered on 8200 cal yr BP. Lake sediment and phytoplankton abundance reflect rapid changes after the deposition of Mazama tephra around

7700 cal yr BP. Late Holocene levels of diatoms, carbonates, and illite suggest frequent changes in moisture levels at century scale, possibly associated with ocean-atmosphere tele-connection to the changes in solar irradiance. In contrast, early-Holocene lake level changes are associated with long-term climate forcing, like solar insolation, and reveal a much longer periodicity of around 5000 years. This research points out that natural forcing mechanisms affecting the sea surface temperature of the Pacific Ocean, similar to the solar insolation, solar irradiance and glacial melt water fluxes can affect the regional climate and the lake productivity by affecting the moisture transport to the region.

xvii

CHAPTER 1

INTRODUCTION

Research Questions and Objectives

Severe droughts have affected the northwestern United States since instrumental records started in the 17th century, as well as in the recent past. More intense droughts of longer duration were recorded during the period from 900 to 1300 AD, consistent with the Medieval Warm

Period (Cook et al., 2004). The most recent drought from 1999 to 2004, which affected most parts of the western U.S. and Canada, is considered to be the worst drought since the 1930’s

“dustbowl” (Cook et al., 2004). The northwestern region of the interior Pacific Northwest had also been identified as a core region for drought where droughts tend to originate, expand and persist even during years of regional aridity contraction (Knapp et al., 2004). Therefore, it is important to study the drought frequency in the Pacific Northwest because the vastly growing population over the area forces overexploitation of the limited water resources and can worsen the severity of rainfall deficits. Studying of such past events will further allow us to predict future droughts, as well as floods. Such information is also important for water resource management, agricultural and urban planning authorities to better prepare for future droughts.

This research will attempt to understand the frequency, distribution and the driving forces of the pluvial and drought variability in the western United States during the Holocene.

Carbonate precipitation in freshwater lakes can take place when the lake water becomes supersaturated with respect to calcium carbonate, either through biological or physicochemical processes (Nelson et al., 2009). Hydrology of the lake (input and output of surface water), rainfall and groundwater, sediment input, bedrock of the watershed, and temperature are the major factors governing the lacustrine carbonate precipitation (Tucker and Wright, 1990; Platt and Wright 1991; Gierlowski-Kordesh, 2010). In general, climate is a primary factor governing carbonate precipitation (Gierlowski-Kordesh, 2010). Temperature and precipitation determine the biological productivity of the lakes and therefore affect the biological carbonate deposition rates (Platt and Wright 1991). Authigenic carbonates are widely used in paleoclimatic studies assuming the carbonates precipitate in isotopic equilibrium (Hammarlund et al., 2003). Higher authigenic carbonate precipitation is observed in summer months due to phytoplankton productivity, in most temperate and high latitude lakes. The major advantage of using authigenic carbonates is they can generate high resolution paleoclimatic records up to decadal to annual scale (Leng and Marshall, 2004).

Study Area

Cleland Lake (50.82° N, 116.38° W, 1158 m asl) is located in the north-south trending

Columbia valley just west of the continental divide in the Rocky Mountains of southeastern

British Columbia (Figures 1.1, 1.2). The lake, which extends over 0.235 km2 (23.5 ha) and is covered with ice from November to May, has a main basin that is 31.7 m deep. Meteorological data from the Brisco weather station 10 km east of the lake document a mean annual temperature of ~11.24 °C over the period from 1984 to 2003. The January mean temperature is -3.7 °C and the July mean temperature is ~24.8 °C (Figure 1.3).from November to May. The lake is covered

with ice The mean annual precipitation during the period from 1924 to 2003 is 426 mm. The surface inflow comes from its 2 km2 watershed, while water loss is restricted to evaporation and groundwater seepage. According to the Köppen–Geiger climate classification system, the area has Highland climates (H; Strahler, 1959).

Geology of the Study Area

The watershed of Cleland Lake consist of Proterozoic clastic aregillaceous and calcareous sedimentary rocks, which belong to the Purcell Supergroup and Mount Nelson

Formation (geological Map of British Columbia, 1962). The bedrock is covered with glacial till.

The major rock types include quartz arenite, limestone, and quartzite. The south end of the lake consists of a magnesite ridge composed of coarse grained sparry magnesite deposit (Simandl and

Hancock, 2002). This magensite deposit consist of 60 to 95% magnesite, 3 to 40% sparry dolomite, <5% silica, and <0.5% pyrite (Simandl and Hancock, 2002). Approximately 620 m southeast of the magnesite deposit are metaliferous horizons consisting of sphalerite and bornite

(Simandl and Hancock, 2002).

Figure 1.1 Relief map of Western Canada, showing major mountain systems and ranges. Blue shading shows present day ice cover. Red square shows the location of Cleland Lake (modified from Menounos et al., 2009).

A B

Figure 1.2 A. Map of British Columbia showing the location of Cleland Lake (red dot) and selected sites referenced in the text (black triangles), 1: Eleanor Lake, 2: Windy Lake, 3: Frozen Lake, 4: North Crater Lake, 5: Lake of the Woods. B. Bathymetry map of Cleland Lake (lake depth in meters), showing locations of core sites. C. Shaded relief map of watershed in the Cleland Lake locale.

60 20

50 15

40 10

C 

30 5 Precipitation(mm) 20 0 Temperature

10 -5

0 -10 Jan Feb March April May June July Aug Sept Oct Nov Dec Month Mean monthly Rainfall (mm) Mean monthly snowfall (mm) Mean monthly Temp (°C)

Figure 1.3. Top. Annual cycle of temperature and precipitation in Brisco, 10 km from Cleland Lake. Mean monthly snow fall and rain fall is the average from 1924 to 2003 and the mean monthly temperature is the average from 1984 to 2003. Bottom. Monthly minimum and maximum precipitation between 1924 and 2003. (Data source http://climate.weather.gc.ca/)

Lake Hydrologic Properties

The lake water is alkaline with an average dissolved bicarbonate and carbonate concentrations of ~530 mg/l and an average pH of 9.6. This chemistry allows summer time precipitation of micritic calcite during whiting events (Figure 1.4). Total dissolved solids (TDS) concentration vary between 510 mg/l and 550 mg/l from to surface to the lake bottom. The salinity of water varies from 0.42 ppt at the surface to 0.44 ppt (420 mg/l and 440 mg/l) at the bottom (Appendix 1).

Temperature

The temperature profile of Cleland Lake demonstrates that the lake is strongly stratified in the summer. The average temperature of the surface layer (epiliminion) is 19.7°C. The temperature of the transition zone (metalimnion, 5 to 12 m) drops rapidly from 19.18 to 5.84 ºC.

The water below 12 m (hypolimnion) has very low temperatures, with an average of 4.2 ºC

(Figure 1.4).

Distribution of oxygen

Figure 1.4 shows the vertical distribution of dissolved oxygen (DO) in summer. The surface oxygen concentrations are relatively high (average 9.2 mg/l) within the top 5 m.

Concentration rapidly drops from 9.61 to 0.42 mg/l within the metalimnion (5 to 15 m). Lake water is anoxic below 12 m. DO concentrations that are relatively high at the surface and lowest in the hypolimnion, constitiute a characteristic clinograde oxygen profile (Kalf 2009). The epilimnion is nearly saturated with DO due to mixing with atmospheric oxygen and photosynthetic oxygen. Increase in organic matter input from the epilimnion and the resulting

higher respiration rates and organic matter decomposition gradually lower the hypolimnetic DO concentrations.

T (°C) DO (mg/L) SC (ms/cm) pH 0

15 Depth (m)

25 10 20 5 10 1.1 1.2 9.6 9.8

Figure 1.4. Hydrological properties of Cleland Lake in July 2009, T; Temperature in Celsius, DO; dissolved oxygen (mg/l), SC; specific conductivity (ms/cm), and pH.

Research Hypothesis

H1.a -Visible derivative spectroscopy can be used as a proxy for lake phytoplankton abundance

H1.b. Phytoplanktom abundance at Cleland Lake should represent regional paleoclimatic changes

Literature study shows that DSR can be used to obtain information about clay minerals, iron oxides and hydroxides, and plant pigments (Ortiz et al., 2009, Wolfe et al., 2006; Ji et al.,

2005). Fossil pigments are an important constituent of the organic matter preserved in lake

sediment and can be used as a proxy for primary productivity within the lake and changes in phytoplankton community structure (Peteet et al., 2003; Das et al., 2005; Wolfe et al., 2006;

Bouchard et al., 2013). Concentrations of sedimentary photosynthetic pigments have long been used as a direct and accurate method to reconstruct the response of lake phytoplankton to environmental changes (Sanger and Gorham 1972; Sanger and Crowl 1979; Sanger 1888; Nara et al., 2005). Phytoplankton composition in temperate lakes tends to follow a succession pattern similar to that of plants on the landscape and is strongly influenced by water chemistry and lake level. Biological communities in drought-sensitive lakes are therefore a reflection of water balance changes, such that algal pigment analysis of sediment from these lakes should reflect past hydroclimate variability.

H2- Lake level changes depicted from multiple proxies should correlate with existing paleoclimatic reconstructions of the Pacific Northwest

Annually laminated lake sediments are excellent archives of paleoclimatic conditions.

Lake carbonates should reflect the changes in precipitation to evaporation ratio of the lake watershed and hence would indicate pluvial and drought conditions during the Holocene. Rate of carbonate precipitation, C and O isotope ratios and magnetic susceptibility of authigenic carbonates will be used to reconstruct the paleo-lake levels. Accuracy of the reconstructed lake levels and the depicted precipitation to evaporation ratios can be tested by comparing the reconstructed precipitation to evaporation ratios with the available paleoclimatic records.

H3- Long-term variations of lake levels reconstructed from multiple proxies should represent variations of Holocene ocean-atmosphere circulations.

Decadal variability of the climate in the Pacific Northwest is mainly a result of variations in sea surface temperature (SST), salinities and ocean topographies of the Pacific and Atlantic

Ocean (McCabe et al., 2004). Literature suggests the multi-decadal drought frequency over the conterminous United States is attributable to Pacific Decadal Oscillation (PDO), Atlantic Multi- decadal Oscillation (AMO), and Northern hemisphere temperature (McCabe et al., 2004).

However, the physical mechanism for prolonged droughts remain poorly defined. Therefore, the reconstructed Holocene lake levels will be correlated against PDO, AMO and Pacific and

Atlantic Ocean SST data in order to understand the major driving forces of the climatic variability of the Western Pacific.

Holocene Climate of Southern British Columbia: A literature Review

Eastern British Columbia is a transitional region in between climatic regimes and air mass dominance where winter onshore flow collides with polar air masses (Gavin et al., 2011).

Such regions are considered to be suitable for studying Holocene abrupt climatic changes (Gavin et al., 2011). Most of the paleoclimatic reconstructions in British Columbia are based on pollen analysis. The area's many glacial lakes provide excellent site for preservation of pollen suitable for paleo-ecological and paleoenvironmental analysis (Figure 1.5). Among all these reconstructions, Dog Lake pollen record is the spatially closest lake to the Cleland Lake. The most recent pollen and diatom based paleoclimatic reconstruction comes from Felker Lake,

British Columbia, located 400 km northwest of Cleland Lake (Figure 1.5).

Despite the presence of abundant lacustrine deposits, temporal resolution available in pollen analysis is lower than the high resolution lacustrine isotope and DSR methods. Plant pollen abundances heavily depend on precipitation and temperature. But, other environmental factors like slope, aspect, and soil texture are also important (Walker and Pellatt, 2003).

However, extensive lags can exist between climatic changes and vegetation response as vegetation respond slowly(Walker and Pellatt, 2003). As a result, amplitude and timing of individual rapid climatic changes recorded in paleoclimatic proxies in response to the same forcing mechanism may differ over short distances. This is mainly due to the topographical changes between coastal and interior, as well as lowlands and mountains (Mathews and Heusser,

1981).

Figure 1.5. Map of BC showing the locations of previous palynological reconstructions, with Cleland Lake shown for reference (red dot). 1. Marion and Surprise Lake, 2. Horseshoe Lake, 3. Phair Lake, 4. Chihil Lake, 5. Pinecreset and Squeah Lake, 7. Big Lake, 8. Red Mountain Lake, 9. Pemberton Lake, 10. Mahoney Lake, 11. Kipoola Lake, 12. Thunder Lake, 13. Windy Lake, 14. Dog Lake, 15. Felker Lake. (After Galloway, 2011).

During the past decade, aquatic midges (chironomidae) have emerged as a leading paleotemperature proxy (Walker and Pellatt, 2003, Figure 1.6). As the life cycle of these insects is shorter, they respond rapidly to climatic changes compared to trees, making midge-based reconstructions favored by aquatic biologists (as opposed to pollen based reconstructions).

Midge reproduction depends on summer air temperature and, therefore, midge abundance is a reflection of summer temperature (Chase et al., 2008). The midge life cycle, however, is also dependent on other environmental factors, such as pH, salinity, competition and predation

(Chase et al., 2008). Resolutions of midge reconstructions are at century or longer scales and limited by the sedimentation rate and bioturbation by midges and other fauna (Walker and

Pellatt, 2003).

Figure 1.6. Locations of midge based paleotemperature reconstructions in British Columbia 1. Hippa Lake, 2. Misty Lake, 3. Marion and Mike Lake, 4. Frozen Lake, 5. Cabin Lake and 3M Pond, 6. Crater Lake and Lake of the Woods, 7. Eagle Lake, 8. Windy Lake, 9. Thunder Lake, 10. Redmountan Lake. (Modified from Chase et al., 2008)

Late Glacial Climate of British Columbia

During the Last Glacial Maximum (LGM) the area was extensively covered by the

Okanagon lobe of the Cordilleran ice sheet (Figure 1.7). The ice sheet was constrained by the topography in the area: the Coast and Elias Mountains to the west and the Rocky and Mackenzie

Mountains and the much larger Laurentide ice sheet to the east. The high mountain ranges of

British Columbia and southern Yukon were the major accommodation regions for the

Cordilleran ice sheet. The easternmost lobe of the ice sheet—the Pend Orielle lobe—reached its maximum southern limits prior to 12500 14C years B.P (14,700 cal yr BP; calibrated from Calib

7.02; Clague and James, 2002). The ice sheet showed a rapid growth between 18000 and 14000

14C years B.P., (21800 to 17000 cal yr BP). However, the extent of the Cordilleran ice sheet was much more limited at 18000 14C years BP (21,800 cal yr BP) than the Laurentide ice sheet

(Booth et al., 2004). The ice reached its maximum extent by 14000 14C years BP (17000 cal yr

BP), covering the entire southern British Columbia and Seattle area with a thickness of about 2 km (Booth et al., 2004). This lagged 4000 14C years behind the glacial maximum in other parts of the northern hemesphere (Clague and James, 2001; Booth et al., 2004). The actual reason for this lag is not yet understood, indicating the complexity of the climate system in the Pacific

Northwest. Two possibilities are (1) the decay of the Laurentide ice sheet might have provided moisture for the growth of Cordillarian ice sheet, and/or (2) changes to the atmospheric circulation pattern in the western North America and the northern Pacific Ocean (Clague and

James, 2001).

The ice remained at its maximum position for several centuries before it retreated north of Seattle by 13600 14C yrs BP (16575 cal yr BP). The rapid retreat of the ice sheet was facilitated by calving into proglacial lakes and the sea (Clague and James, 2001). The glacial retreat was interrupted by several minor advances of the ice front collectively known as “Sumas advance”, between 13000 and 10500 14C yrs B.P.(15000 and 12000 cal yr BP) (Clague and

James, 2001; Menounos et al., 2009). The exact chronology of the glacial retreat in interior

British Columbia is poorly known, but the extent of the glaciers were perhaps limited to the present Alpine ranges by 9500 14C yrs BP (10500 cal yr BP).

Figure 1.7. Extent of the Cordilleran ice sheet during the last glacial period (Booth et al., 2003). The red square shows the location of the study area.

Early Holocene Climate of British Columbia

The Late Glacial-early Holocene have been identified as the interval experecincing the most rapid and significant floral and climatic fluctuations in southern British Columbia over the past 14000 cal yrs B.P. (Pellatt et al., 2002). The deglaciation, Sumas ice advance, Younger

Dryas cooling, and rapid early Holocene warming all took place during this 2000 year period

(Pellatt et al., 2002).

Paleotemperature and precipitation reconstructions based on pollen abundance from

Marion Lake (700 km west of Cleland Lake; Figure 1.5) show lowest postglacial July temperature and moderately high (1900±620 mm) precipitation between 14000 and 12000 cal yr

BP (12000 and 10500 14C years B.P., Figure 1.9.A and B: Mathews and Heusser, 1981). Interior

British Columbia was abundantly covered by Pinus parkland during the post-glacial period when the prevailing cold and dry climatic conditions and poor soil conditions limited the growth of other taxa (Galloway, 2011; Hallet and Hills, 2006). Pollen reconstruction for deglaciation and early Holocene and around southwestern British Columbia shows elevated pollen levels for western hemlock and spruce during the interval from 11400 to 11000 cal yr BP. which suggests high precipitation but a warmer climate (Pellatt et al., 2002).

Fossil midge-based paleotemperature reconstructions from Windy Lake support a rapid transition to a warm early Holocene from a 1-2 °C colder than present deglacial period in southern British Columbia (Figure 1.8; Chase et al., 2008). Precipitation rapidly dropped to 1450

±620 mm and temperature rapidly increased by to 2°C ±1.1 °C between 12100 to 11550 cal years B.P (10500 B.P and 10000 14C years B.P.; Mathews and Heusser, 1981). The temperature remained high while precipitation remained low between 11550 to 8400 cal yr BP (10000 and

7500 years 14C years B.P.) during this dry period, commonly known as the “xerothermic interval” (Figure 1.8, Chase et al., 2008; Figure 1.9.A,B, Mathews and Heusser, 1981).

Declining western hemlock and spruce was replaced by douglas-fir and red alder after

11000 years BP, further supporting the inference of a warm and dry early-Holocene (Chase et al.,

2008). The earliest portion of the pollen record from a peat bog in the Okanagan valley of

Southern British Columbia placed the transition from the wet deglacial period to an arid early-

Holocene after 9300 cal yr BP (8400 14C years BP.; Alley, 1976). An increase in bracken

(Pteridium) spores (Pellatt et al., 2002), and very low abundance of western hemlock (Tsuga heterophylla), western red cedar (Thuja plicata), true fir (Abies) and spruce is characteristic of a warm, dry climate frequently associated with fire (Walker and Pellatt, 2004). Elevated percentages of xerohpyte plant taxa, such as Pinus juniperus, Artemisia, and Poaceae also provide paleobotanical evidences for a drier climate prior to 8500 cal yr BP.

Paleobotanical and paleoenvironmental reconstruction from Dog Lake, British Columbia, located 30 km east of Cleland Lake, also supports an early Holocene dry climate characterized by a long-term decrease in lake levels until 8500 cal. yrs B.P. (Hallet and Hills, 2006 and references therein). Diatom-based lake level reconstruction from Felker Lake, British Columbia, identified the period from 8,140 to 6,840 cal yrs BP as a period with low lake levels, in response to warm summer air temperatures leading to a well-stratified, nutrient-limited epilimnion. The presence of tree stumps at elevations higher than present suggests tree line in British Columbia was higher than the present limit during early Holocene (Chase et al., 2008 and references therein). Evidence for a warm early Holocene was recorded across the Northern Hemisphere. For instance, increased aeolian dune activity and prolonged periods of aridity has been recorded from the central Great Plains of North America (Hallet and Hills, 2006 and references there in).

3.5

3 C) ° 2.5 2 1.5 1 0.5 Anomaly (750-yr 0 smoothed) Anomaly (300-yr Temperature Anomaly ( Anomaly Temperature -0.5 smoothed) -1 -100 1900 3900 5900 7900 9900 11900 Age (cal yr BP)

Figure 1.8. Midge inferred July summer temperature anomaly from four sites in Southern British Columbia; Frozen Lake (Rosenberg et al., 2004), North Crater Lake and Lake-of-the-Woods (Palmer et al., 2002), and Windy Lake (Chase et al., 2008) (re-drawn using the data from Gavin et al., 2011).

Figure 1.9. Mean annual precipitation (A) and mean July temperature (B) reconstructions based on pollen abundance from Marion Lake from Mathews and Heusser, 1980.

Although the evidences for a warm dry early Holocene have been recorded across southern British Columbia, evidence for a xerothermic episode (warm and dry climate) is sparse and needs further study in northern British Columbia (Walker and Pellatt, 2004). The most likely mechanism linked to this climatic change is the increased northern hemisphere summer insolation at this time. The northern hemisphere summer insolation was at its maximum and was

8% higher than the present insolation by 13000 cal yr BP (Galloway et al., 2011). However, the maximum early Holocene temperature in Western Canada was recorded between 11500 and

8400 cal ye BP (Mathews and Heusser, 1981). Continued presence of Laurentide ice sheet caused the maximum temperature to lag behind the insolation peak by several thousand years in the region (Galloway et al., 2011; Berger and Loutre 1991).

Mid Holocene Climate of British Columbia

The Mid-Holocene is identified as a period of gradual transition from a very dry warm early Holocene to a cool moist mid Holocene (Lowe et al., 1993, Galloway et al., 2011); however, the timing of the cooling is reported as a time-transgressive event across British

Columbia (Walker and Pellatt 2003). Highly variable climatic conditions with higher effective moisture as well as low effective moisture have been reported across British Columbia during the mid-Holocene (Hallet and Hills, 2006). Moist conditions were recorded from coastal sites, while arid conditions were recorded from interior British Columbia sites located closer to the Rockies and are less influenced by the moist Pacific air masses (Hallet and Hills, 2006).

According to pollen-based reconstructions from coastal British Columbia, between 7500

14C years BP (8400 cal yr BP) and 6000 14C years BP (6800 cal yr BP) temperature gradually dropped, and the precipitation increased from low early Holocene values to higher modern values (Mathews and Heusser, 1981). Temperatures fluctuated around 14.5°C between 6000 and

4000 14C years BP (6900 to 4525 cal yr BP; Mathews and Heusser, 1981). Fossil midge-based reconstructions from the Windy Lake, southern British Columbia, suggest summer temperatures remained 1-2°C higher than the present temperatures between 8000 to 6500 cal yr BP, while temperatures close to the present temperatures have been recorded by 6000 years B.P (Chase et al., 2008). Continuous summer temperature decrease associated with the summer insolation decrease has been reported throughout the western United States until 4000 cal yr BP (Chase et al., 2008 and references there in). Multi-proxy evidence from Big Lake, British Columbia, suggests higher effective moisture conditions from 8500 to 6700 cal yr BP (Hallet and Hills,

2006) with the highest lake levels between 7600 and 6700 cal yr BP. The deposition of laminated sediment in Mahoney Lake supports an interpretation of wet climatic conditions between 6300 and 5650 cal yr BP in southern central British Columbia. Increase in mesotrophic diatoms species at Felker Lake, further suggests long, cool, springs and/or periodic windy conditions that induce mixing and a redistribution of nutrients and heat throughout the water column from 6700 to 5700 cal yr BP (Galloway et al., 2011). The expansion of Salix and Betula commonly found in moist swamps and other moist habitats occurred throughout the region by 6290 cal yr BP

(Galloway et al., 2011). The decrease in diatom inferred depth of the Felker Lake and decrease in pollen and spore accumulation rates of Betula, Salix, Alnus, Picea and Abies after 5810 cal years

B.P. suggest a return to a period with low effective moisture. A gradual rise in western hemlock was recorded 8850 cal yr BP (8000 14C years BP) at Marion Lake (Mathews and Heusser, 1981).

This rise was followed by true fir and birch 8000 cal yr BP (7200 14C years BP). Fossil midge- based reconstructions from Cabin Lake recorded a rapid cooling around 8400 cal yr BP (7500

14C years BP), while 3M Pond recorded a gradual decline in temperature between 7500 and 4000

14C years BP (8400 and 4525 cal yr BP). Considering the average recoded values, one can

conclude there was a gradual temperature decline in the southern interior of British Columbia compared to the rapid decline in coastal British Columbia (Walker and Pellat 2003).

Advancement of glaciers in the Canadian Cordillera between 7360 and 6450 cal yr BP provide further evidence for cool and moist climatic conditions (Galloway et al., 2011 and references therein). Increase in aeolian activity reported from lake and dune records from Great Plains is interpreted as indications of strengthened Westerlies around 6000 years BP (Galloway et al.,

2011).

Late Holocene Climate of British Columbia

The Late Holocene is described in the literature as a period of variable climatic conditions in southern British Columbia. Several late Holocene lake level fluctuations associated with regional glacial advances have been identified from the region (Gavin et al., 20011). A shift in moisture regime with frequent droughts are reported from Dog Lake (Hallet and Hills, 2006)

Big Lake (Grimm et al., 1992), and several other lakes in western Canada between 2400 to 1200 cal yr BP (Laird et al.,2003; Benett et al., 2001; Cumming et al., 2002). Increased fire activity and transition of vegetation from wet, closed forests to dry, open forests around Dog Lake further strengthen the interpretation of a dry period. Evidence of high fire frequency is also reported from the coast and Cascade Mountains of British Columbia between 2400 to 1200 cal yr

BP. (Hallet and Hills, 2006 and references therein). This period of high fire frequency with frequent summer droughts is named “the Fraser Valley Fire Period” (Hallet and Hills, 2006; and references therein). The Late Holocene is identified as a period of considerable variation in pollen accumulation rates and lake levels in Felker Lake (Galloway et al., 2011). Diatom-based lake levels from Felker Lake report three events with extremely low lake levels, high salinity,

and high specific conductivity during late Holocene (1910 to 1800, 1030 to 690 cal yr BP;

Galloway et al., 2011). Similar fluctuations in lake levels associated with regional glacial advances are recorded from the Eleanor Lake during the late Holocene (2400, 1400 and 800 cal yr BP; Gavin et al., 2011). Century-scale lake level fluctuations have also been recorded from saline Mahoney Lake during the past 1500 years, with six short periods of low effective precipitation alternating with periods of high effective precipitation (Lowe et al.,1997). Several paleoclimatic reconstructions from the Canadian Rocky Mountains suggest solar minima are associated with glacial advances and high lake levels (Lowe et al., 1996; Galloway et al., 2011.

Multiple studies explain the relationship between solar irradiance and the climate of the Pacific

Northwest. One possible mechanism to explain this pattern in the Pacific Northwest is the link between the solar minimum and the sea surface temperature of the Pacific Ocean (Mann et al.,

2005; Marchitto et al., .2010). It has been suggested solar activity is inversely correlated with sea surface temperature (SST) of the Eastern Tropical Pacific Ocean (ETP). As a result, periods of low solar activity create a warm ETP giving rise to warm-dry El Niño-like conditions in the

Pacific Northwest, in contrast with wet conditions in south-western America (Marchitto et. al.,

2010) (Steinman et al., 2014).

Table 1.1. Summary of previous paleoclimatic reconstructions from British Columbia (BC)

Period Recorded climatic conditions Recorded locations (within BC) 14000 - Low July temperature (1-2 °C Marion Lake (Mathews and 12000 cal yr colder than present), moderately Heusser,1981), Windy Lake (Chase et BP high (1900 mm ±620 mm) al., 2008) precipitation

Early Precipitation dropped down to Marion Lake (Mathews and Holocene 1450 (±620 mm), temperature Heusser,1981), Frozen Lake, North 12100- increase up to 2°C (±1.1 °C) Crater Lake, Lake-of-the-Woods, 11550 cal Windy Lake (Chase et al., 2008) years B.P

11500 to Higher July temperature Frozen Lake, North Crater Lake, Lake- 8500 cal yr of-the-Woods, Windy Lake (Chase et BP al., 2008) Drop in Tsuga heterophylla, Coastal BC (Alley 1996; Walker and

Thuja plicata, Abies), increase Pellat, 2004) in xerophytic taxa (Pinus Okanagan Valley, Southern British juniperus, Artemmisia, Columbia (Pellat et al., 2002) Poaceae, Pseudotsuga menziesii, Alnus rubra (warm dry climate) Higher tree line in BC Castle Peak in the southeastern Coast Mountains of British Columbia (Chase et al., 2008) Decrease in lake levels Dog lake (Hallet and Hills, 2006) Increased aeolian dune activity Central Great Plains of North America (Hallet and Hills, 2006)

Mid Holocene 8500-6700 Moist (wet) Big Lake (Grimm et al., 1992) 8400-4525 Decline in temperature Cabin lake, 3M pond (Walker and Pellat, 2004) 8400-6800 Increase in precipitation, decrease Marion Lake (coastal BC, Mathews and in temperature Heusser, 1981) 6300-5650 Moist (wet) Mahoney Lake 6700 -5700 Long, cool, springs and/or Felker Lake (Galloway et al., 2011) periodic windy conditions, expansion of moist habitat trees 7360 - 6450 Advancement of glaciers Canadian Cordillera (Galloway et al., 2011) Late Frequent droughts and wet events Dog Lake (Hallet and Hills, 2006), Big Holocene varying in century scale Lake (Grimm et al., 1992), Mahoney Lake (Lowe et al.,1997), Laird et al.,2003; Benett et al., 2001; Cumming et al., 2002

2400-1200 cal Frequent droughts Dog Lake, Coast and Cascade Mountains yr BP Increased fire frequency, open (Hallet and Hills, 2006); Big Lake forests (Grimm et al., 1992) 1910 -1800, Low Lake Levels Felker Lake (Galloway et al., 2011) 1030 - 690, 250- 140 cal

1500 cal yr Six periods of low lake levels, and Mahoney Lake (Lowe et al.,1997) BP - present high lake levels

CHAPTER 2

Reconstruction of Late Quaternary paleohydrologic conditions in southeastern British Columbia

using visible derivative spectroscopy of Cleland Lake sediment

(Manuscript in review in Quaternary Research)

Abstract

Visible Derivative Spectroscopy (VDS) is a quantitative, non-destructive method of sediment analysis. VDS analysis of sediment from Cleland Lake, Southeastern British Columbia provides a reconstruction of paleolake productivity and hydrologic change during the past 14000 calibrated years before present (cal yr BP). The first five principal components (PC) of the VDS data explain 95% of the variance in the sediment reflectance. Four PCs correlate with standard reflectance curves for diatom, dinoflagellate algae, and cyanobacteria pigments that record ecological change, while two PCs are paleohydrologic indicators. Dinoflagellate algae are predominant from 11600-8600 cal yr BP then decrease to low levels after ~8600 cal yr BP. PCs

3-5 represent variations in cyanobacteria abundance and exhibit peaks from 14000-11600,

14000-9500 and 6100-5400 cal yr BP, respectively. Conditions promoting cyanobacteria shifted toward those favoring diatoms around 9400 and lasted until 170 cal yr BP. Higher dinoflagellate- related pigment concentrations suggest a lower lake level from 11600- 8600 cal yr BP, followed by higher water levels and wetter conditions after 8500 cal yr BP. We propose that drier

24 conditions transitioning from the late glacial into the Holocene were caused by insolation-driven, non-linear feedbacks between the Pacific subtropical high-pressure system, vegetation, and soil moisture.

Introduction

In western North America large droughts similar to those of the last decade were a common occurrence during the 20th century (Cook et al., 2004). Studies of paleoclimate proxy data have revealed that more intense droughts of longer duration occurred in the desert

Southwest during the Medieval Warm Period (~900 to 1300 AD) (Cook et al., 2004). Multi- decadal drought variability in western North America is largely attributable to Pacific ocean- atmosphere dynamics involving the El Nino southern Oscillation (ENSO) and the Pacific

Decadal Oscillation (PDO) (McCabe et al., 2004; Mann et al., 2005; Nelson et al., 2011;

Steinman et al. 2012), as well as the teleconnected influence of the Atlantic Multidecadal

Oscillation (AMO) and the North Atlantic Oscillation (NAO) (McCabe et al., 2004; Cook et al.,

2004). Although drought variability during recent centuries is relatively well understood

(McCabe et al., 2004; Cook et al. 2004; Steinman et al. in press), the frequency, duration and forcing mechanisms of hydroclimate variations on centennial to millennial timescales remains poorly defined, particularly during the Pleistocene-Holocene transition (Galloway et al., 2011).

Holocene climate is thought to have been relatively stable in comparison to the dramatic changes that occurred during the late glacial period (Mayewski et al., 2004; Overpeck and Cole

2006). However, recent findings indicate that substantial climate variability occurred during the

Holocene over timescales important to humans (years to decades) (e.g. Alley et al., 2003). Such abrupt events cannot be resolved using the relatively short instrumental climate record; therefore

it is imperative to study longer, highly resolved climate proxies such as those attainable from lake sediment.

In semi-arid regions, small, closed-basin lakes are sensitive to changes in regional precipitation/evaporation (P/E) balance resulting from climate change. In these lakes, P/E balance variability produces hydrologic instability, inducing water level variations that are reflected in the physical, biological, and chemical composition of sediment accumulating in the basin (Talbot 1990; Battarbee, 2000; Leng and Marshall 2004: Steinman et al., 2010ab;

Pompeani et al., 2012). Large magnitude changes in sediment characteristics can occur in response to several factors, most notably large magnitude water level variations resulting from non-linear responses to climate system feedbacks (deMenocal et al., 2000; Fritz, 2008). Previous work has demonstrated that spectral reflectance, particularly Visible Derivative Spectroscopy

(VDS), is an efficient and non-destructive method of quantitatively analyzing organic matter content (including plant pigments), clays, carbonates and iron oxides in sediment (Ortiz et al.,

1999; 2009 and 2011) for the development of paleoclimate reconstructions.

Fossil pigments are an important constituent of the organic matter preserved in lake sediment and can be used as a proxy for primary productivity within the lake and changes in phytoplankton community structure (Peteet et al., 2003; Das et al., 2005; Wolfe et al., 2006;

Bouchard et al., 2013). Concentrations of sedimentary photosynthetic pigments have long been used as a direct and accurate method to reconstruct the response of lake phytoplankton to environmental changes (Sanger and Gorham, 1972; Sanger and Crowl, 1979; Sanger, 1888; Nara et al., 2005). Phytoplankton composition in temperate lakes tends to follow a succession pattern similar to that of plants on the landscape and is strongly influenced by water chemistry and lake level. Biological communities in drought-sensitive lakes are therefore a reflection of water

balance changes, such that algal pigment analysis of sediment from these lakes can be used to investigate past hydroclimate variability. The aim of this study is to understand the timing and magnitude of environmental change in the Pacific Northwest, particularly during the late

Pleistocene and early Holocene by reconstructing paleo-lake productivity and hydrologic balance variations. To this end we present the succession of phytoplankton communities at Cleland Lake, an alkaline, surficial, closed-basin lake basin in southeastern British Columbia, using VDS analysis of lake sediment cores.

Study Area

Cleland Lake (50.82° N, 116.38° W, 1158 m asl) is located in the north-south trending

Columbia Valley just west of the continental divide in the Rocky Mountains of southeastern

British Columbia (Fig. 2.1). The lake has a surface area of 0.235 km2, has a maximum depth of

31.7 m, and is typically covered with ice from November to May. Meteorological data from

Meteorological Services of Canada’s Brisco weather station (ID 1171020, 10 km east of the lake, elevation 823 m) document a mean annual temperature of 11.2 °C over the period from 1984 to

2003 AD. January mean temperature was -3.7 °C and the July mean temperature was 24.8 °C.

The mean annual precipitation amount during the period from 1924 to 2003 AD was 426 mm.

Surface inflow is limited to the relatively small, immediate catchment, while water losses are restricted to evaporation and presumably to groundwater seepage.

B A

Figure 2.1 Cleland lake location map A. Map of British Columbia showing the location of Cleland Lake (red dot) and selected sites referenced in the text (black triangles), 1: Eleanor Lake, 2: Windy Lake, 3: Frozen Lake, 4: North Crater Lake, 5: Lake of the Woods. B. Topographical map of watersheds in the Cleland Lake locale. C. Limnological measurements of the Cleland Lake water column in July 2011: temperature (T) dissolved oxygen (DO), specific conductivity (SC), and pH.

Methods

Sediment Core Collection

Four sediment cores (B-09, C-09, E-09, and F-09) were collected from Cleland Lake in

May and July 2009. Cores F-09 and E-09 were collected using a square rod, Livingston piston corer at water depths of 7.0 m and 12.8 m, respectively. The uppermost sediment depth of cores

B-09, E-09, and F-09 is 30 cm below the sediment water interface. Core E-09 (shallow water) consists of four overlapping drives extending to a sediment depth of 2.6 m. Core F-09

(intermediate water) consists of six overlapping drives that extend 3.3 m below the sediment- water interface. Core drive overlaps were identified by visual examination of sedimentological features. Core B-09 (deep water) is 2.5 m long and was collected with a polycarbonate piston corer at a water depth of 20.2 m using a percussion corer. Core C-09 was recovered from a water depth of 24 m using a freeze corer filled with a mixture of ethanol and dry ice. Individual

Livingston core drives were wrapped in polyvinyl plastic, and sealed in PVC casing in the field.

All cores were transported to the Department of Geology and Planetary Science at the University of Pittsburgh. Cores were stored at 4ºC (the freeze core was stored at -10º C) prior to sampling and analysis. Water temperature, pH, dissolved oxygen, conductivity, and alkalinity were measured using a Hach® Hydrolab water quality sonde and Hach® Digital Titrator.

Geochronology

Bulk sediment samples were digested in 7% H2O2 for ~12 hours and sieved at 63 µm.

Terrestrial macrofossils (> 63 µm) were picked using a small brush under a light microscope for radiocarbon dating. Samples were treated with an acid-base-acid wash and rinsed with deionized water to a neutral pH prior to being analyzed at the W.M. Keck Carbon Cycle Accelerator Mass

Spectrometry Laboratory at the University of California, Irvine (UCI) (Abbott and Stafford,

1996). The F-09 core chronology is based on four AMS 14C dates on terrestrial macrofossils and two tephra layers of known age (Table 1). A linear point-to-point age model was generated using the CLAM 1.0.2 software with IntCal09 calibration curve (Blaauw, 2010; Reimer et al., 2009).

For each age control point, a 95% confidence interval was also calculated. The basal age was extrapolated following the trend established by the two oldest age model control points (Figure

2.2 ).

Visible Derivative Spectroscopy

Visible derivative spectroscopy (VDS) is used to define sediment pigment content based on the shape of the diffuse spectral reflectance (DSR) of components extracted from the derivative spectrum for each sample. Diffuse spectral reflectance was measured at 0.5-cm intervals on the wet, split-core sediment surface of core F-09 using a handheld Minolta CM-

2600d spectrophotometer (Konica-Minolta USA, Ramsey, NJ, USA). This instrument measures the reflectance over the visible range of the electromagnetic spectrum (380-700 nm) at 10 nm resolution. Sediment reflectance of the dried, freeze core samples was measured using an ASD

LabSpec Pro FR UV/VIS/NIR spectrometer using a high intensity contact probe specifically designed for measuring powdered samples. This instrument measures diffuse reflectance over the visible and near infrared range of the electromagnetic spectrum (350-2500 nm) at ~4 nm resolution in the visible and 10 nm resolution in the near infrared (NIR). The resulting spectrum, which is oversampled by the instrument and reported at a 1-nm interval, was band-averaged to

10-nm reflectance for direct comparison with the Minolta data. We analyzed data in the visible range (400-700 nm) to avoid noise in the ultraviolet (UV) and differences in the reflectance and absorption processes in the NIR relative to the Visible (Ortiz, 2011). Dry sediment is brighter than wet sediment; therefore we compared wet and dry sediment measurements to quantify the offset. The measured reflectance of the powered samples had to be adjusted for the dryness of the samples by multiplying each spectrum by a constant scaling factor of 0.5, to directly compare the ASD measurements with the Minolta measurements (Appendix 1).

Principal component analysis

(Supplementary Figure 2). VPCA was conducted in SPSS™ 14.0 using the full data set (1751 reflectance measurements) based on all four cores to allow comparison of results between cores.

The components were identified by comparing the principal component loading spectra for each of the PCs (Supplementary Figure 3) with center-weighted reflectance derivative spectra for previously published pigment or known mineral standards in the USGS spectroscopic library or minerals measured in Ortiz’ lab at Kent State University (Gantt, 1975; Robertson et al., 1999;

Graham and Wilcox, 2000; Schagerl and Donabaum, 2003; Schagerl et al., 2003; Toepel et al.,

2005; Clark et al., 2007; Ortiz et al., 2009; Yurco, 2010; Ortiz, 2011). We quantified the relative importance of each pigment or mineral in the PC by fitting a weight-average of the empirically selected standards to each principal component-loading pattern. The quality of the fit was determined by linear correlation. This process is identical to the approach previously applied to marine cores (Ortiz et al., 2009; Ortiz 2011; Bouchard et al., 2013) and similar to spectral matching methods used in X-ray diffractometry (Eberl, 2003 and 2004).

X-ray Fluorescence

The bulk elemental composition of the lake cores was measured using the ITRAX X-ray

Fluorescence (XRF) core scanner at the University of Minnesota, Duluth. The scanner has the potential to scan cores at 0.2 mm resolution and consists of a molybdenum X-ray source. Cleland

F-09 core was scanned at 5 mm resolution at 30 mA and 45 kV over a scan time of 60 seconds.

The Si:Ti ratio determined using the scanning XRF elemental data correlates with biogenic silica abundance and hence can be used as a proxy of diatom abundance (Brown et al., 2007; Brown

2011; Johnson et al., 2011).

Results

Lake Limnological Properties

Cleland Lake is stratified during the summer with a thermocline depth of ~5 m (Figure

2.1). In August 2011 the epilimnion had an average temperature of 19.7°C, dissolved oxygen concentrations of 9.2 mg/l and a pH of 9.75. The hypolimnion (i.e. water deeper than 5m to the lake floor at 29 m) had an average temperature of 5.8°C, average dissolved oxygen concentration of 1.6 mg/l and a pH of 9.5. Total dissolved solids (TDS) concentration vary between 510 mg/l and 550 mg/l from to surface to the lake bottom. The salinity of water varies from 0.42 ppt at the surface to 0.44 ppt (420 mg/l and 440 mg/l) at the bottom (Appendix 1).

The lake water is alkaline, with an average dissolved bicarbonate and carbonate concentration of ~530 mg/l, which promotes the production of micritic calcium carbonate in the water column during summer whiting events. The lake water is anoxic below 12 m, which reduces sediment bioturbation and aids preservation of plant pigments within sub-mm laminations.

Age Model

The age model for Core F-09 spans the period from 11300 ± 1050 cal yr BP to present

(Table 1.1, Figure 2.2). Linear extrapolation of the last two age control points gave an age of

14305 ± 1300 for the base of the core. The thickness of the error envelope ranges from ±50 to

±100 years at the top of the core, increasing to ±200 to ±500 years between the depths of 225 to

250 cm. Uncertainty in the basal ages increases from ±1000 cal yr BP at an age of ~11300 cal yr

BP to ±1300 cal yr BP at the base of the core, which we estimate has an age of ~14100 cal yr BP.

Figure 2.2. Age vs. depth model of Cleland Core F-09 (0 to 361.5cm) based on four radiocarbon dates and two tephra layers. Red dotted lines represent the 95 % confidence interval.

Table 2.1 AMS Radiocarbon and tephra dates for Cleland Lake Core F-09. * Omitted 14C ages.

14C yr Median Core 2σ lower 2σ upper UCAIMS before calibrated Material Depth (cal yr (cal yr No. present age (cm) BP) BP) (BP) (cal yr BP)

84885 charcoal 35.5 210±80 200 -10 430

Mt. St.

Helens 66.5 480 470 490

tephra layer

84851 charcoal 162.5 3660±20 4000 3920 4100

Mazama 204-208 6850±100 7600 7400 7760 Tephra layer

84852 charcoal 214 7435±30 8100 8030 8280

99884 charcoal 286 9800±350 11300 10280 12390

extrapolated 357.5 N/A 14100 12800 15500 basal age

84886 charcoal 287 780±20* 700 680 730

Interpretation of the Principal Components

The first five principal components (i.e. PCs) of the reflectance data explain 97% of the variance (Table 2.2). Principal component 1 explains 28% of the variance and is interpreted as a mixture of illite (75%) and sphalerite (30%) (Figure ). Source of this sphalerite is a metaliferous horizons consisting of sphalerite and bornite (Simandl and Hancock, 2002),

located approximately 620 m southeast of the southern lake margin. This PC can be used as a proxy for fluvial input to the lake. Illite, which is present throughout the record, likely represents the autochthonous clay mineral within the lake basin. Principle component 2 explains 24% of the variance in the reflectance data and reflects a mixture of dinoflagellate algae pigments (peridinin,

Pheophytin-A, 60% and 33% respectively) and goethite (7%: Figure 2.3). Major photosynthetic pigments of flagellated algae belonging to the division Pyrrophyta include chlorophyll-a, chlorophyll-c, and peridinin (Sze, 1993). The presence of peridinin and Pheophytin-a (a degredation product of chlorophyll-a) in PC 2 identifies this component as a proxy for dinoflagellate abundance in the lake (Sze, 1993). Dinoflagellate algae are common in oligotrophic, high latitude/polar lakes, or in water with low light intensities, such as snow covered lakes and turbid water (Sze, 1993). Large, positive PC 2 values indicate greater dinoflagellate abundance and low nutrient conditions, or low light intensity. Principal components 3, 4 and 5 contain variable amounts of blue-green algal pigments and hence may represent the succession of three different blue-green algal communities in Cleland Lake. PC 3 explains 23% of the variance and is a mixture of phycocyanin (80%), a pigment resulting from cyanobacteria (cyanobacteria), and the clay minerals smectite and chlorite (20%, Figure 2.3).

Smectite and chlorite, which are present largely near the base of the core may indicate climate- related, allochthonous clay minerals transported into the basin during deglciation of the

Cordilleran Ice Sheet (Rhoton et al., 1979). Climate is one factor among many, however, that govern the formation and deposition of clay minerals in lake basins (e.g. parent material, sedimentary facies, and diagenetic alterations), such that interpretations of clay concentration data should be supported by other proxies (Liu et al., 2007; Ruffel et al., 2002; Folkoff and

Meentemyer, 1987). Principal component 4 explains 16% of the variance and is a mixture of the

spectral signatures for Bacillariophycea (70%) and Phycocyanin (30%) (i.e. diatoms and cyanobacteria photosynthetic pigments, Figure 2.3). The spectral signature for Bacillariophycea has a positive correlation with PC 4, indicating greater diatom abundance when the PC 4 scores increase. In contrast, phycocyanin reflects a negative correlation with PC 4, indicating lower levels of cyanobacteria when PC 4 increases and vice versa. This defines the inverse relationship and contrasting appearances of diatoms versus cyanobacteria in sediment from Cleland Lake. PC

5 explains six percent of the variance of reflectance data is related to the cyanobacteria accessory pigments, allophycocyanin (54 %) and pheophytin-A (46%) which is a degradation product of chlorophyll –A, (Figure 2.3). Therefore this component is considered as a proxy for a third cyanobacteria assemblage. For the remainder of this study, we focus our discussion on PC 2 and

PC 4, which are proxies for paleolake productivity and water level, respectively. Interpretation of the down-core patterns associated with the remaining components (PC’s 1, 3, and 5) is beyond the scope of this study, although we make reference to them as warranted (Figure 2.3).

Table 2.2. Total variance explained by the first four components from the Principal component analysis

Component Percent of variance explained Cumulative percent of variance explained

1 27.95 27.95

2 23.94 51.89

3 23.17 75.06

4 16.32 91.38

5 5.7 97.08

1 A 0.06 0.02 0.8 B 0.8 0.04 0.6 0.01 0.6 0.02 0.4 0.4 0.2 0

0 sphalerite)Illite+ ( 0.2

PC 1componentscore PC 0 PC 2 component 2 score PC

0 -0.02 -0.2 -0.01 peridinin Pheophytin +Pheophytin Geothite peridinin 400 500 600 700 400 500 600 700 wavelength(nm) wavelength (nm) 1 -0.015 1 0.02 C D 0.8 0.5 0.6 0.005 0.4 0 0 0.2 0.025 -0.5

component score component 0

PC 4 component 4 score PC -1 -0.2 0.045

-0.4 -1.5 -0.02 Bacillariophyceae+Phycocyanin

400 500 600 700 Phycocyanin+Smectitie+chlorite 400 500 600 700 wavelength (nm) wavelength (nm) 0.7 0.02

a -

0.3 0.01

-0.1 0 PC 5 component 5 score PC

-0.5 -0.01 +pheophytinallophycocyanin 400 500 600 700 wavelength (nm)

Figure 2.3. Comparison of first five Principal components with their reference spectra; primary y axis represent component scores while the secondary y axis represent the first derivative of the percent reflectance of the reference spectra, black line represent the principal component while the red line is the correlating reference spectrum for each component. A) Graph showing PC 1 (black line) and the first derivative spectra of the 95 % illite+5% sphalerite (red line) mixture. B) Graph showing PC 2 (dinoflagelltae algae scores, black line) and the first derivative spectra of the 60 % peridinin + 33 % phaeophytin-a + 7 % goethite (red line) mixture. C) Plot of PC 3 scores (clay and blue green algae, black line) and the first derivative spectra of the 80% phycocyanin + 20% (smectite+chlorite) mixture; (red line), D) Plot of PC 4 scores (Diatoms and blue green algae); black line and the first derivative spectra of the 70% bacillariophycea + 30% phycocyanin mixture (blue line). E) Graph showing PC 5 ( black line) and the first derivative spectra of the 54 % allophycocynain + 46% pheophytin-a mixture (red line)

Several factors promote pigment preservation in sediment. For example, greater lake depths increase stratification, promote low oxygen levels (or anoxia) and decline in bottom water temperature and inhibit resuspension of sediment and degradation of organic matter, thereby promoting pigment preservation (Sanger 1988). In addition, low light conditions and high sedimentation rates promote the preservation of pigments by reducing photochemical decomposition (Sanger 1988). Alkaline waters tend to preserve pigments as opposed to acidic waters, which degrade chlorophyll (Sanger 1988). Previous research documents high pigment preservation potential in shallow, productive lakes (Sanger and Gorham 1972; Sanger and Crowl

1977; Sanger 1988). At Cleland Lake, cold (~6°C), alkaline (pH ~9.5), and anoxic water conditions occur below 5 m depth (Figure 2.1). Such conditions enhance the preservation of pigments at all three core sites. The presence of laminated sediment throughout the record (with the exception of between ~8000-10000 cal yrs BP) suggests minimal sediment resuspension and improved pigment preservation. It is possible that past variability in pigment preservation affected sediment composition, particularly from 8000 to 10000 cal yr BP when fewer laminae exist. However, if pigment preservation was reduced during this period, our estimates would not suggest an increase in abundance of dinoflagellate and blue-green at that time. We infer that there was little or no difference in pigment preservation at the different core sites and that this did not vary through time, because there is no evidence to support differential preservation.

Discussion

Evolution of the Lake Phytoplankton Community over the Holocene

Because Core F-09 is the only core that extends through the Holocene with reliable age control (to 11300 (±1000) cal yr BP), we use it to study the evolution of the phytoplankton

community in the lake. Variations in PC 2, 3, 4 and 5 suggest a post-glacial succession of various classes of algae (Figure 2.5). The earliest part of the record, prior to 11800 (±1100) cal yr BP (i.e the deglacial period), is characterized by low diatom, cyanobacteria, and dinoflagellate abundances. Dinoflagellate algal pigments record a gradual increase in abundance starting around 12400 (±1100) cal yr BP based on increases in PC 2 values (Figure 2.3 and Figure 2.5).

Dinoflagellate levels rise rapidly between 11850 and 11600 (±1200) cal yr BP, reach the highest values in the record by 11600 (±1100) cal yr BP, and remain high until 8600 (±200) cal yr BP with excursions centered around 10700 (±900), and 8750 (±400) cal yr BP. Dinoflagellate productivity remained low during the middle and late Holocene with several millennial- and centennial-scale fluctuations centered around 7500 (±200), 6400 (±100), 3250 (±75), 2450 (±50),

1550 (±20) and 800 (±10) cal yr BP. Principal component 3 represents an early successional cyanobacteria community (which we denote B-G community 1), that was prominent in the lake prior to 11600 (±1200) cal yr BP (Figure 2.5). Based on increasing PC 3 component scores, we suggest that this community rapidly decreased after 11600 cal yr BP and reached low levels by

10550 cal yr BP. The abundance of this community remained low throughout the rest of the record (10500±800 to 170±200 cal yr BP). The early Holocene drop in B-G community 1 was simultaneous with the rapid increase in dinoflagellate algae abundance inferred from PC 2.

Principal component 4 also represents an early successional blue green alga community (B-G community 2), which appeared in the lake during the late glacial period but remain abundant longer than B-G community 1 (14100±1300 to 9500±500 cal yr BP). Principal component 5 represents a third cyanobacteria community (B-G community 3) that become dominant in the lake during mid Holocene (6100±100 to 5400±100 cal yr BP) (Figure 2.5). Low PC 5 scores suggest that, B-G community 3 was low in abundance during the deglacial period, but gradually

increased after 11500 cal yr BP, then peaked during mid-Holocene. B-G Community 3 gradually declined after 5400 cal yr BP, then remained low in abundance throughout the late Holocene

(1850±50 to 170±200 cal yr BP).

Diatom abundance suggested by PC 4 is low prior to 11800 (±1200) cal yr BP, then increases rapidly during three intervals centered on 11550, 11400 and 11200 (±1200) cal yr BP

(Figure 2.3 and Figure 2.5). Diatom levels gradually drop to their lowest values after 11100

(±1000) and remain lower between 10800 (±700) to 9500 (±400) cal yr BP, while cyanobacteria community 2 levels rise to their highest Holocene levels. Pigment concentrations indicate an increase in diatoms after 9500 (±500) cal yr BP and high values from 8500 (± 200) to 170 (±200) cal yr BP with several millennial- and centennial-scale oscillations. Overall, diatom levels follow an increasing trend during the middle to late Holocene, while cyanobacteria community 2 remains low from 9000 (±400) to 170 (±200) cal yr BP.

Phytoplankton abundances reconstructed by PC 2 and PC 4 significantly correlates with the Northern hemisphere summer insolation trend. PC 2 has a positive correlation and exhibits a more abrupt response (r = 0.6, n = 59, p < 0.05) to insolation, indicating that higher dinoflagellate abundance 11600 (±1100) to 8600 (±200) cal yr BP is correlated with greater summer insolation. PC 4 shows a negative correlation with insolation (r = -0.7, n = 73, p < 0.05), indicating low diatom abundance at lower insolation values (Figure 2.5).

Biogenic silica concentrations in lake sediment provide insight into lake diatom productivity and can be related to millennial scale climatic changes (Williams 1995; Brown

2011; Johnson et al., 2011). Comparison of the PC 4 down core record with the XRF-derived

Si:Ti ratio supports the VDS interpretation (Figure 2.5). The Si:Ti ratio is lower during the late glacial period (prior to 12000 cal yr BP) indicating low diatom abundance. The ratio gradually

increases between 12000 to 10500 cal yr BP and rapidly decreased after 10300 cal yr BP. The

Si:Ti ratio remain low between 10300 and 9500 cal yr BP simultaneous with the period of low diatom abundance. The ratios gradually increase after 9500 cal yr BP until 170 cal yr BP, corroborating with PC 4-derived diatom abundance (Figure 2.5). The transitions between low and high diatom periods inferred from the Si:Ti ratio are synchronous with the diatom abundance derived from PC 4. However the two records are not perfectly matched because post-depositional processes may affect pigment and silica preservation, and PC 4 is a mixture of 70% diatom pigment and 30% cyanobacteria pigments.

Paleohydrological Conditions Inferred from Phytoplankton Abundance

Water level influences light penetration depth, and is related to nutrient influx and water column stratification, which affects nutrient supply to the epilimnion (Nõges and Nõges, 1999;

Battarbee 2000). These factors in turn affect the biological productivity in the lake. Hence, water level changes in temperate lakes are a strong control on lake biological productivity (Nõges and

Nõges, 1999; Battarbee, 2000; Battarbee, 2000; Nõges et al., 2003; Heinsalu et al., 2008). At lower water levels, in the presence of improved light conditions, nutrient limitation controls phytoplankton availability (Nõges et al., 2003). Such conditions lead to an increase in species adapted to low nutrient availability such as dinoflagellates (Sze, 1993, section 4.3). Diatoms are one example of lake primary producers that are sensitive to changing hydrologic conditions, such as water level changes, salinity and nutrient concentrations (Barker et al., 1994; Nõges et al.,

2003; Heinsalu et al., 2008). Total diatom abundance (planktonic and littoral) increases exponentially with water depth (Baker et al., 1994). Diatom abundance is a common proxy used to reconstruct lake water depth (Barker et al., 1994; Brugam et al., 1998; Gavin et al., 2011), with increases in lake level associated with higher diatom abundance in sediment (Barker et al.,

1994; Heinsalu et al., 2008). Variations in the diatom species composition, as well as the ratio between the littoral/planktonic species, can indicate water level changes that shift littoral sedimentary environments away from or towards the center of the basin (Barker et al., 1994;

Brugam et al., 1998; Gavin et al., 2011). However, because the VDS approach only identifies the total diatom (bacillariophycea) pigment concentration and does not distinguish between benthic and planktonic species.

In addition to water level, temperature and nutrient cycling also plays a major role in controlling phytoplankton composition (Nõges et al., 2003; Grover and Chrzanowski, 2006 and references therein). Cyanobacteria predominate during the summer months and in warm climates due to their tolerance for higher temperatures. In contrast diatoms are more prevalent in cold water (Nõges et al., 2003). Diatom abundance is also influenced by the availability of nutrients.

For example, in dimictic lakes, higher temperatures can lead to lake-water stratification, producing nutrient depletion in the epilimnion (Barker et al., 1994; Kalf, 2002). Lakes in temperate regions like southern British Columbia tend to be stratified in the summer, and often overturn and mix twice a year, resulting in the resuspension of nutrients and two prominent periods of diatom blooms during spring and fall. When the availability of nutrients becomes limited, diatoms can be replaced by taxa such as green algae or cyanobacteria (Sze, 1993; Nõges et al., 2003; Grover and Chrzanowski et al., 2006 and references therein). There are several other factors, however, like hydrology, herbivore biomass, turbidity, and light that can also affect phytoplankton abundance (Flores and Barone, 1998; Fabbro and Duivenvoorden, 2000; Kruk et al., 2002). This diversity of controlling mechanisms not only highlights the complexities of using biological proxies to measure climate, but also shows how the biological conditions of the lake are intricately linked to many aspects (rather than primarily just one) of the ecosystem and

climate. Despite the many potential responses of phytoplankton to environmental changes, we use phytoplankton pigments to infer the paleo-environmental conditions at Cleland Lake, and support our interpretation using multiple proxies and cores from different water depths in the basin. In summary, PC 2 from Core F-09 is used as a proxy for lake depth, where higher PC 2 values indicate greater dinoflagellate abundance and a shallower lake, and conversely low PC 2 values indicate decreased dinoflagellate abundance and a relatively deeper lake. PC 4 is used as a temperature and nutrient proxy. Higher PC 4 values indicate a deeper lake and/or cold, nutrient- rich conditions, with greater diatom abundance and less cyanobacteria. Low PC 4 values indicate a lower diatom and greater cyanobacteria abundance typical of a nutrient-depleted stratified lake.

Vertical and horizontal heterogeneity in the spatial distribution of phytoplankton in lakes has been studied since the mid-20th century (Baldi 1941; George and Heaney 1978; Heaney and

Talling 1980; Wu et al., 2010). Several investigations support the idea of pronounced horizontal variability of phytoplankton abundance associated with wind-induced water movement (George and Heaney, 1978; Heaney and Talling, 1980; Wu et al., 2010). Considerable variation of phytoplankton amounts in the presence of buoyant cyanobacteria and dinoflagellate algae

Ceratium was reported from Esthwaite Lake in the Lake District of England (George and

Heaney, 1978; Heaney and Talling, 1980), indicating that freshwater algae can develop vertical and horizontal zonation even within small lake environments. A scatter plot of PC 2 vs. PC 4 from Cores B-09 (deep water), F-09 (intermediate), and E-09 (shallow) demonstrates that component PC scores vary with lake depth (Figure 2.4, Table 2.3). Sediment from the shallow water core contains high concentrations of dinoflagellate and cyanobacteria related pigments, and low concentrations of diatom pigments; while the deep-water sediment core indicates a similar cyanobacteria and diatom abundance, but lower dinoflagellate pigment concentrations.

The intermediate core (F-09) contains a lesser amount of dinoflagellate pigment than the shallow water core, but greater amounts of dinoflagellate pigment than the deep-water core. The distribution of PC 2 and PC 4 in Core F-09 corresponds with the trends at the other two core sites from 8800 to 170 cal yr BP. However, prior to 8800 cal yr BP, diatom and cyanobacteria values deviate from the general trend observed in the other cores (Figures 2.4 and 2.3) likely due to the facies changes associated with early Holocene lake level variability. Between 11600 and 8600 cal yr BP, cyanobacteria and dinoflagellates values are high while diatom values are low.

Conversely, prior to 11600 cal yr BP, diatom and dinoflagellate abundances are low, but cyanobacteria abundances are high. The comparisons of PC 2 and PC 4 from the three cores at different water depths suggests that PC 2 (dinoflagellate abundance) at core site F-09 is influenced by lake depth, reflecting changes from shallow to deep water facies when the dinoflagellate abundances increase. Principal component 4 (diatom and cyanobacteria abundance) does not show such a clear connection with lake depth. For example, although Core

B-09 (i.e. the deep water site) has higher diatom abundance than the shallow core (E-09), it contains fewer diatoms than the intermediate water core (F-09), complicating the use of PC 4 as a lake-level proxy. This may be due to the absence of benthic diatom species in the deep-water sediment or redox-related preservation issues as a result of physiochemical disparities between the different core sites.

The triangular distribution and spread of the data points in the cross plot between PC 2 and PC 4 for Core F-09 indicates that the abundance of the three primary producers is affected by several factors, including temperature and nutrient availability (Figure 2.4, Grover and

Chrzanowski, 2006; Noges et al., 2003 and references therein). Nonetheless, the scatter plot of

PC 2 and PC 4 demonstrates that water depth is an important factor that affects pigment/phytoplankton abundance.

The reflectance-based reconstruction of Cleland Lake productivity shows three anomalous periods from the deglaciation through the transition into the Holocene (Figure 2.5).

The first stage can be characterized as a period of low productivity that occurred during local deglaciation from ~14000 to ~11800 cal yr BP. Low abundances of diatoms and dinoflagellates, and higher amounts of cyanobacteria are recorded during this period. Increased runoff from glacial melt water might have contributed to higher water levels in the lake at this time. A higher percentage of mineral matter content obtained from LOI data and higher illite input suggested from PC 1 supports greater runoff-related mineral input to the lake derived from the local watershed catchment (Figure 5.13). During this period, the lake was young with limited productivity and in its early successional stages (Sze, 1993). During the second period (between

11600 to 8600 cal yr BP), cyanobacteria and dinoflagellate abundance peaked, likely in response to the maxima in incoming solar radiation. Lesser precipitation amounts and higher temperatures at this time likely limited nutrient input and produced lower lake levels, which favored the growth of cyanobacteria over diatoms. Higher temperatures led to stratified lake conditions, possibly leading to further nutrient depletion, and thereby reducing diatom abundance. Reduced runoff to the lake is also suggested by a rapid drop in illite concentration between 11750 and

11150 cal yr BP. Illite levels remain low between 11100 and 8600 cal yr BP simultaneous with the dry period.

The third productivity period starts around 8500 cal yr BP with a rapid drop in dinoflagellate levels suggesting a transition to greater precipitation amounts and an increase in nutrient fluxes, which produced higher diatom concentrations and an increase in lake-level. The

synchronous increase in illite (8500±200 cal yr BP) further supports an increase in runoff to the lake. After this rapid transition from a warm, dry early Holocene to a wet middle to late

Holocene, several decadal- to century-scale changes occurred around 7500, 6400, 3250, 2450,

1550 and 800 cal yr BP. Moreover, the transition from the deglaciation to early Holocene low stand, followed by the shift back to higher water levels implies that large P/E changes likely occurred over the span of several centuries (or perhaps less time) (e.g. 11850 to 11600 and from

8600 to 8500 cal yr BP).

The presence of an early Holocene dry period is also supported by sedimentological analysis of the intermediate core. Finely laminated organic-rich sediment and higher organic matter percentages at the top of the core (between 30 and 207 cm) indicate the occurrence of a high stand during the middle and early Holocene. Sediment composition transitions from being poorly laminated to massive and carbonate-rich between 227 and 290 cm. Corresponding with these depths, LOI-derived calcium carbonate percentage and XRF derived Ca concentrations rapidly rise between 12800 and 11400 (from 16% to 40% and 16100 to 117100 cps) and remain high until 8700 cal yr BP (Figure 5.14). This sedimentological evidence supports a low stand between 11400 and 8700 cal yr BP.

Figure 2.4. Scatter plot of Dinoflagellate algae (DFA, PC 2) versus Diatoms and Cyanobacteria (cyanobacteria, PC 4) for Cores B-09, E-09 and F-09. Fields of color-coded points are enclosed in circles corresponding to water depth. Green diamonds: Core E-09 shallow water (7 m); gray triangles: Core B-09 deep water (25m); blue dots: Core F-09 intermediate water depth (12.8 m) from 8750 to 150 cal yr BP. Red triangles or orange squares: core F09 intermediate water depth (12.8m) from 8750 to 11750 cal yr BP, or 11750 to 14300 cal yr BP respectively. Large symbol in each cluster represent the mean PC 2 and PC 4 values for each of the core/water depth and time intervals.

Table 2.3 Average values of PC 2 and PC 4 for all three sediment cores.

8800 to 11800 cal yr 11800 to 14000 170 to 8800 cal yr BP Water depth BP (prior to 12000) cal Core (m) yr BP

PC 2 PC 4 PC 2 PC 4 PC 2 PC 4

B-09 20.2 deep -0.79 0.014

12.8 F-09 -0.27 1.14 2.32 -0.43 -0.77 -0.16 intermediate

E-09 7 shallow 0.29 -0.43

Regional Comparison

Additional support for the reflectance-based, paleohydrological and paleoproductivity reconstruction at Cleland Lake can be found in the existing paleoclimate literature from British

Columbia. During the late glacial and early Holocene, rapid and significant vegetational and climatic fluctuations occurred in southern British Columbia (Pellatt et al., 2002). Deglaciation, alpine glacial advances, Younger Dryas cooling, and rapid early Holocene temperature changes occurred over this period (Pellatt et al., 2002). Palynological and paleoenvironmental climatic reconstructions suggest the Holocene climate of southern British Columbia can be divided into three distinct zones consisting of a cold and dry deglacial period, early Holocene dry period, and a cold, wet middle to late Holocene (Alley, 1976; Mathews and Heusser, 1981; Pellatt et al.,

2002; Walker and Pellatt, 2003; Hallet and Hills, 2006; Galloway et al., 2011; Gavin et al.,

2011). In Cleland Lake sediment, the same Holocene climatic pattern appears to be expressed as variations in pigment concentrations that occurred at similar times.

Figure 10 Figure 2.5. Down core variation of PC 2, PC 4, Si:Ti ratio from core F-09 compared against summer insolation at 60°N and three independent paleoclimatic reconstructions from southern British Columbia. A. Age model control points of the Cleland lake F-09 core, (filled triangles) 14C and tephra ages, (open triangles) extrapolated age. B. Northern hemisphere June incoming solar radiation at 600N (Berger and Loutre 1991). C. Dinoflagellate algae abundance derived from diffuse spectral reflectance PC 2. D. Diatom and Cyanobacteria (blue green algae) abundances derived from diffuse spectral reflectance PC 4. E. Si:Ti ratios for the F-09 core obtained from XRF data is a proxy of biogenic silica and accompany the diatom abundances derived from PC 4. Note that the y-axis of Si:Ti plot is represented in a log scale to enhance the visibility of data. Periods of highest dinoflagellate and cyanobacteria abundance are consistent with the peak summer insolation (gray shaded box). E. Si:Ti ratios for the F-09 core, obtained from XRF data is a proxy of biogenic silica and accompany the diatom abundances derived from PC4. F. Annual precipitation anomaly from central boreal Canada (Viau, A.E. and K. Gajewski. 2009) G. Percent biogenic silica record from Eleanor Lake (Gavin et al., 2011). H. Combined chironomid-inferred summer temperature anomalies from four sites; Frozen Lake (Rosenberg et al., 2004), North Crater Lake and Lake-of-the-Woods (Palmer et al., 2002), and Windy Lake (Chase et al., 2008) (after Gavin et al., 2011).

Deglacial Period

Fossil midge-based temperature reconstructions and several pollen analyses from locations around Cleland Lake suggest that the deglacial period in southern British Columbia was 1 to 2 °C colder than present (Matthews and Heusser, 1980, Chase et al., 2008). The cold, moist deglacial climate between 12000 and 10500 14C yr BP (recalibration using CLAM 1.0.2 yields ages of 14000 and 12150 cal yr BP; Mathews and Heusser, 1981) was followed by a rapid decrease in precipitation and increase in temperature from 10500 to 10000 14C years BP

(recalibrated ages 12150 to 11500 cal yr BP; Alley, 1976; Mathews and Heusser, 1981; Chase et al., 2008). This timing is synchronous with the rapid onset of drier conditions at Cleland Lake

(11800 to 11700 cal yr BP).

Early Holocene

Several modeling and pollen-based temperature and precipitation reconstructions throughout southern British Columbia suggest summer temperatures were 5 to 10°C warmer than

present and that precipitation was about 30% lower than present during the early Holocene, consistent with the low stand inferred from pigments at Cleland Lake (Mathews and Heusser,

1981; Bartlein et al., 1998; Chase et al., 2008). This rapid transition to a warm early Holocene is also suggested by a midge-based reconstruction (Chase et al., 2008). Other studies indicate that summer temperatures from 10500 to 9000 cal yr BP were likely 3 to 4°C warmer than present

(Chase et al., 2008). At Big Lake, 30 km east of Cleland Lake, a lake level reconstruction based on Charophyte accumulation rates indicates lower water level from 10300 to 8500 cal yr BP

(Hallet and Hills, 2006). Sediments from Eleanor Lake (250 km northwest of Cleland Lake) dating to 10200 cal yr BP, contain abrupt increases in biogenic silica, indicating that an arid climatic event occurred at this time (Gavin et al., 2011, Figure 2.5). Pollen-based, regional climatic reconstructions from Canada reveal a warm, dry early Holocene with higher precipitation (relative to overall dryness during this time) between 12000 and 10000 cal yr BP in the central Canadian region (50–70°N, 80–120°W, Viau, and Gajewski. 2009, Figure 2.5). In addition, elevated percentages of xerohpyte plant taxa from the same location (Pinus juniperus,

Artemmisia, Poaceae) provide evidence that a drier climate existed prior to 8500 cal yr BP

(Hallet and Hills, 2006). A much longer warm interval between 10000 to 4900 cal yr BP was recorded by sediments from Mahoney Lake; located in the semi-arid Okanagan valley south west of Cleland Lake (Lowe et al., 1997). The presence of tree stumps at elevations higher than the present tree line in British Columbia, further suggests that temperatures were warmer at this time

(Chase et al., 2008 and references therein). In general, these studies and others (e.g. Matthews and Heusser, 1980) indicate that the early Holocene was warmer and drier than present, which is supported by the spectral reflectance-based reconstructions at Cleland Lake.

End of the Early Holocene Drought

The rapid drop in dinoflagellate values in Cleland Lake around 8600 yr BP suggests a return to wetter conditions. This observation is supported by the arrival of Pseudotsuga/Larix pollen to the Kootenay valley (Hallet and Hills, 2006). The sharp decrease in fire frequency around Dog Lake, southern British Columbia (32 km west of Cleland Lake) after 9000 cal yr BP could also be interpreted as an increase in effective moisture, although many other factors may have contributed to this change (Hallet and Hills, 2006). Rapid increases in summer moisture and lower temperatures between 9300 and 8500 cal yr BP are suggested by a reduction of biogenic silica abundances at Eleanor Lake (Gavin et al., 2011, Fig. 4). At Cleland Lake, this wet period is concurrent with an increase in diatom abundances as inferred from pigments and Si/Ti ratios. .

Although paleorecords from the region are not entirely consistent, some of the differences can be explained by disparity in lake hydrologic setting, intra-regional climate heterogeneity, and seasonal partialities of different paleoclimate proxies. In general, the timing of the reflectance-based reconstructions from Cleland Lake broadly corresponds with other climate reconstructions from the region.

Possible Mechanism for the Late Quaternary Drought

The prolonged period of heightened dryness in southern British Columbia during the late

Pleistocene-early Holocene was likely driven by nonlinear feedbacks related to insolation-related amplification of the subtropical, North Pacific high-pressure system (Knapp et al., 2004; Bartlein et al., 1998). Modeling results from Bartlein et al., 1998 suggest that the Pacific Northwest surface-air temperature is largely controlled by air temperature in the upper troposphere.

Simulations of sea level pressure suggests that an ~8% increase in northern hemisphere summer insolation resulted in a strengthening of the eastern Pacific and Bermuda sub-tropical high-

pressure systems (Bartlein et al., 1998) (Figure 6.1). The development of such a (blocking) pressure system would have reduced moist westerly air entering the interior regions of the

Pacific Northwest (Stahl et al., 2006). The spatial and temporal expression of this period of dryness was likely highly variable, in part due to large geographic disparities (i.e. elevation, topography, climate) and numerous microclimates that exist across the larger Pacific Northwest region (Shinker and Bartlein, 2010). We propose that insolation-driven changes in air temperature moderated moisture delivery to Cleland Lake by altering the configuration of the subtropical high-pressure systems (Bartlein et al., 1998).

A relatively warm climate during the middle Holocene thermal maximum (HTM) is recorded by various paleoclimate proxies from the middle and high-latitudes of the Northern

Hemisphere (Viau et al., 2003; Renssen et al., 2009). The timing and the magnitude of this event varied substantially between regions (Kaufman et al., 2004; Renssen et al., 2009 and references therein). For example, in Alaska and northwestern Canada the HTM occurred from 11000 to

9000 cal yr BP (Kaufman et al., 2004; Renssen et al., 2009 and references therein), where as in north central Canada the warming occurred between 8000 and 5000 cal yr BP (Kaufman et al.,

2004 and references therein). In contrast in Northwestern America it occurred between 8000 and

5000 cal yr BP (Stone and Fritz 2006; Whitlock et al., 2012). The HTM in Northwestern Canada was apparently simultaneous with the Holocene summer insolation maximum of the Northern

Hemisphere; whereas the delayed warming in central and eastern Canada (longitudinal variability) likely resulted from atmospheric and ocean circulation changes associated with the residual Laurentide Ice Sheet (LIS).

The Cleland Lake sediment core data suggests a pattern of climate change in

Southwestern British Columbia during the Pleistocene-Holocene transition similar to that of

other paleoclimate records from the region. The abrupt changes (possibly over several centuries or less) that marked the beginning and end of the HTM are likely the result of an insolation- driven, non-linear response as the climate system passes through a threshold resulting from a positive feedback between vegetation, and the land surface, ocean and atmosphere. For example, climate models simulating atmospheric conditions have demonstrated that a 40% increase in

North African rainfall can result from an 8% increase in summer radiation (DeMenocal et al.,

2000). Paleoclimate data from western North America as well as other parts of the world provide evidence of prolonged droughts in the early-middle Holocene that lasted for centuries to millennia (Overpeck and Cole, 2006 and references therein). These prolonged and severe droughts resulted from non-linear responses of the climate system to radiative forcing (Overpeck and Cole, 2006 and references therein; Viau et al., 2003; Clegg et al., 2011) and therefore hint at the possibility that similar climate responses could occur in the future in response to anthropogenic, greenhouse gas forcing. The combined effect of a warming ocean and a positive feedback from reduced soil moisture in an anthropogenically warming world could enhance the probability of prolonged droughts (Overpeck and Cole, 2006), although the details of the response may depend on the interaction of greenhouse forcing and solar insolation.

Conclusions

Cleland Lake is a small, alkaline, closed-basin lake that is hydrologically and geochemically sensitive to drought. Phytoplankton communities are influenced by lake responses to hydroclimate variability; therefore we used VDS to measure fossil phytoplankton pigments in

Cleland Lake sediment cores to infer algal abundance and distribution, and to thereby reconstruct lake productivity and hydrologic balance throughout the late Quaternary. Our analysis suggests

that lake productivity was low and mostly limited to early successional, cyanobacteria communities during the late Pleistocene deglaciation (prior to 11600 cal yr BP). Dinoflagellates rapidly increased after 11800 and remain high between 11600 and 8600 cal yr BP, indicating warm, low lake-level conditions. Low diatom and high cyanobacteria concentrations between

11500 and 9500 cal yr BP further indicate a warm, shallow lake at this time. The magnitude of early Holocene lake level change was likely greater than that of the middle and late-Holocene.

Dinoflagellate algae pigments decrease rapidly after 8600 cal yr BP, suggesting a transition to a wetter climate and higher lake levels. This interpretation is further strengthened by illite input to the lake as represented by VDS PC 1, sedimentological evidence from the comparison of PCs between cores from different water depths in the lake as well as comparison with other paleoclimate records from the region, which show similar patterns.

In summary, lake level changes inferred from variations in the concentration of dinoflagellate algae and cyanobacteria in Cleland Lake sediment cores indicate that lower lake levels were coincident with the Northern Hemisphere summer insolation maximum, which occurred between ~11500 to 8700 Cal year BP at this latitude. Lake levels increased after 8600 cal yr BP and remained relatively higher until present, implying wetter conditions more consistent with the modern climate of the region. The abrupt shift to low lake levels and relatively dry conditions that characterized the late Pleistocene-Holocene transition (and the corresponding rapid return to wetter conditions thereafter) were likely driven by non-linear climate responses to orbital forcing, and a resulting amplification of the subtropical, North

Pacific high-pressure system, which reduced moist westerly air flow from the Pacific Basin.

CHAPTER 3

MODERN HYDROLOGY IN THE PACIFIC NORTWEST DURING THE MID-HOLOCENE

Abstract

In small closed-basin lakes in semi-arid regions, variations in precipitation/evaporation (P/E) balance affect the physical, biological, and chemical composition of the lake water and sediment.

This study presents visible derivative spectroscopy, XRF derived elemental concentrations and

18 18 δ O values of carbonates (δ Ocarb) in sediment cores from Cleland Lake, British Columbia. The data provide insight into paleolimnological variations during the past 7500 years, based on a 2.5 m long sediment core obtained from the deep lake basin (core B09, water depth of 24m).

18 18 δ Ocarb of closed basin lakes are sensitive to P/E balance, giving rise to enriched δ Ocarb values during arid periods and depleted values during wet events. Principal Component (PC) 1 of the reflectance data, i.e. Illite+sphalerite is used as a clay mineral proxy, and PC 4, diatoms+ cyanobacteria, is used as a paleoproductivity proxy. Diatom abundances have a positive correlation with illite (r= 0.79, n=73 α=0.01) throughout the record. In contrast, negative

18 correlations exist between δ Ocarb and diatom abundances, indicating higher diatom abundance

18 during wet periods. Variability in diatom abundances is greater than that of the δ Ocarb values, indicating that factors other than the P/E balance affects phytoplankton abundance. Late

Holocene is characterized by high frequency moisture variability with three rapid increases in

56 diatom abundances centered at 2600 cal yr BP, 1200 cal yr BP, 950 cal yr BP, and 650 cal yr BP and three periods (centered at 2500 cal yr BP, 2100 cal yr BP, and 1400 cal yr BP) of predominantly low diatom abundance. Increase in diatom abundances are associated with

18 depleted δ Ocarb values, suggesting a groundwater associated inflow of nutrients to the lake. The largest and most abrupt event during the mid and late Holocene is marked by a 5 ‰ depletion in

18 δ Ocarb around 2600 cal yr BP. Prominent transitions in lake sediment chemistry are synchronized with this event. Concentrations of Ca, Fe, Ti and Mn transitions from high to low levels while, concentrations of Cu, Ni, and Cr transitions from low to high levels. These transitions in trace metal concentrations would indicate a transition from a shallow lake that undergo frequent lake level changes and seasonal mixing (7500 to 2600 cal yr BP) to a well stratified deep lake after 2600 cal yr BP. Late Holocene Ni and Cu enrichment could be a result of intense organic matter decomposition under reducing conditions. Most recent advancement of the lake levels is simultaneous with the Medieval Climatic Anomaly, which could be associated with the prevailing warmer central and western Pacific ocean SST giving rise to La Niña-like conditions and a warm (positive) phase of PDO (Cobb et al., 2003; Mann et al., 2009; Steinman et al., 2014).

Introduction

Isotopic data from hydrologically sensitive lakes can be used as proxies for paleo- moisture regime (Hammarlund et al., 2003). High resolution moisture records up to decadal to multi-decadal scale can be obtained from well preserved annually varved lake sediment, based on the hydrological budget of the lake under study. δ13C ratio of lacustrine inorganic carbon depends mainly on the isotopic composition of inflowing water, CO2 exchange between the

atmosphere and the water column, total dissolved inorganic carbon (TDIC) of water and the biological activity (photosynthesis) of the lake (Leng and Marshall, 2004). Increased fresh water input during wet climate produces higher lake levels, increase in outflow of the lake, or open lake conditions leading to low salinity (low specific conductivity) (Li et al., 2008). Such changes would give rise to depleted δ18O and δ13C (Li et al., 2008).

In contrast, relatively dry periods dominated by decreased precipitation/evaporation (P/E balance) and low lake levels will give rise to enriched δ18O and δ13C due to the preferential evaporation of the light isotope (16O and 12C respectively) (Hammarlund et al., 2003; Leng et al.,

2005; Li et al., 2008). Previous research in the area suggests that closed basin lakes around

Canadian Arctic and Pacific Northwest (including Cleland Lake) have a higher degree of precipitation than groundwater recharge (Anderson et al., 2005). δ18O and δ1D of the surface water samples of these closed basin lakes lie offset of the Global meteoric water line (GMWL), but plot on the local evaporative line (LEL), representing their higher sensitivity to regional droughts and precipitation (Anderson et al., 2005, figure 3.1). In hydrologically closed basin lakes δ18O and δ13C covary indicating that lake δ13C and δ18O are functions of climate change (Li et al., 2008). High δ13C values indicate different degrees of equilibration of total dissolved inorganic carbon (TDIC) with atmospheric CO2. A strong rise in lake levels are reflected by a rapid decrease in δ18O (>1‰ change over 50 years) while a rapid enrichment in δ18O (>1‰ change over 50 years) indicate a strong drop in lake levels (Li and Ku, 1996). Therefore this study proposes to reconstruct the Holocene paleoclimatic variations in the Pacific Northwest, using stable C and O isotopes of lake sediment from Cleland Lake, British Columbia (BC). We supplement the isotopic data with additional paleoclimate proxies (trace metal concentrations and visible derivative spectroscopy) to place additional constraints on the isotope data.

D ‰ VSMOW ‰ D 

Figure 3.1. Oxygen and hydrogen isotope ratios of surface water and precipitation samples from the closed basin lakes of Southern Yukon Territory. LEL is the local evaporation line, a linear regression of the lake data (Anderson et al., 2005).

Trace metals can be dissolved in lake water or as adsorbed on to particles (Tribovillard et al., 2006). They can be removed from the water column by biotic or abiotic processes

(Tribovillard et al., 2006). Biotic processes are prominent in oxic environments and include

absorption by primary producers during photosynthesis (Tribovillard et al., 2006). Abiotic processes dominate in suboxic and anoxic environments. In oxygenated water Mn is oxidized to insoluble Mn+2 or Mn+4 (Tribovillard et al., 2006). As a result Mn is enriched in oxidized sediments (Brown et al., 2000; Naeher et al., 2013). Mn can become mobilized under reducing conditions and can be available in the water column in its soluble Mn-2 form resulting sediments depleted in Mn (Tribovillard et al., 2006). Mn cycling is also associated with other trace elements like Ni, Cu, Zn, Co,V, Pb and Cr. Mn involves in transferring these trace metals from the water column to the sediment by trapping them to Mn oxyhydroxide (Tribovillard et al.,

2006). In addition to that Ni and Cu are delivered to the sediment by organic matter decay. Here we assume that higher Mn concentrations indicate a shallow lake where mixing of the water column results oxidized sediments, enriched in Mn. In contrast a deep lake permits permanent stratification in the lowe-water column giving rise to anoxic bottom water depleted in Mn. In addition, higher Ni and Cu levels indicate higher organic matter flux and organic matter (OM) decomposition under reducing conditions, while low Ni and Cu levels may indicate anoxic conditions developed due to restricted water mass renewal or stagnation (Tribovillard et al.,

2006).

Methods

Geochronology

Sample preparation for carbon dating was done at University of Pittsburgh, prior to being analyzed at the W.M. Keck Carbon Cycle Accelerator Mass Spectrometry Laboratory at the

University of California, Irvine (Abbott and Stafford, 1996). Bulk sediment samples were digested in 7 % H2O2 for nearly 12 hours and sieved at 63 µm. Terrestrial macrofossils for

radiocarbon dating were then handpicked under a light microscope. Samples were treated with acid-base-acid wash and rinsed with deionized water for pH neutrality. The B-09 core chronology is based on eleven AMS 14C dates of terrestrial macrofossils and two tephra layers

(table 4.1). A linearly-interpolated age model was prepared by point-to-point calibration using the CLAM age modeling software (Blaauw, 2010).

Visible Derivative Spectroscopy

Visible derivative spectroscopy (VDS) is used to define sediment pigment content, based on the shape of the diffuse spectral reflectance (DSR) of components extracted from the derivative spectrum for each sample. Diffuse spectral reflectance was measured at 0.5 cm intervals on the wet, split core sediment surface of core F-09 using a handheld Minolta CM-

2600d spectrophotometer. This instrument measures the reflectance over the visible range of electromagnetic spectrum (400-700 nm) at a 10 nm resolution.

Principal Component Analysis

Varimax-rotated, principal component analysis (VPCA) of the correlation matrix derived from the center-weighted derivative of the DSR data extracts orthogonal components that can be related to sediment composition (Ortiz et al., 1999; Peteet et al., 2003; Ortiz et al., 2009). The derivative transformation minimizes the scattering effects, which influence the raw reflectance spectra (Ortiz et al., 1999; Ortiz et al., 2009). VPCA was conducted in SPSS™ 14.0 using the full data set (1751 reflectance measurements) based on all four cores raised from Cleland Lake.

This paper is based on the data from the deep basin core (B09).The components were identified by comparing the spectral signatures from each PC with known pigment and mineral standards

(Gantt, 1975; Robertson et al.,1999; Graham and Wilcox, 2000; Schagerl et al., 2003; Schagerl

and Donabaum, 2003; Toepel et al, 2005 (Figure 2.3)). This process is similar to the spectral matching method used in XRD or previously applied to marine cores (Ortiz et al., 2009; Ortiz

2011).

Carbon and Oxygen Isotope Analysis

Carbon and oxygen isotope analyses were performed at 1-2 mm resolution on the bulk sediment samples of the B-09 core at the University of Pittsburgh. Sediment were first disaggregated in 7% hydrogen peroxide overnight at room temperature. This process also removes organic matter in the sediment. Then the fine fractions (<63 m) were recovered by sieving through a 63 µm sieve. The recovered fines are then bleached in diluted NaClO (bleach) to oxidize the organic matter for 6-8 hours. Finally the sediment were centrifuged and rinsed in de-ionized water three times. Stable isotope measurements were performed at the University of

Pittsburgh using the Finnigan Delta XL mass ratio spectrometer.

Isotopic ratios of the samples are measured as the ratio of a heavy isotope to a light isotope and compared to a universally accepted standard (Vienna Pee Dee Belemnite standard;

VPDB). Values are expressed in parts per thousand (‰), which is more useful than parts per hundred when ratios are small. This comparison can be expressed as follows for carbon

( 18 O/16 O) - ( 18 O/16 O) d 18 O = sample standard ´1000 (in ‰) ( 18 O/16 O) standard

18 16 18 16 18 16 ( O/ O)sample is the isotopic ratio of O to O in the sample and ( O/ O)standard is the oxygen isotope ratio in a reference calcite belemnite fossil from the Pee Dee Formation, South Carolina. Therefore, values are reported as ‰ Pee Dee belemnite (PDB) (Palmer, 2002).

X-ray Fluorescence

The bulk elemental compositions of core samples were measured using the ITRAX X- ray Fluorescence core scanner at the University of Minnesota, Duluth. The scanner has the potential to scan cores at 0.2 mm resolution. Cleland F-09 and E-09 cores were scanned at 5 mm resolution, while the core B-09 was scanned at 0.4 mm resolution. Scanning at a coarser resolution allows a longer dwell time, and thus, a higher signal to noise ratio. The scanner, which is powered by a molybdenum X-ray source, was operated at 30 mA and 45 keV over a scan-time of 60 seconds. The XRF core scanner provides the relative abundance of bulk elements, in terms of counts per second (cps). PCA was performed on stacked XRF data from the B-09, E-09, and

F-09 cores.

Results

Present Isotopic Properties of the Lake Water

Modern 18O and 2D of Cleland lake deviates from the general trend of the local meteoric water line of Calgary and Alberta (LMWL) and lie offset to the global meteoric water line (GMWL, Figure 3.2). Enriched water isotope values represent that the lake is sensitive to the evaporation.

-50

-100

-150 Cleland Lake

GMWL

H (‰ VSMPW) (‰ H 2

 -200 LinearLMWL (EDMONTON INDUSTRIAL (ALBERTA)) LEL -250

-300 -35 -30 -25 -20 -15 -10 -5 0 18O (‰VSMOW)

Figure 3.2. Oxygen and hydrogen isotope ratios of surface water and precipitation samples from Western Canada. LEL is the local evaporation line, a linear regression of the Cleland Lake water isotope data (red filled diamonds). LMWL is the local meteoric water line; a linear regression of the isotope data from the precipitation of Calgary and Edmonton, Canada (open diamonds) (Data source IAEA/WMO, 2014). GMWL is the global meteoric water line; a linear regression of the global precipitation data, defined by Craig, 1961.

Age model

The core B-09 age model extends from 470 to 7697 cal yr BP. The age model is well constrained with a total of 13 age control points, giving an average of one age control point per each 16 cm (Figure 3.3). The thickness of the two sigma error envelope is in the range of 50 cal yr BP to 150 cal yr BP (Table 3.1).

Table 3.1. AMS Radiocarbon and tephra dates for the Cleland lake B-09 core. *Omitted radiocarbon ages

Depth Material 14C yr before Median 2σ lower 2σ lower (cm) present (BP) calibrated age (cal yr BP) (cal yr BP) (cal yr BP) 44 tephra n/a 470 468 471 54 718 640 795 65 1055 921 1189 73.5 leaf 1610 1492 1355 1629 95 insect 2090 2060 1994 2127 115 pine needle 2270 2320 2301 2346 153 wood 3195 3410 3369 3455 182.5 bark 3945 4390 4340 4444 208.5 charcoal 5085 5830 5732 5920 220.9 charcoal 5575 6355 6311 6399 244 charcoal 6290 7215 7170 7259 245 wood 6420 7350 7268 7425 249.5 tephra 6850 7700 7606 7788 38* pine needle 545 550 530 553 215* seed casing 5565 6355 6308 6364

230

180

Linear interpolated age 130 C-14

Depth (cm) Depth tephra Omitted 80 Min 95% CI Max 95% CI 30 -1000 1000 3000 5000 7000 9000 Age Cal years B.P. Figure 3.3. Age model of the Cleland C-09 and B-09 cores. C-09 age model consist of 13 210Pb dates from 1 to 13.5 cm. Core B-09 consist of 11 14C dates and two tephra layers between 30 cm and 250 cm.

Visible Derivative Spectroscopy and Principal Component Analysis

The first five principal components (PC) of the reflectance data explain 97% of the variance in the reflectance data set (Figure 2.3). PC 1 correlates with the combined reflectance spectrum of illite and sphalerite (r= 0.91, n=31, α = 0.001), while PC 2 correlates (r=0.87, n=31,

α = 0.001) with dinoflagellate algae (DFA) pigments (peridinin, Pheophytin-A,) and goethite

(Figure 2.3). Principal component 3 correlates (r= -0.92) with montmorillonite and cyanobacteria pigments (phycocyanin+chlorophyllide-b). Principal component 4 correlates (r= 0.91) with a mixture of bacillariophycea and phycocyanin (i.e. diatoms and cyanobacteria photosynthetic pigments, Figure 2.3). Principal component 5 correlates with allophycocyanin (accessory pigment in cyanobacteria) and pheophytin-a (Figure 2.3.). This chapter further discuss the down core variation of PC 1 and PC 4 in the core B-09. PC 1 is considered as a proxy for runoff and

sediment input, where higher PC 1 score representing higher runoff and nutrient input to the lake. PC 4; a proxy for diatom and cyanobacteria abundance is also considered as a proxy for lake level. Increases in lake level are associated with higher diatom abundance in sediment

(Barker et al., 1994; Brugam et al., 1998; Heinsalu et al., 2008; Gavin et al., 2011), therefore higher PC 4 scores indicate increased diatom abundance and high lake levels.

PC 1 of the deep B-09 core has a positive linear correlation with PC 4 (r= 0.77, n=143,

α=0.01) throughout the record (Figure 3.4). Values of PC 1 vary between high and low values of

1.3 and -1.5 respectively with an average of -0.62. Average PC 1 score (-0.69) remain close to the average of the whole record, from 7500 to 5000 cal yr BP (Figure 3.4). PC 4 scores vary between 1.5 and -0.75 with prominent peaks and troughs between 7100 to 6800 cal yr BP and

6250 to 5900 cal yr BP, respectively. Prominent enrichments in PC 1 scores occur centered on

6950, 6400 and 5800 cal yr BP respectively. PC 1 demonstrate comparatively lower values during mid Holocene, (average -0.79) from 5000 to 2500 cal yr BP, except briefly at 4450, 4100,

3450, 3100 and 2600 cal yr BP, where high PC 1 scores are recorded. PC 4 scores do not show dramatic fluctuations between 5000 to 2500 cal yr BP, except around 4100 cal yr BP, when high

PC 4 scores are recorded. Both PC 1 and PC 4 scores rapidly increase between 2750 and 2550 cal yr BP and rapidly drops after that. Both the PCs gradually rise between 2200 and 1200 cal yr

BP. PC 1 scores vary between ~0.71.2 between 1250 and 400 cal yr BP, with prominent peaks centered around 1100, 950 and 600 cal yr BP. PC 4 values gradually increase between 1050 and

400 cal yr BP (Figure 3.6).

δ18O and δ13C show a linear correlation (r=0.7, n= 143, α=0.01, Figure 3.5) throughout the record. Values of δ18O vary between -2.7 ‰ and -12.9 ‰ and remain close to the average (-

6.2 ‰), except during 6650, 2500 and 650 cal yr BP (Figure 3.5). Relatively depleted δ18O

values (-9.8 ‰ and -12.9‰) are recorded during two former events while relatively enriched values are recorded during the latter event. δ13C record covary with the δ18O record showing similar fluctuations with an average value of 3.4‰ (Figure 3.5).

2.5 High Diatom

r² = 0.6113

1.5

0.5 Phycocyanin) -0.5

PC 4 (70% Bacillariophycea+ 30% Bacillariophycea+ (70% 4 PC High Illite -1.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 PC 1 (95% Illite+ 5 % Sphalerite )

Figure 3.4. Scatter plot between PC 1 versus PC 4. Symbols color coded for three different time periods, purple diamonds; 7500 to 5000 cal yr BP, green triangles; 5000 to 2500 cal yr BP, red squares; 2500 to 400 cal yr BP;. Black line; linear trend between PC1 and PC 4 indicating higher diatom abundance when the illite (clay) input to the lake is high.

C (VPBD) C 2

13 δ R² = 0.549 0

-2 -10 -8 -6 -4 -2 0

18 δ O (SMOW)

Figure 3.5. Plot of 18O versus 13C from the Cleland Lake sediments

Elemental Concentrations

Concentrations of more than 20 elements were measured using the ITRAX core scanner.

However, only the 11 that exhibit considerable variation are explained in this chapter (Figure

3.7). P, S and, Si concentrations remain high between 7500 and 2500 cal yr BP with century scale oscillations. Concentrations of P,S, and Si rapidly decrease between 2400 and 2000 cal yr

BP and remain low till 1200 cal yr BP. The levels rapidly rise to higher concentrations after that and remain high during the late Holocene (1100 to 400 cal yr BP; Figure 3.7). Highest Holocene

Si, P, and S levels are recorded between 800 and 900 cal yr BP. Ti and Fe levels are relatively high between 7200 and 2500 cal yr BP, but rapidly drops to lowest Holocene levels after 2500 cal yr BP and remain low till 400 cal yr BP. Concentration of Cu, Ni, and Cr remain low or below detectable levels prior to 2500 cal yr BP. Concentrations gradually increases after 2000

cal yr BP and reach highest levels around 800 cal yr BP. Ni and Cu concentrations remain high only for several years (2600 to 1800 cal yr BP) and drop back to low levels. Ca levels remain high between 3000 and 7500 cal yr BP and drop to low levels around 2800 cal yr BP. Mn concentration is low (below 200 cps) during early-Holocene, gradually increases after 6500 cal yr BP and remains high (400 to 600 cps) from 5000 to 2500 cal yr BP. Mn concentration rapidly drops between 2850 and 2500 cal yr BP and remains low after 2500 cal yr BP.

In summary, P and S have relatively high concentrations between 7500 and 2500 cal yr

BP, abruptly decrease at 2500 cal yr BP and rapidly rise to highest levels after 1200 cal yr BP.

In contrast, Cr, Cu, Ni, and Zn levels are low between 7500 and 2500 cal yr BP and rapidly increase around 2700 cal yr BP. Fe and Ti concentrations are highest from 7500 to 7400 cal yr

BP, but remain low from 5000 to 400 cal yr BP. Si concentrations rapidly drop by 2700 cal yr

BP, remain low till 900 cal yr BP and rapidly increase after that. Ca concentrations are higher between 7500 and 2500 cal yr BP and gradually drops to lowest levels after 2500 cal yr BP.

PC 1 PC 4 d18O LOI 550°C Mineral matter %CaCO3 (illite) (diatoms+BGA) 2 1 0 -1 -2 2.5 1.0 -1.0 -14 -10 -6 -280 55 30 50 25 0 50 25 0 250

1250

2250

(Cal yr BP)(Cal 3250 Age 4250

5250

6250

7250 High low High low Wet dry max.: 7550 Cal yr BPIllite Diatoms

Figure 3.6. Down core variation of reflectance PC 1 (illite+sphalerite), reflectance PC 4:diatom+ cyanobacteria pigments (BGA), δ18O of carbonates. Black vertical bars represent Holocene glacial advances in Western Canada (Menouns et al., 2009).

Si P S Ca Ti Cr Mn Fe Ni Cu Zn

0 200 400 0 150 300 0 150 300 0 15000 30000 0 500 1000 0 100 200 0 500 1000 0 10000 20000 0 100 200 0 100 200 0 200 500 0

1000

2000

3000

(cal yr BP) yr (cal Age

4000

5000

6000

7000

max.: 7658.1 cal yr BP Figure 3.7. Variation of major and trace metal concentrations between 7500 to 400 cal yr BP, with the generalized stratigraphic column of the deep core (B-09). Concentrations are in counts per seconds. Trace metals (Cr, Ni, Cu, Zn) rapidly increases while Ti, Fe, Si, Mn and Ca rapidly drops after 2500 cal yr BP. Black bars represent Holocene glacial advances in Western Canada (Menouns et al., 2009). Note that the scale of the X-axes are different from one element to the other for the clarity of visualization.

Discussion

PC 4 covaries with PC1 throughout the record (7500 to 400 cal yr BP), indicating a positive correlation between diatom abundance and illite concentration in Cleland Lake

(Figure 3.4). Time series of PC 1 and PC 4 also share a similar trend with Si, P, and S. This suggest that the diatom levels are associated with the availability/abundance of the major nutrients (P,S and Si).

Cleland Lake δ13C records covary with the δ18O, with an r value of 0.7 (Figure 3.5). This suggests that the lake remained hydrologically closed throughout the record (7500 to 400 cal yr

BP; Leng et al., 2005). δ18O values remained stable between 7500 to 400 cal yr BP, suggesting stable lake levels except around 6650, 2500 and 650 cal yr BP. Rapid drop in δ18O (-5.3 to -9.8

‰ ) between 6700 and 6650 cal yr BP suggest a rapid rise in lake levels or a ground water input to the lake. Simultaneous with this rise diatom abundance and illlite (derived from PC 1) as well as mineral matter content (from LOI) increase but the organic matter content and percent CaCO3 content decrease. This suggests higher detrital input to the lake associated with enhanced ground water inflow. Synchronous increase in illite concentration (derived from PC 1) further supports enhanced mineral matter supply to the lake. Concentrations of P, S, and Si remain high during this period (6600 to 6250 cal yr BP). A drop in carbonate percentage further strengthens a lake level rise and development of lake anoxia where preservation of carbonate is limited.

Observed illite, diatom and lake level increases are simultaneous with five periods of neoglacial advances throughout Western Canada (7360 to 6450, 4400 to 3970, 3540 to 2770,

1710 to 1300, and 700 cal yr BP: Menounos et al., 2009). Glacial advances are directly influenced by a combination of climatic conditions including temperature, precipitation, and solar radiation (Holzhauser et al., 2005). Although the present Cleland Lake watershed is not

directly fed by a glacier (Figure 1.2), this coexistence might indicate that lake level changes and lake productivity are influenced by regional climate (Holzhauser et al., 2005). Similar lake level increases simultaneous with neoglacial advances have been reported from several independent studies from interior British Columbia (Lowe et a., 1996; Hallet and Hills, 2006; Galloway et al.,

2011).

Late-Holocene PC 1 and PC 4 scores show dramatic fluctuations indicating frequent changes of illite input to the lake and lake diatom abundances. Both illite and diatom levels rapidly increase between 2700 and 2550 cal yr BP. This rapid increase is simultaneous with the largest recorded drop in δ18O values (-6.3 to -13 ‰). This depletion in δ18O values between 2700 and 2500 cal yr BP suggest, a rapid rise in lake levels. Simultaneous with this lake level rise, concentrations of several major and trace elements show significant changes. Concentrations of

P, S, and several trace metal concentrations (Cu, Ni, Cr, Pb) rapidly increase. A similar trend is shared by the LOI-derived mineral-matter content and Ti, Fe, and As, confirming their detrital origin. This suggests higher detrital as well as clay mineral input to the lake associated with enhanced ground water inflow. Concentrations of Ca, Si, Al, As, Ti, Fe, and Mn rapidly drop after 2500 cal yr BP and remained low till 400 cal yr BP.

This anomalous depletion in δ18O could be a result of higher lake levels and an associated lake overflow. The idea of lake overflow cannot be confirmed due to the lack of present ground water flow data in the catchment area. However the lake hydrologic flow model (prepared using

Arc GIS; Figure 3.8) suggest that this lake is a ground water recharge lake (Sharpley et al., 2008) and hence it is possible that the lake overflowed into Jade Lake during periods of higher lake levels (Figure 3.8).

Rapid change in the Cleland Lake around 2500 cal yr BP also coincides with the solar irradiance minimum around 2700 cal yr BP. Several paleoclimatic reconstructions from the

Canadian Rocky Mountains suggest that solar minima are associated with glacial advances and high lake levels (Gavin et al., 2011). Also some other studies explain the relationship between solar irradiance and the climate of Pacific Northwest. Solar activity is inversely correlated with sea surface temperature (SST) of the eastern Tropical Pacific Ocean (ETP: Marchitto et al.,2010). As a result periods of low solar activity, create a warm ETP giving rise to El-Niño like conditions in the Pacific Northwest (Mann et al.,2005; Marchitto et al.,2010; Steinman et al.,

2014). However, in contrast to the observed increase in lake level at Cleand Lake, dry conditions are recorded associated with warm-El Niño-like conditions in Pacific Northwest (Steinman et al.,

2014).

A Flow lines Lakes B Watershed

D C

Figure 3.8. Cleland Lake watershed showing the flow lines. A. Cleland Lake. B. Jade Lake, C, D and E are unknown (unnamed) lakes. Arrowhead show the direction of flow.

Trace metals in core B-09 show prominent transitions in their concentrations after 2600 cal yr BP. Mn and Fe levels are high while Ni and Cu levels are relatively low prior to 2600 cal yr BP. Fe and Mn levels rapidly decrease, while Ni and Cu levels rapidly increase after 2600 cal yr BP. This transition could represent a change in redox conditions in the water column between mid and late Holocene (Naeher et al., 2013; Lenz et al., 2014). High Mn levels in sediment during mid-Holocene should represent periods of oxygenated bottom water. The lake transition to more anoxic conditions (preventing leaching of Mn from reduced sediemnt into reduced bottom water) is signaled by lower levels of Mn, during late-Holocene after 2600 cal yr BP

(Naeher et al., 2013; Lenz et al., 2014). This transition can be further interpreted in terms of lake

level changes between mid and late Holocene. Low lake levels during mid-Holocene would have promoted mixing, giving rise to oxygenated bottom water. In contrast, high water level during late Holocene would have helped to maintain a well-stratified water column with less mixing and permanent anoxia.

An alternative explanation for the increase in trace metal concentrations is prehistoric mining activities. Enrichment in Cu, Ni and Zn simultaneous with the period of rapid advancement of copper usage in the Roman empire (Hong et al., 1996). Enrichment in Cu concentration was also recorded from central Greenland Summit ice core, beginning about 2500 years ago figure (Hong et al., 1996). The origin of the pollutants are assumed to be from the main production areas, Spain, Cyprus and central Europe (Hong et al., 1996). With the increase in copper usage for military and civilian purposes, cumulative copper production was calculated as 500,000 metric tons, between 4000 and 2700 years ago (Hong et al., 1996). The peak copper production was estimated to be ~15000 metric tons per year during the peak production and rapidly decline thereafter (Hong et al., 1996). This peak Cu production lags ~500 years behind the observed peak in trace metal concentrations at Cleland Lake. Considering the ~350 year uncertainty in ice age-gas age differences, it is possible that this peak in trace metals is a product of atmospheric air pollution from metal smelting during the Roman empire. After the demise of the Roman empire world copper production remained low. So do the metal concentrations in

Cleland Lake. The world copper production peaked again during the Northern Sung period

(around 9000 years ago) based in China. Calculated average production was at a similar magnitude (15000 metric tons per year) to that during the Roman empire. Cu, Ni, and Zn concentrations in Cleland Lake also show an increase between 1200 and 1000 cal yr BP.

However the other two cores in Cleland Lake do not show a dramatic increase in the above

metals during any of those periods (Appendix 3). If the metal increase is caused by an atmospheric source, it should be recorded in the other parts of the lake as well. The absence of such a signal from the shallow or the intermediate cores suggest that the peaks in trace metals are just a coincidence. We can further infer that the observed trace metal enrichments are results of redox changes within the lake as explained in the previous paragraph.

Rapid changes in lake levels and diatoms around 2600 cal yr BP is followed by a rapid drop in illite and diatoms (around 2550 cal yr BP) and remain low till 2250 cal yr BP. Diatom and Illite levels gradually increase between 2300 and 1100 cal yr BP with three prominent peaks in illite and diatom abundances centered on 1250, 1100 and 950 cal yr BP. Simultaneous with the diatom and illite increase, (from 1250 to 950 cal years BP) 18O values show a rapid 2‰ depletion, suggesting an increase in lake level. Another prominent lake level rise is indicated by increases in diatoms and illite as well as depletion of δ18O between 1000 and 750 cal yr BP and again between 600 to 650 cal yr BP. Simultaneous with this increase P, S , Si, and Cr levels rapidly increase, suggesting an increase in nutrient input to the lake. These changes further support the idea of high frequency moisture variability during late-Holocene as discussed in chapter 5.The latter moist period suggested by an isotopic depletion between 600 and 750 cal yr

BP, i.e. 1200 to 1350 AD coincide with the Medieval Climatic Anomaly (Mann et al., 2009;

Steinman et al., 2014). Recent research in the area has suggested moist climatic conditions in the

Pacific Northwest as opposed to recorded drier conditions across the other parts of North

America during this period (Mann et al., 2009; Steinman et al., 2014). Modeling studies have suggested a prevailing warmer central and western Pacific ocean SST giving rise to La Niña-like conditions and a warm (positive) phase of PDO as the cause of this wetness (Cobb et al., 2003;

Mann et al., 2009; Steinman et al., 2014).

Similar transition in geochemical parameters are observed in the Elk Lake, Minnesota during the transition from the cool and moist early Holocene to the warm and dry mid Holocene

(Dean et al., 1993). This transition in moisture regimes in Elk Lake is characterized by a transition from a Mn rich lake to low Mn concentrations (Dean et al., 1993). Increase in Mn concentrations in Elk Lake is considered as a response to increased mobilization of organometallic complexes through humus accumulation in the pine dominated vegetation (Dean et al., 1993). Transition in vegetation to prarie-savanah during the mid-Holocene might have limited the mobility of these metal complexes resulting a drop in Mn concentration in the lake.

Si, Al, Na, Mg K increase as a response to increased wind erosion during the dry prairie period from 8000 to 4000 varve years BP (Dean et al., 1993). In contrast to the observed increase in diatom with increase in lake level at Cleland Lake, increase in diatom abundances are recorded during the prairie dry period in Elk Lake, Minnesota (Bradbury and Dieterisch-Rurup, 1993).

Increase in diatom productivity is a result of increase in nutrient influx from terrestrial sources

(Bradbury and Dieterisch-Rurup, 1993; Dean et al., 1993).

Conclusions

Diatom and illite levels of Cleland Lake were reconstructed using the first and the fourth principal components of sediment color reflectance data. Illite levels covary with diatoms thought out the record indicating diatom blooming with increased nutrient loading. Diatom and illite records are inversely correlated with the δ18O records, suggesting higher diatom abundance during periods of positive P/E balance and vice versa. Periods of enhanced diatom productivity and positive P/E conditions recorded during middle and late Holocene (4400, 3550, 3150, 2700,

1150, 950 and 600 cal yr BP) are simultaneous with the neoglacial advances in western Canada.

Hence these episodes could be a response to the cold, humid regional climate during those events. Diatom and illite levels and δ18O values show high frequency variability during late

Holocene, supporting frequent changes in moisture levels. The largest observed 18O

(hydrologic) depletion occurred centered on 2550 cal yr BP, indicating a rapid rise in lake level.

Contemporaneous with this depletion, nutrient (P, S) levels increase, suggesting a higher rate of groundwater inflow and mineral delivery. This rapid depletion is simultaneous with the solar irradiance minimum centered around 2700 cal yr BP and could be linked to the SST increase in

Eastern Tropical Pacific Ocean and resulting El-Niño like conditions. Trace metal concentrations reflect a transition at 2600 cal yr BP, from a period of high Mn and low Cu and Ni concentrations to a period of low Mn and high Cu and Ni levels. This indicates the transition from a lake that underwent frequent lake level changes and seasonal mixing during mid-

Holocene to a deep, well stratified lake during late time. Late Holocene Ni and Cu enrichment further indicate intense organic matter decomposition under reducing conditions. Increase in lake level, illite, and diatom levels from 600 to 750 cal yr BP coincide with the Medieval Climatic

Anomaly. This wet period could be associated with the prevailing warmer central and western

Pacific Ocean SST giving rise to La Niña-like conditions and a warm (positive) phase of PDO

(Cobb et al., 2003; Mann et al., 2009; Steinman et al., 2014).

CHAPTER 4

EFFECT OF MAZAMA TEPHRA ON LAKE SEDIMENTARY COMPOSITION AND

PRIMARY PRODUCTIVITY

Abstract

Diffuse Spectral Reflectance (DSR) data defines the spectral shape of components in the samples, enabling the quantitative analysis of sediment composition. Principal component analysis of DSR data from Cleland Lake derived five components explaining 97% of the variability. PC 1 and 3 correlate with illite+sphalerite and smectite+chlorite, respectively. Three of the PCs (PC 2, 4, and 5) correlate with the standard reflectance curves of dinoflagellate algae, diatoms and cyanobacteria, and were used as a proxy for paleoproductivity. Lake sediment and phytoplankton abundance reflect considerable changes after the deposition of Mazama tephra around 7700 cal yr BP. Pre-Mazama Cleland Lake was dominated by illite, which was probably derived from the weathering of argillaceous bed rock. Illite content rapidly declined around 7760 cal yr BP and was replaced by smectite and chlorite. Smectite levels remain high during the mid-

Holocene (7700 to 4400 cal yr BP) and gradually decrease after that. Post-Mazama smectite and chlorite could be a result of weathering of mafic tephra in the alkaline lake setting.

During the past 14000 years, the lake was dominated by three communities, each of which included cyanobacteria. Late glacial lake primary productivity (14000 to 11500 cal yr BP) was limited to cyanobacteria community 1. Dinoflagellate algae and cyanobacteria community 2 became prominent during early-Holocene (11500 to 8500 cal yr BP), when the lake level was much lower and regional climate was warm and arid. The abundance of diatoms and cyanobacteria community 3 gradually increased after 9000 cal yr BP, as a response to increased regional moisture. A rapid decline in dinoflagellates, diatoms and cyanobacteria community 3 is recorded from 8500 to 7700 cal yr BP, simultaneous with the global cooling event centered on

8200 cal yr BP. A rapid recovery in all three primary producers is recorded within several decades after the Mazama tephra deposition, which was possibly due to the increase in nutrient and silica input from tephra weathering. The abundance of primary producers gradually declined after 6800 cal yr BP, probably in response to the decline in the nutrient input from tephra weathering. In contrast, diatom abundance in the intermediate core exhibits a long-term increasing trend, initiated during the early Holocene. This trend shows a negative correlation with the northern hemisphere summer insolation at 65N. Recorded changes in the clay mineral composition and phytoplankton abundance in Cleland Lake suggest that the lake is sensitive to long term regional and global climatic changes, as well as to rapid environmental changes such as volcanic tephra deposition.

Introduction

Explosive volcanic eruptions can eject large amounts of silicate rock fragments

(pyroclastics) and tephra into the atmosphere. Tephra deposition can cause significant physical, chemical, and biological impact in the areas where they fall. In the case of small-scale eruptions,

the effects are limited to a few tens of kilometers. In contrast, large eruptions, such as the Mount

Mazama eruption, can disperse tephra over a million square kilometers (Kittleman 1973; Ayris and Delmelle, 2012). At the local scale, hot tephra falls can have immediate ecological impacts and are more well-studied (Telford et al., 2004) than the more common dispersal of thin layers of tephra in distal areas (Telford et al., 2004).

Effects of tephra deposition in lakes can have immediate impact, as well as trigger long- term ecological changes (Hickmon and Reasoner, 1994; Telford et al., 2004; Ayris and Delmelle,

2012). The most common immediate response is an increase in turbidity (Hickmon and Reasoner

1994; Telford et al., 2004; Ayris and Delmelle, 2012). Increased turbidity reduces light penetration and can lower primary productivity (Ayris and Delmelle, 2012); however, Stoke's

Law-based calculations suggest tephra with average fine sand size will settle down to the lake bottom within a few hours (Telford et al., 2004; Ayris and Delmelle, 2012). When deposited on water bodies, soluble elements can enhance the growth of phytoplankton (Telford et al., 2004;

Ayris and Delmelle, 2012). Several studies have reported changes in diatom composition or diatom concentration changes in response to tephra deposition. A decline in phytoplankton abundance can result immediately after the tephra deposition due to growth inhibition by soluble toxic metals, such as Al, As, Ba, Cd, Co, F, Mn, Ni, Pb and Zn (Hickmon and Reasoner, 1994).

In contrast, enhanced silica input from tephra weathering can increase diatom abundance, as well as their diversity (Ayris and Delmelle, 2012). Enhanced diatom concentrations can last from several decades to several hundreds of years depending on the size of the tephra fall relative to the lake size and turnover rate. Increased diatom abundance lasting up to 300 years have been reported from Alpine lakes in Alberta, Canada (Telford et al., 2004 and references therein; Ayris and Delmelle, 2012). A rapid increase in algal species with only a three to four month lag has

been reported from lakes around Mount St. Helens tephra after its eruption in 1980 (Hickmon and Reasoner, 1994). The shallow areas of the lakes were colonized rapidly (Hickmon and

Reasoner, 1994). The eruption of Mount Mazama, (~7700 cal yr BP) deposited volcanic ash that carpeted much of North American continent. This event serves as a stratigraphic marker for age determination in lake deposits in the western US (Hallet et al., 1997), but little work has been conducted on the impact of the event on the paleoecology of lake systems. This chapter focuses on the changes of lake sediment and phytoplankton composition at Cleland Lake during the past

14000 years, focusing specifically on the mid-Holocene Mazama tephra deposition.

Methods

Preparation of a Common Age Model

Both the B-09 deep basin core and the F-09 intermediate depth core from Cleland Lake have well-constrained age models. The core chronology of the F-09 core is based on four AMS

14C dates of terrestrial macrofossils and two tephra layers (Table 2.1). It was assumed that the surface sediment was deposited in 2009 during the period of sample collection (-59 cal yr BP). A linearly-interpolated age model was prepared by point-to-point calibration using CLAM age modeling software (Blaauw, 2010). The basal age was extrapolated following the trend established by the two oldest recorded events (Figure 2.2). The B-09 chronology is based on eleven AMS 14C dates of terrestrial macrofossils and two tephra layers (Table 3.1). A linearly- interpolated age model was prepared following the same calibration procedure in the intermediate core (Figure 3.3).

The initial age model for the shallow water core, E-09 consists of six radiocarbon ages.

Rapid transition of carbonate mud to pink, silty-clay with a sharp erosional surface at 151 cm

suggest a period of erosion, and prevents us from interpolating ages between 0 cm and 158 cm.

Therefore, in order to construct a reliable age model, we employed a variation on Shaw’s method of graphic correlation employing geophysical data to create a common stratigraphic framework for the three cores (Boggs, 2011). The F-09 core, which extends the full length of the Holocene and into the deglacial period (170 to 14000 cal yr BP), was used as the reference depth scale to prepare a common age model for all three cores. The B-09 core was preferred over the well age- constrained deep-basin core (B-09), as the latter only extends back to 7600 cal yr BP. By placing all of the cores in a common stratigraphic framework, we can build a common age model that makes use of all the available ages.

Three control points were initially selected based on the magnetic susceptibility records from the shallow and the intermediate cores (Figure 4.1). Using these three control points, a transfer function was created to linearly interpolate depth in E-09 to depth in F-09. This depth scale was refined by the addition of several more depth control points picked by comparing the peaks and troughs of PCs of reflectance data from E-09 core with that of the E-09 core (Figure

4.2). A final, revised transfer function to relate depth in E-09 to depth in F-09 was generated by interpolating the depths in between these age control points (Figure 4.2, Table 4.1). Similarly, a new depth scale was prepared for the deep core (B-09) by comparing the reflectance PCs from the three cores (Figure 4.3). The PCs of the reflectance data from all three cores were plotted on the depth scale of the intermediate core (Figure 4.4).

After calibrating both the shallow and the deep cores to a common depth scale (depth scale of the intermediate core), the radiocarbon and tephra ages from the shallow and deep cores

0.8 15

0.6 13

- - 11 0.4 9 0.2 7 0 5 -0.2 3 -0.4

09 Magnetic susceptibility Magnetic (*10^ susceptibility 09

- E09 Magnetic succesptibility succesptibility (10^ Magnetic E09 -0.6 -1 F

-0.8 -3 0 50 100 150 200 250 300 Depth (cm)

0.8 E-09 F09 15

0.6 13

- - 11 0.4 9 0.2 7 0 5 -0.2 3 -0.4

09 Magnetic succesptibility Magnetic (10^ succesptibility 09

09 Magnetic susceptibility Magnetic (*10^ susceptibility 09

- F E -0.6 -1

-0.8 -3 0 50 100 150 200 250 Depth (cm)

Figure 4.1. Magnetic susceptibility of the shallow (E-09) and the intermediate (F-09) cores. Top - plotted on their original depth scales; bottom magnetic susceptibility of the shallow core plotted on the depth scale of the intermediate core. Green diamonds are the initial age control points used for wiggle matching.

were projected onto this common depth scale (Figure 4.4). Several new age models were prepared combining the radiocarbon and tephra ages from all three cores. Out of these models, a model which included eleven 14C dates from the deep core, four 14C and two tephra dates from the intermediate core and two 14C dates from the shallow core was selected as the model which gave the most plausible ages for the shallow core (Figure 4.5). When selecting the age control points, the dates that do not follow the linear sedimentation rate were avoided. The model that gave the most reasonable age constraints for stratigraphically known events as the Mazama tephra deposition was given the priority.

3 6 E-09 PC 2 F-09 PC2

4 1

09 PC 2 PC 09

09 PC 2 PC 09 - F E -1 0

-3 -2 0 50 100 150 200 250 300 Depth (cm) 3 6 E-09 PC 2 F09 PC2

4 1

2 2 PC 09

- F

09 PC 2 PC 09 -

E -1 0

-3 -2 0 50 100 150 200 250 300 Depth (cm)

Figure 4.11 Reflectance PC 2 of the shallow (E-09) and the intermediate (F-09) cores. Top: plotted on their original depth scales. Bottom: PC-2 of the E-09 core plotted on the depth scale of the intermediate core. 87

2 4 B-09 PC 2 F09 PC 2

09 PC 2 PC 09

09 PC 2 PC09

- F

B 0

-2 -2 0 50 100 150 200 250 300 Depth (cm) 2 4 B-09 PC 2 F09 PC 2

09 PC 2 PC 09

09 PC 2 PC09

- F

B 0

-2 -2 0.00 50.00 100.00 150.00 200.00 250.00 300.00 Depth (cm) Figure 4.3.Reflectance PC 2 of the deep (B-09) and the intermediate (F-09) cores. Top - plotted on their original depth scales. Bottom: PC-2 of the B-09 core plotted on the depth scale of the intermediate core.

0 50 100 150 200 250 300 3

1 PC 1 PC

-1

F-09 E-09 B-09

-3 4

2.5

1 2 PC

-0.5

-2 5

1 PC 3 PC

-1

-3 3

1 PC 4 PC

-1

-3

2.5

0.5 PC 5 PC

-1.5

-3.5 0 50 100 150 200 250 300 Depth (cm)

Figure 4.4 First five principal components of reflectance data from the three cores of Cleland Lake plotted on a common depth scale. Shallow core E-09 (red line), intermediate F-09 core (green line), and deep B-09 core (green line).

Table 4.1 Age control points used for calibration of the depths of the shallow core to the depth scale of the intermediate core.

depth in E core depth in F core 30.5 32 38.25 39.75 45.75 49.25 55 59.75 82.75 75.25 108.5 90.25 116.5 122.25 132 145.5 158 193.75 225.5 212.5 242 225.75 246 235.75 276.75 265.75 292 277.25

Results and Discussion

Interpretation of the Principal Components

The first five PCs of the reflectance data explain 97% of the variance in the derivative transformed, reflectance data set. Principal component 1 explains 28% of the variance in the data set and is interpreted as a mixture of illite (75%) and sphalerite ( 30%) (Figure 2.3). This PC can be used as a proxy for fluvial input to the lake. Illite, which is present throughout the record, likely represents the autochthonous clay mineral within the lake basin. Principle component 2 explains 24% of the variance in the reflectance data and reflects a mixture of dinoflagellate algae

pigments (60% peridinin and 33% pheophytin-a) and goethite (7%; Figure 2.3). Major photosynthetic pigments of flagellate algae belonging to the division Pyrrophyta include: chlorophyll-a, chlorophyll-c, and peridinin (Sze, 1993). The presence of peridinin and pheophytin-a (a degredation product of chlorophyll-a) in PC 2 identifies this component as a proxy for dinoflagellate abundance in the lake (Sze, 1993). Dinoflagellate algae are common in oligotrophic, high latitude/polar lakes, or in water with low light intensities, such as snow covered or turbid lakes (Sze, 1993). Large, positive PC 2 values indicate greater dinoflagellate abundance and low nutrient conditions or low light intensity. Principal components 3, 4 and 5 contain variable amounts of cyanobactera pigments and hence may represent the succession of three different communities that include cyanobactera in Cleland Lake (interpreted as cyanobacteria communities 1, 2 and 3, respectively, hereafter). PC 3 explains 23% of the variance and is a mixture of phycocyanin (80%), the primary accessory pigment found in fresh water cyanobacteria and the clay minerals smectite and chlorite (20%, Figure 2.3.4.c). Principal component 4 explains 16% of the variance and is a mixture of the spectral signatures for bacillariophycea (70%) and phycocyanin (30%) (i.e. diatom and cyanobacteria photosynthetic pigments, Figure 2.3.4.d). The spectral signature for bacillariophycea has a positive correlation with PC 4, indicating greater diatom abundance when the PC 4 scores increase. In contrast, phycocyanin reflects a negative correlation with PC 4, indicating lower levels of cyanobacteria when PC 4 increases, and vice versa. This defines the inverse relationship and contrasting appearances of diatoms versus cyanobacteria in sediment from Cleland Lake. PC 5 explains six percent of the variance of reflectance data is related to the cyanobacteria accessory pigments, allophycocyanin (54%) and pheophytin-A (46%), (Figure 2.3); therefore, this component is considered as a proxy for a third cyanobacteria assemblage.

Reconstruction of the Sedimentary Composition and the Lake Phytoplankton Composition

The eruption of the Mazama tephra layer was accurately dated as 6730 ± 40 14C years BP

(7697 cal yr BP)(Hallet et al., 1997). In the Cleland Lake intermediate core, the Mazama tephra layer is represented by a five centimeter thick, reddish brown layer composed of sand-size ash

(209 cm to 213 cm). This layer is characterized by low Ca and Ni concentrations, and high Si, K,

Ti, Mn, Fe, Zn and Rb concentrations (Figure 4.5). In the shallow core, a distinct tephra layer was not visually identified. A nine centimeter thick carbonate mud and silty clay layer (228 cm to 219 cm) was identified as corresponding to the Mazama tephra layer in this core. This layer shares similar chemical (high Ti, Fe, Zn and K) and physical properties (high magnetic susceptibility and dry bulk density) with the Mazama ash layer in the intermediate core (Figures

4.1 and 4.5).

As reflected by PC 1 scores from the intermediate core, the lake was dominated by illite during the deglacial period (prior to 11500 cal yr BP; Figure 5.6.A and B). Higher clay input during this period could have been a result of higher runoff associated with glacial meltwater input during the deglaciation. Higher percent mineral matter content (75% to 85%) and higher concentration of detrital elements (Si, K, Ti, Fe, Ni and Rb; XRF-PC 1) further supports a higher mineral input to the lake during this period (14000 to 11700 cal yr BP) (Figure 5.9). The low abundance of diatoms and dinoflagellates, and higher amounts of cyanobacteria are recorded during local deglaciation from ~14000 to ~11800 cal yr BP.

100 B09 14C ages F 09 14C ages 150 E09 14C ages B-09 tephra ages 200 F-09 tephra ages

250 Depth in core F 09 F core in Depth 300

350

400 -500 1500 3500 5500 7500 9500 11500 13500 15500

Age (Cal yr BP)

50 Max 95 % CI 100 Min 95 % CI

150 B-09 F-09 200

E-09 Depth (cm) Depth 250

300

350

400 -500 4500 9500 14500 Age (Cal yr BP)

Figure 4.5. Top: All the available age data from Cleland Lake plotted on the depth scale of the intermediate core. Bottom: Age model for the shallow core, prepared using selected 14C and tephra ages from deep, intermediate and shallow cores.

age Si Cl K Ca Ti Mn Fe Ni Zn Rb Mo inc

6000 0 10000 40 1000 75000 400000 40000 30000 1200000 4000 10000 300010000 200

202

204

206

(cm) 208

depth 210

212

214

216

218

220 max.: 220 cm Figure 4.6. Elemental concentrations of the Mazama tephra layer (209 cm to 213 cm) in the intermediate core.

100 2100 4100 6100 8100 10100 12100 3 F-09 E-09 B-09 High Illite

1 PC 1 PC

-1

-3 4 High Dinoflagellates

2 PC 2 PC

5 -2 100 2100 4100 6100 8100 10100 12100 High Smectite 3

1 PC 3 PC

-1

-High3 Diatom 3 100 2100 4100 6100 8100 10100 12100

1 PC 4 PC

-1 High cyanobacteria -3

2.5 High cyanobacteria

0.5 PC 5 PC

-1.5

-3.5 100 2100 4100 6100 8100 10100 12100 Depth (cm)

Figure 4.7. First five principal components of reflectance data on the age scale (see text for an explanation of each PC). Shallow core E-09 (red line), intermediate F-09 core (green line), and deep B-09 core (green line), black shaded line represents the Mazama tephra layer.

Illite content rapidly drops around 11600 cal yr BP. As the illite input decreased, dinoflagellate and cyanobacteria concentration (reconstructed by PC 2) increased to their highest

Holocene levels. Dinoflagellate algae levels remain high between 11600 and 8500 cal yr BP in both the shallow and the intermediate cores. This period, which has higher dinoflagellate algae, low illite and low diatom in the intermediate core, has been identified as a period of low lake levels. The early-Holocene rapid drop in illite concentration in the intermediate core is a response to the reduced runoff simultaneous with the dry period. However, illite content and the detrital mineral matter (40-50%) in the shallow core remain high between 9500 and 8000 cal yr

BP., suggesting the shore line was closer to the core site during this period(Pompiani et al.,

2012). Illite concentration gradually decreases in the shallow core between 8400 and 7000 cal yr

BP and was replaced by high levels of smectite and chlorite around 7700 cal yr BP, coincident with the 7697 cal yr BP Mazama eruption, within the margin of the age error estimates.

Clay mineral levels (reconstructed from PC 1 and PC 3) show considerable variations between the three different core sites. The amount of clay minerals is higher in the shallow core than in the intermediate core. This could be due to association of detrital minerals in the shallow water areas from shoreline reworking and rapid lake level fluctuations.

Lake sediment is dominated by illite prior to the deposition of Mazama tephra. This illite could have derived from the weathering of argillaceous (aluminum bearing) sedimentary rocks and quartzite in the watershed (Poppe at al., 2001). Alkaline water in the lake might also have favored the formation of illite within the lake. The dominant sediment composition changes to smectite and chlorite following the deposition of Mazama tephra. The composition of the

Mazama tephra has been documented as consisting mainly of mafic minerals, plagioclase feldspar (73.1±1.2%), hypersthene (9.4±0.5%), magnetite (10.2±1.1%), hornblende (3.9±0.4%) and clinopyroxne (2.8±0.3%) (Kittleman, 1973). Smectite commonly forms from the weathering of mafic rocks in areas with poor drainage (Poppe at al., 2001). The alkalinity of the Lake and the availability of Ca further support the formation of smectitie. Chlorite is also another weathering product of mafic minerals. Therefore, it is likely that the chlorite and smectite found in Cleland Lake were secondary products of weathering of the Mazama tephra that deposited on the lake 7700 cal yr BP. No other bedrock source is likely as a source of the smectite and chlorite, and no other plausible event is observed to provide an alternative phyothesis.

Smectite levels remain high in the shallow core longer than the deep or the intermediate core. In the shallow core, smectite levels abruptly increase around 7600 cal yr BP and remain relatively high until 4200 cal yr BP. In the deep core, smectite levels remain high from 7500 to

6000 cal yr BP and then rapidly drop after that. Smectite levels increase more gradually in the intermediate core compared to the other two cores. Increasing PC 3 scores are recorded from this core from 11500 cal yr BP, simultaneous with the early Holocene dry period. PC 3 scores gradually decrease after 7550 cal yr BP. A rapid increase is again recorded around 7400 cal yr

BP and remain high until 6400 cal yr BP. The early Holocene gradual increase (11500 to 7550 cal yr BP) in PC 3 could be a result of an increase in cyanobacteria abundance in response to the warm regional climate rather than an increase in smectite content. Cyanobacteria predominate during the summer months and in warm climates due to their higher temperature tolerance

(Chapter 2 this dissertation; Nõges et al., 2003). In contrast, the mid-Holocene (7400 to 6400 cal yr BP) abrupt increase could be a result of increase in smectite, following the Mazama tephra deposition.

At the end of the early-Holocene dry period (9000 ± 500 cal yr BP), diatom abundance in the shallow and the intermediate core gradually increasesd in response to the increase in lake levels. An abrupt drop in diatom abundance, however, was recorded from 8500±150 to 7705 ±

200 cal yr BP in the shallow core. Rapid drop in the abundance of cyanobacteria and dinoflagellate algae were also recorded in both the shallow and the intermediate core between

8500±150 to 7705 ± 200 cal yr BP. In contrast, the diatom and cyanobacteria community 3 abundance in the intermediate core does not show a simultaneous, rapid drop. Instead, they follow increasing trends initiated during the early Holocene (8900 and 11600 cal yr BP, respectively). Given the short range of uncertainty, we can infer this decline in phytoplankton was likely initiated prior to the Mazama tephra deposition (7708±200 cal yr BP), as a response to the regional climatic changes. This period (8200 to 8500 cal yr BP) experienced abrupt climatic change throughout the northern hemisphere (Alley et al., 1997; Mayewski et al., 2004; Gavin et al., 2011). A significant cooling is recorded from Greenland ice cores and North Atlantic records

(Alley et al., 1997; Mayewski et al., 2004; Gavin et al., 2011). Several paleoclimatic reconstructions from Western Canada record simultaneous climatic changes (Hallet and Hills,

2006; Gavin et al., 2011). Simultaneous cooling in the region is suggested by glacial re-advances and declines in tree lines (Hallet and Hills, 2006; Gavin et al., 2011). It is suggested this global event was a response to the reorganization of the climate system after the collapsing of the late

Pleistocene ice sheets (Alley et al., 1997; Hallet and Hills, 2006; Gavin et al., 2011).

All three phytoplankton groups record a rapid recovery in their abundance within a few decades after the Mazama tephra deposition (7650±200). Dinoflagellate algae and diatoms abundance remain high until 6,800 cal yr BP and gradually decrease after that. Cyanobacteria community 3 became prominent during the mid-Holocene (7650 to 5500 cal yr BP). Their

abundance became prominent first in the shallow core (7650 to 6800 cal yr BP) and then in the intermediate core (6800 to 5500 cal yr BP). It is possible this community became prominent in the shallow core in response to the nutrient input from the Mazama tephra deposition and then were later distributed into the deeper part of the lake. Similar increases in cyanobacteria and benthic diatom species abundance have been reported elsewhere (Hickmon and Reasoner, 1994 and references therein). Hickmon and Reasoner, (1994 and references there in) reported the shallow areas of a lake were the first to be colonized by algae during a similar event. When the nutrients from tephra weathering become depleted, the cyanobacteria community gradually declined in the lake. In the intermediate core, a continuous increase in diatom abundance was recorded from mid-Holocene (9000 cal yr BP) until the late Holocene (170 cal yr BP), with the highest abundance after 1050 cal yr BP. These long-term increasing trends in the diatom and cyanobacteria community 3 in the intermediate core could be a response to the increased effective summer moisture in the region (Chapter 2, this dissertation; Hallet and Hills, 2006;

Gavin et al., 2011). Diatom abundance in the intermediate core also correlates with northern hemisphere insolation trends (Figure 2.5); therefore it is also possible this decreasing trend in diatoms was driven by declining insolation.

The decline in the phytoplankton abundance is also consistent with the rapid drop in percent organic matter in both the shallow and the intermediate cores (Figure 5.1 and 5.8). In the intermediate core, a rapid drop in organic matter percentage was recorded at 7050 cal yr BP.

Organic matter percentage gradually increases after 7050 cal yr BP, but does not reach the pre-

Mazama level until 5600 cal yr BP. A simultaneous rapid decline was also observed in the shallow core at 7708 cal yr BP. In contrast to the slow recovery in the intermediate core, a rapid

increase was observed in the shallow core, reaching the highest Holocene levels (57%) within a few decades after the tephra deposition (7550 cal yr BP).

Conclusions

Down-core variations of first five principal components of sediment reflectance data reveal the variations of sediment composition and lake paleo-productivity during the past 14000 years. The pre-Mazama lake was dominated by illite derived from bedrock weathering. Sediment composition transitioned to smectite and chlorite after the deposition of the Mazama tephra. Lake productivity was low and mostly limited to early successional, cyanobacteria communities

(community 3) during the late Pleistocene deglaciation (prior to 11600 cal yr BP). Dinoflagellate algae and cyanobacteria community 2 became dominant, while diatoms decreased during the early Holocene (11600 to 8600 cal yr BP) in response to lower lake levels. Diatom and cyanobacteria community 3 gradually increased after 9000 cal yr BP. A rapid decline in diatoms, cyanobacteria and dinoflagellate algae is recorded from 8500 to 7700 cal yr BP, simultaneous with the global cooling event centered on 8200 cal yr BP and the subsequent Mazama tephra deposition around 7700 cal yr BP. Primary productivity rapidly increased after the tephra depositions, possibly in response to an increase in silica and nutrient input. Higher levels of impact from tephra are observed in the shallow core than the intermediate or the deep core.

Dinoflagellate algae, diatoms and cyanobacteria community 3 remain high until 6800 cal yr BP and gradually decrease after that. In contrast, the diatom abundance in the intermediate core does not indicate dramatic changes simultaneous with the tephra deposition. Instead, the diatoms follow a long-term increasing trend associated with the decreasing northern hemisphere summer insolation. This further indicates that the diatom abundance in the shallow core is controlled by

100

the availability of nutrients. In summary, the observed clay mineral and lake primary productivity changes of the Cleland Lake suggest the lake was considerably influenced by the

Mazama tephra deposition. Long-term fluctuations in algal and cyanobacteria compositions were also influenced by regional climatic changes.

101

B A CHAPTER 5

HOLOCENE PALEOPRODUCTIVITY AND LAKE LEVEL CHANGES AT CLELAND

LAKE

Abstract

Lake level changes are directly related to chemical composition of lake sediment and can lead to

facies changes. Sedimentological and chemical analyses of three sediment cores from Cleland

Lake, British Columbia, were used to reconstruct paleolake levels and precipitation/evaporation

balance in southeastern British Columbia. Presence of unlaminated, carbonate-rich sediment in

the intermediate depth core (from a water depth of 12 m) suggest low lake levels between 11800

and 8600 cal yr BP. Deposition of well-laminated, organic matter-rich sediment from 7600 to

3600 cal yr BP indicates higher lake levels during the mid-Holocene. During the late Holocene,

percent carbonate and Ca concentrations reflect significant variations between high and low

levels. The five periods of low lake levels (from 3600 to 3200, 2300 to 2200, 2100 to 1800,

1300 to 1200, and 700 to 600 cal yr BP) alternate with periods of high lake levels. These climatic

changes are also documented in the previous independent lake-level reconstructions from British

Columbia. Spectral analyses further suggest low frequency variations in lake-level changes

during the early Holocene and low frequency changes during the late Holocene.

102

Introduction

Lake sediment normally consists of detrital material, algal or terrestrial organic matter, as well as inorganically precipitated carbonates and authigenic minerals (Schnurrenberger et al.,

2003). The hydrology of the lake plays a dominant role in determining the mineralogy and facies development in a carbonate lake (Tucker and Wright, 1990; Platt and Wright, 1991).

Hydrologically-open lakes have permanent outflow. Inflow from the surrounding drainage basin and the precipitation is balanced by evaporation and surface or subsurface outflow (Tucker and

Wright, 1990; Platt and Wright, 1991). In contrast, hydrologically-closed lakes have no surface outflow and the water level is controlled by the balance of inflow (precipitation), and evaporation and the groundwater balance (Tucker and Wright, 1990; Platt and Wright, 1991).

Provided that the groundwater contribution is small, in these lakes, variable P/E balance

(precipitation/evaporation) produces hydrologic instability. Such instability produces water level variations that are reflected in the physical, biological, and chemical composition of sediment accumulating in the basin (Talbot, 1990; Battarbee, 2000; Leng and Marshall, 2004: Steinman et al., 2010ab; Pompeani et al., 2012).

A lake’s chemical composition is directly affected by its depth. Lake-level changes can shift littoral sedimentary environments away from, or towards, the center of the basin, ultimately leading to changes in depostional facies (Barker et al., 1994; Brugam et al., 1998).

Sedimentological information (grain size, composition, color, and sedimentary structures) can provide valuable information about such changes, allowing scientists to interpret paleodepositional environments. Loss on ignition (LOI) is a traditional and economical method of determining the carbonate and inorganic content of calcareous sediment, which are poor in clay (Heiri et al., 2001). LOI results can be inaccurate when appreciable amounts of clay are

103

present. Such inaccuracies are due to the chemical decomposition of the clay minerals, resulting in an underestimation of the detrital component, and an overestimation of the carbonate or organic components. Here, we present lake-level changes identified from LOI results and Initial

Core Descriptions (ICD) of lake sediments recovered from Cleland Lake, British Columbia.

These data are compared with XRF elemental data, obtained by a whole core, scanning XRF system.

Carbonate precipitation in freshwater lakes can occur when lake water becomes supersaturated with calcium carbonate. Supersaturation can occur through biological processes, physicochemical processes, or the interaction between physicochemical and biochemical processes (Duston et al., 1986; Nelson et al., 2009). The hydrology of the lake (input and output of surface water), sediment input, biogenic production, and temperature are the major factors governing lacustrine carbonate precipitation (Tucker and Wright, 1990; Platt and Wright 1991;

Gierlowski-Kordesh, 2010). In general, climate is a primary factor governing carbonate precipitation (Gierlowski-Kordesh, 2010). Temperature and precipitation determine the biological productivity of the lakes and therefore, affect the rate of deposition of biological carbonate (Platt and Wright, 1991).

Physiochemically-derived carbonates include the reworked detrital carbonates imported from hinterlands (Tucker and Wright, 1990). Inorganic carbonate precipitation can also occur when Ca-rich water falls into lake water (Tucker and Wright, 1990). Inflow can occur either as spring water or river flow. Carbonates precipitated in this manner are common in the glacial terrains of the northern US and Canada, where dolostone and limestone bedrocks provide a substantial contribution to the till (Brown et al., 1992).

104

Carbonates can deposit primarily by evaporation in relatively shallow lakes in arid and semi-arid environments. Carbonates precipitated in this way are impure and contain considerable amounts of clay, dolomite, and evaporates such as gypsum, halite and brine salts (Murphy et al.,

1996). Carbonate production in marl lakes is the direct result of three features common to all carbonate lakes: (1) lake basins are developed on units containing limestone or dolomite, either as clasts or bedded strata; (2) a considerable volume of water that enters the lake basin is

+2 - enriched with Ca and HCO3 during its movement through CO2-charged soil zones and calcareous substrate; and (3) the climate of the region causes a considerable change in the temperature of the epilimnotic waters (a maximum in the summer and a minimum in the winter)

(Duston et al., 1986) and ultimately leads to stratification, which separates the epilimnion (the near surface water column of a lake) from the hypolimnion. The first two features are a result of the geologic setting of the lake and merely sets the basis for carbonate deposition. The third feature, climate, which creates a temperature gradient, is the main process involved in marl production (Duston et al., 1986).

Biogenic carbonates derive either from the skeletal remains of various organisms, (such as molluscs, ostracodes, charophytes and phytoplankton) or from biogenically-induced carbonate precipitation. Organisms play an important role in this process by removing CO2 during photosynthesis, resulting in biogenically-induced precipitation (Tucker and Wright, 1990). The main contributors of direct carbonate precipitation are charophytes (small aquatic algae), microscopic algae, and cyanobacteria (Tucker and Wright, 1990). Charophyte remains are common in many lake deposits and occur as both calcified reproductive structures and plant stem encrustations (Tucker and Wright, 1990). Charophyte remains are good indicators of fresh water

105

environments in paleolake deposits (Tucker and Wright, 1990). Cyanobacteria and green algae form extensive carbonate deposits, including tufas, mounds, stromatolites, and oncoids.

Lake Sedimentary Structures

Lake sedimentary structures can provide important information about depositional environments, including depositional hiatuses, slumps or turbidites, and sediment reworking.

Laminae present in the sediment have regular to irregular thicknesses and can be continuous or discontinuous sediment lenses. They consist of varying percentages of silicates and carbonates deposited in the profundal zone (aphotic zone below the thermocline) (Tucker and Wright, 1990;

Gielowskie–Kordesch, 2010). Laminae are referred to as varves only if the lamination patterns represent true annual (seasonal) sedimentation events. Various processes can give rise to lamination in lakes. For instance, seasonal variation in precipitation of carbonates produces couplets of dark and light laminae. During spring and summer, carbonate sedimentation is high and light-colored layers enriched with calcite and aragonite are typically deposited. During periods of high runoff, darker laminae consisting of clastic material including clay minerals, quartz, detrital calcite, and dolomite precipitate. Couplets of carbonates and siliciclastics could derive from variations in sediment fluxes to the lake. Siliciclastic layers are formed during periods of enhanced runoff or tubidites (Tucker and Wright, 1990). These couplets do not have as regular a thickness as seasonal couplets (Tucker and Wright, 1990). In freshwater lakes, preservation of laminations occurs when anoxia reduces the bioturbation, or when the deep-water sediment is protected from winds and currents (Gielowskie–Kordesch 2010).

106

Two main facies groups have been identified from the analysis of ancient and modern carbonate lake environments; the lake margin (littoral) and the lake basin (pelagic) (Tucker and

Wright, 1990).

Lake Margin (littoral zone)

Biogenic or bioinduced carbonate production predominates in shallow water less than 10 m in depth. Inorganic carbonate precipitation can also take place due to warming and wave agitation, or by mixing Ca-rich stream inflow with Ca-rich lake waters (Platt and Wright, 1991).

Mudstone facies dominate in shallow lakes with lower wave action (Platt and Wright, 1991).

Evaporites, soil, or peat can be depositted during low lake stands (Platt and Wright, 1991).

Littoral carbonates can laterally extend to the flood plain where clastic fluvial deposits dominate.

Littoral zones are inhabited by rooted aquatic plants and charophyte algae. Therefore, the zone is dominated by bioclasts and algal sediment (Platt and Wright, 1991). Carbonate production in this zone is induced by the photosynthetic CO2 uptake (Platt and Wright, 1991).

Decrease in CO2 cause an increase in pH leading to saturation of carbonate-ions. These chemical shifts result in subsequent carbonate encrustation of reeds and charophyte stems as well as the calcification of charophyte oogonia (reproductive structures). The encrusted carbonates rapidly break down in-situ into carbonate mud (clay-sized carbonate grains). Carbonates produced by mollusk or ostracode shell fragments, as well as small bioherms built by algae and cyanobacteria also contribute to littoral carbonates (Platt and Wright, 1991).

107

Lake Basin

Carbonate sedimentation rates can be higher within the lake basin, where they are dominated by bioinduced carbonate precipitation by pelagic phytoplankton. Sediment can also be supplied from redeposition, while inorganic precipitation can take place during lake turnover

(Platt and Wright, 1991). Fine-grained siliciclastic material can also accumulate more rapidly within the lake basin compared to the lake margin. Distribution of the facies in this zone is controlled by lake dynamics, such as the stratification of the water column and lake circulation

(Platt and Wright, 1991). Strong stratification in deep lakes would develop anoxic bottom water conditions. Absence of bioturbation under these conditions would help preserve well-laminated sequences. CO2 uptake during algal blooms results in precipitation of thin carbonate layers.

Subsequent settling of these phytoplankton will generate organic-rich layers, giving rise to alternating carbonate and organic-rich layers. These layers reflect the seasonal variability of phytoplankton productivity (Platt and Wright, 1991). In contrast, poorly-laminated facies predominate in shallow lakes. Periodic oxygenation of the bottom waters supports bioturbation of bottom sediment and enhances the decay of organic matter as well as the destruction of laminae (Platt and Wright, 1991).

Methods

Sedimentological Analysis

Initial core descriptions were made based on general identification of the compositional variability of the core and the recognition of sedimentary structures, including bedding structures and potential unconformities (Schnurrenberger et al., 2003). Sediment was identified using the classification system developed by the Lake Research Center (LRC) at the University of

108

Minnesota. Classification was based on two primary observations: the macroscopic structure of the sediment and the major and minor components of the sediment (Schnurrenberger et al.,

2003). Macroscopic structures include bedding features, texture, and color (Schnurrenberger et al., 2003). Major and minor components explain the composition of the sediment, such as clay, carbonate, or organic matter (Schnurrenberger et al., 2003).

Sediment color was observed on the freshly opened surfaces of the cores while they were still wet. Texture was determined using grain-size folders and by rubbing small quantities of sediment between fingers (Schnurrenberger et al., 2003). Beds are defined as layers with thicknesses greater than 1 cm, whereas laminations are layers with thicknesses less than 1 cm.

Laminations are common in lakes exhibiting large compositional changes over a relatively short interval of time (Schnurrenberger et al., 2003).

Loss on Ignition (LOI)

Total inorganic and organic carbonate content was measured by loss on ignition at the

University of Pittsburgh, following the methods explained by Heiri et al., (2001). The Cleland

Lake sediments were measured at 2 cm intervals for LOI. Sequential loss on ignition is a traditional and widely used method for quantitative determination of the carbonate and the organic carbon content of sediment. The method is particularly well-used in lacustrine settings.

Sediment was first oven dried at 105oC for 12 to 24 hours, until it reached a constant weight. This process removes the water contained in the sediment and gives the dry weight of the sediment (DW105) (Heiri et al., 2001). Dry sediment is then heated to 550ºC in a muffle furnace.

Organic matter is oxidized to CO2 and ash during this step (Heiri et al., 2001). Therefore, the weight loss at this step is proportional to the organic matter content of the sample and calculated using equation 1 (LOI550) (Heiri et al., 2001): 109

LOI550= (DW105-DW550/DW105)*100 (Eqn. 1)

The LOI550 is LOI at 550°C (as a percentage), DW105 represents the dry weighting of the sample before combustion, and DW550 is the dry weight of the sample after heating to 550°C (Heiri et al., 2001).

The remaining material is then oxidized at a temperature of 1000ºC. During this step,

CO2 is evolved from carbonates, leaving behind an oxide (Heiri et al., 2001). Weight loss at this step is proportional to the carbonate content of the sample (LOI1000). The initial weight of the sediment and weight loss at each step is measured and used to calculate the organic carbon, the calcium carbonate, CaCO3 and the residual (mineral matter) content using the following equations (Heiri et al., 2001):

LOI1000= (DW550-DW1000/DW105)*100 (Eqn. 2)

LOI1000 is LOI at 1000°C (as a percentage), and DW1000 represents the dry weight of the sample after heating to 950-1000°C (Heiri et al., 2001). Assuming that the only source of carbon is from carbonates, and using a molar weight of CO2 as 44 g and the molar weight of CaCO3 as

100 g, weight percent of CaCO3 in the sample is calculated by multiplying LOI1000 by 2.274.

Finally the percentage of mineral matter in the sample is provided by the following equation

(Eqn 3):

Mineral matter = 100 – (LOI550 + LOI1000 *2.274) (Eqn. 3)

Ostracode analysis

Sediments from the two shallow water cores (E-09 and F-09) were analyzed for the abundance of ostracodes at selected depths based on the percent carbonate content at the Kent

110

State University paleolimnology laboratory. Seventeen samples were selected from each core for a total of 34 samples. Wet sediments samples were washed through a stack of sieves with 150 and 63 micrometer openings. Residues from the 150 µm sieves were decanted into Whirlpack® bags, frozen and freeze-dried before analyzing for the osrtracode abundance. Freeze-dried sediments were then observed and counted under a light microscope.

Geochemical Analysis

Elemental concentrations of lake sediment can be used to reconstruct paleoenvironmental and paleoclimatic conditions. Concentrations of metal elements can be used to understand the redox conditions of the paleoenvironments and, hence, could be used as paleolake-level indicators (Gorham and Swaine, 1965). For instance, enrichment of uranium (U), molybdenum

(Mo), and arsenic (As) imply anoxic conditions. Higher enrichment of detrital elements would indicate periods of higher surface water input. Calcium (Ca) concentration is a commonly used paleoproductivity and paleoclimate proxy. Ca concentrations in lake sediment have been interpreted in contradicting ways in the literature (Brown, 1996; Ricket and Jhonson, 1996;

Finnely et al., 1996). Ca concentrations closely follow the total inorganic carbon (TIC) of the sediment. Generally higher inorganic carbonate precipitation is observed during dry periods due to the concentration of dissolved ions under higher evaporative conditions as well as higher calcium carbonate preservation (Benson et al., 2002; Brown, 2011). In contrast to this observation, enhanced Ca+2 input from ground water recharge during periods of higher rainfall resulted higher carbonate precipitation, in small groundwater-dominated lakes (Shapley and

Donovan 2005). Such seemingly inconsistent results arise when individual proxies do not fully constrain all potential interpretations. Thus it is important to evaluate a suite of indicators; that allows researchers to determine a more unique interpretation for their proxies.

111

X-ray Fluorescence

The bulk elemental compositions of core samples were measured using the ITRAX X-ray

Fluorescence core scanner at the University of Minnesota, Duluth. The scanner has the potential to scan cores at 0.2 mm resolution. Cleland F-09 and E-09 cores were scanned at 5 mm resolution, while the core B-09 was scanned at 0.4 mm resolution. Scaning at a coarser resolution allows a longer dwell time, and thus, a higher signal to noise ratio. The scanner, which utilizes a

Molybdenum X-ray source, was operated at 30 mA and 45 keV over a scan-time of 60 seconds.

Principal Component Analysis (PCA) was performed on stacked XRF data from the B-09, E-09, and F-09 cores. The XRF core scanner provides the relative abundance of bulk elements, in terms of counts per second (cps).

Age Models

The age models explained in Chapter 2 and Chapter 3 were used to infer the ages of the intermediate core and the deep core, respectively. The combined age model prepared by the

Shaw’s method of graphic correlation of the reflectance records (chapter 4) from the deep, intermediate, and shallow cores of Cleland Lake, was used to infer the ages of the shallow core.

Interpretation of Paleohydrologic Conditions

We infer the paleolake levels of Cleland Lake using the relationship between sediment type and depositional environments as published in scientific literature. The following assumptions were made about lake depths and paleodepositional environments based on sediment type. Sediment dominated by well-laminated, organic matter and low calcium carbonate represent a deep lake with high effective moisture (P>E). High water levels prevent mixing of water which leads to permanent stratification creating a cold, anoxic hypolimnion.

112

Low temperatures increase the solubility of CO2 in water. This causes water to be depleted in

- -2 HCO3 and CO3 which decreases the pH of water. Lower pH levels (lower than 8) do not support the preservation of calcium carbonates (Dean and Gorham, 1998; Nelson et al., 2009).

Well-laminated sediments with alternating dark and light layers suggest higher seasonal climatic variations. Dark layers represent organic and inorganic matter deposited during colder seasons

(i.e. winter and autumn) (O’Sullivan, 1983). White layers composed of carbonate represent

-2 periods of higher primary productivity. Primary producers consume CO2 leading to CO3 saturation and CaCO3 precipitation. Physical factors, such as an increase in temperature and evaporation can also cause super saturation with respect to CaCO3 triggering carbonate precipitation (Lowe et al., 1997).

Absence of laminations suggests a low lake level, indicating the absence of permanent stratification. Hence the sediments have been disturbed by mixing or bioturbration. Therefore, massive (unlaminated), carbonate mud with high calcium carbonate percentages, Ca concentrations and low-organic matter would indicate periods of relatively low lake stands

Pompiani et al., 2012).

113

Results

Deep Basin Core )B-09)

Core B-09 was taken at a water depth of 24 m has a total length of 257 cm and consists of two drives. The core extends from 400 cal yr BP to 7600 cal yr BP (post-Mazama sediment). The stratigraphic variation was relatively low in this core which was obtained from far below the thermocline (Figure 5.1 and 5.2). The B-09 core has high organic matter content varying between 40% and 80%. In contrast, it has lower calcium carbonate varying between 5% to 45% and mineral matter content lower than 40% throughout the record (Figure 5.1). Sediment between 257 cm, the core bottom and 197 cm consists of very well-laminated, dark-colored organic mud. Sediment between 177 cm and 145 cm consist of dark brown and light brown well laminated organic mud (gyttja) with 3 cm thick grayish-white carbonate layers at 165 cm and

175 cm. The core is very well-laminated with dark brown, light brown, and white colored sediment triplets between 144 cm and 39 cm. Well-laminated, organic-rich sediment of the core is disturbed between 44.5 cm to 47.5 cm, 60 cm to 64 cm, 66 cm to 69 cm and 133 cm to 136 cm. The former three sections are composed of olive green massive organic mud, while the latter consist of greyish-white laminated carbonate mud.

Calcium carbonate content is low (10% to 25%) in the bottom of the core from 259 cm to

219 cm. Carbonate content fluctuates between low values (5 to 10% ) and high values (45% to

35% ) from 219 cm and 133.5 cm. Carbonate content has an increasing trend from 133 cm to 28 cm with an average value of 30%. The values fluctuate between 40% and 20%. In contrast, organic matter content remains close to the average value (56%) without significant fluctuations throughout the record. Mineral matter content varies between 18% and 25% from 259 cm to103 cm. The percentage remain much lower (5% to 15%) with the exception of 257 cm, 223 cm, 140

114

cm, and 44.5 cm, where higher mineral matter percentages are recorded (42%, 35%, 36% and

62% respectively).

Figure 5.1. Percent organic carbon (LOI 550C), percent carbonate (% CaCO3), percent mineral matter (Min Mat), and dry bulk density in g/cm3(Dry BD) compared with the generalized stratigraphic column (right) of the Cleland Lake deep core (B-09).

115

Figure 5.2. High resolution image of the deep basin core B-09. Left: drive 1; 0.3 to 1.2 m, right: drive 2; 1.3 to 2.5 m.

116

Intermediate-Depth Core )F-09)

The intermediate core was obtained at a water depth of 12 m consist of six overlapping drives. Out of the seventeen samples processed for ostracodes, only four samples recorded their presence (Figure 5.14). The only species present in these samples was assumed to be

Limnocythere friabilis (cf) and need further reference for confirmation (Figure 5.3). The basal section of this 361 cm long core consists of grey to brown clayey silt and fine silt from 361cm to

295 cm (Figure 5.5). These sediment are characterized by high mineral matter (50% to 90%), and low organic matter (<10 %) and low carbonate (< 20 %) (Figure 5.4). Mineral matter is highest

(80 to 85%) below 317 cm and gradually decreases up to 25% by 275 cm.

100 m

Figure 5.3. Limnocythere sp. cf friabilis female right valve (Figure courtesy of Dr. A.J. Smith)

Clastic sediment transitions to well-laminated organic mud at 298 cm and extends up to

275 cm. Laminae (0.5 cm thick) consist of dark brown, organic rich sediment and grayish-white, carbonate mud (marl). Organic content and carbonate content rapidly increase after 300 cm, while the mineral content gradually drops to lower levels. Carbonate content remains high (35% to 50%) from 300 cm to 245 cm, while the organic matter content is relatively low (15% to

117

30%). Organic matter content gradually increases from 12% to 50% (from 243 to 218 cm).

Laminated carbonate transitions to a grayish-white, massive (unlaminated) carbonate mud at 275 cm that extends up to 259 cm. Sediment composition changes to a light brown to white, faintly laminated carbonate mud facies at 259 cm, where the mineral matter content rapidly increases to

50% and remain high up to 243 cm. Faintly laminated sediment transitions to well-laminated carbonate mud at 243 cm and extends up to 227 cm. Calcium carbonate content sharply dropped from 50% to 10% during this transition. Very finely laminated (thickness 2-3 mm) sediment triplets of dark brown, light brown, and white extend from 227 cm to 194 cm. They are separated by the yellowish-brown Mazama tephra layer between 209 cm and 214 cm. Sediment immediately following the tephra layer was depleted in carbonate and organics from 214 to 203 cm and gradually increased after that depth. Sediment between 194 cm and 124.5 cm consists of dark brown and light brown laminated, organic mud with relatively low calcium carbonate (25% to 35%). Calcium carbonate percentages sharply increase (from 15% to 50%) between 123 and

115 cm and remain high during the rest of the record. Sediment composition transitions from greenish-brown to pale-white, well-laminated carbonate mud interbedded with organic mud between 124.5 cm and 110.5 cm. Laminae contacts are less distinct and appeared disturbed or bioturbated between 119 cm and 110.5cm. A finely laminated, dark brown, organic mud and gray carbonate mud layer follows this layer from 110.5 cm to 67 cm. The upper most sediment from 30 cm to 66 cm consists of gray, faintly laminated, calcareous mud.

In summary, mineral matter is high at the base of the core from 361 to 298 cm and remains low (20% to 30%) during the rest of the record (except from 259 to 243, 211 to 203 cm and 67 cm). Organics and carbonate remain low )<20%) at the base of the core from 361 to 298 cm. Organic matter content varies between high (70%) and low (35%) values from 207 cm to 30

118

cm. Calcium carbonate content shows a similar fluctuation between high and low values, varying from 50% to 15%, between 169 cm and 30 cm. High carboante percentage periods are simultaneous with low organic matter content and vice versa. Peak organic matter percentages appear centered around 165 cm, 155 cm to 147 cm, 131 cm to 121 cm, 91 cm and 81 cm (or

4150 cal yr BP, 3700 to 3400 cal yr BP, 2800 to 2450 cal yr BP, 1400 and 1000 cal yr BP respectively [Chapter 2.3.2]). The above depths are characterized by low calcium carbonate percentages. Percent organic matter has minimum values (~35%) around 173 cm, 160 cm, 141 cm, 113 cm (or 4800 cal yr BP, 4000 cal yr BP, 3100 cal yr BP, 2200 cal yr BP and 1300 cal yr

BP respectively).

Wavelet analysis of de-trended organic matter shows ~800 year periodicity between 5000 and 3000 cal yr BP and weak high-frequency periodicities (~250 and 350 year) during the latter part of the record (from 1000 to 100 cal yr BP; Figure 5.6). Calcium carbonate records show 800 year periodicities during mid-Holocene and (5000 to 2000 cal yr BP) and 500 year periodicities during late Holocene (from 3000 to 100 cal yr BP; Figure 5.7).

119

Figure 5.4. Percent organic carbon (% OM), percent carbonate (% CaCO3), percent mineral matter (% Mineral), and dry bulk density in g/cm3 (Dry BD), compared with the generalized stratigraphic column (right) of the Cleland Lake intermediate core, (F-09).

120

Figure 5.5. High resolution images of the core F-09, from left to right, Drive 1; 30-110 cm, Drive 2 59-159 cm, Drive 3; 135-227 cm, Drive 4; 206-293cm, Drive 5;239-321.5 cm and Drive 6 320- 361.5 cm

121

Figure 5.6. (a) Detrended percent organic matter content in the intermediate core during the past 8000 years (b) The wavelet power spectrum for organic matter percentages interpolated to 100 year intervals. Contour levels are chosen so that 75%, 50%, 25%, and 5% of the wavelet power is above each level, respectively. The red dotted line represents a significant periodicity in the organic matter record.

122

Figure 5.7. (a) Detrended percent calcium carbonate content in the intermediate core during the past 8000 years. (b) The wavelet power spectrum for calcium carbonate percentages interpolated to 100 year intervals. The contour levels are chosen so that 75%, 50%, 25%, and 5% of the wavelet power is above each level, respectively. Red dotted lines represent significant periodicities in the carbonate record.

Shallow Core )E-09)

Core E-09 was obtained from a water depth of 7 m and predominantly consists of low stand depositional facies (Figure 5.8 and 5.9). Ostracodes were identified only in three samples out of the seventeen samples (Figure 5.16). The identified genera include Limnocythere,

Candona and Cyprideis.

The base of the core (from 292 cm to 287 cm) consists of whitish-grey to brown, poorly- laminated, organic-rich, carbonate mud (Figure 5.9). Lamina contacts are diffusive and deformed. Poorly-laminated sediment transitions to massive carbonate-mud at 287 cm and extends up to 228 cm. Massive carbonates transition to poorly laminated carbonate mud and grey

123

silty clay from 228 cm to 219 cm. This high-organic matter and low carbonate period (from 219 cm to 203 cm) is characterized by massive to poorly laminated, organic-rich carbonate mud.

Well-laminated carbonate mud associated with rusty brown layers from 203 to 182 cm is preceded by black to dark brown and pink, well-laminated organic-rich carbonate mud from 182 to 175.5 cm. The next 20 cm from 175.5 cm to 151 cm) is characterized by pinkish brown to white, poorly laminated to massive, carbonate mud. Carbonate mud transitions to pink, silty clay at 151 cm and extends up to 142 cm. Clastic sediment transitions to pink and white, laminated, organic-rich, carbonate mud at 142 cm and extends up to 111.5 cm. Laminae at the bottom of this layer (142 cm to 126.5 cm) appear reworked or disturbed. Weekly laminated sediments transition to pink and white, well-laminated, carbonate mud from 96 cm to 89 cm. The top of the core (89 to 30 cm) consists of brown and whitish-brown, weakly to poorly laminated, carbonate mud.

Percent organic matter and mineral matter remain high in the bottom half of the core from 290 cm to 170 cm). In contrast the carbonate percentage was low (20% to 30%) in this section. Mineral matter and organic matter fluctuate between high (50% to 56%) and low values

(30% to 40%) alternatively (i.e organic matter is high where the mineral matter is low and vice versa). Carbonate percentage gradually increases after 184 cm and remain high (60% to 80%) between 122 and 78 cm. Simultaneous with this increase, mineral matter and organic matter gradually decrease after 170 cm and remain low between 120 cm and 80 cm. Carbonate content gradually decreases after 80 cm, while organics and mineral matter gradually increase towards the top of the core (30 cm).

124

Variation of Organic Matter and Calcium Carbonate with Water Depth

The plot of organic matter percentage versus the calcium carbonate percentage displays a linear trend between the three core sites )Figure 5.10, Figure 5.11, and Table 5.1). Core E-09

(shallow water core) has the highest carbonate percentage, while the core B-09 has the lowest carbonate percentage. Carbonate content of the core F-09 (intermediate water depth core) is lower than that of the shallow water core, but higher than that of the deep water core after 8000 cal yr BP. However, the values deviate from the observed linear trend prior to 8000 cal yr BP.

Relatively high carbonate percentages close to the range of carbonate percentage of the shallow core are recorded between 8000 to 11700 cal yr BP (Figure 5.14 and Table 5.1). In contrast, very low organic matter and carbonate percentages are recorded prior to 11700 cal yr BP.

125

Figure 5.8. Percent organic carbon (% OM), percent carbonate (%CaCO3), percent mineral matter (% Mineral), and dry bulk density in g/cm3 (Dry BD) compared with the generalized stratigraphic column of the Cleland Lake shallow core, E-09.

126

Figure 5.9. High Resolution Images of the shallow water core E-09: Drive 1; 30.5-101.5 cm, Drive 2; 94-178 cm, Drive 3; 140-213.5 cm, Drive 4; 196- 292 cm

127

Age (cal yr BP)

Figure 5.10. Down core variation in the percentage of organic matter (blue line) and in the percentage of CaCO3 (red line) by weight, in the intermediate (F-09), shallow core (E-09), and the deep core (B-09).

128

80 E09 70 B09 F09 8000 to 170 cal yr BP 60 F09 11700 to 8000 cal yr BP F09 prior 11700 cal yr BP 50

CaCO3 (wt CaCO3 %) 30

0 0 10 20 30 40 50 60 70 80 90

Organic matter (wt % )

Figure 5.11. Scatter plot of LOI derived carbonate and organic matter percentage for cores B-09, E-09 and F-09.

Table 5.1. Average values of percent carbonate and organic matter for all three sediment cores.

Water depth 170 to 8000 cal yr 8000 to 11700 cal yr 11700 to 14000 Core (m) BP BP (prior 12000) cal yr % OM CO3-2 % OM CO3-2 % OM CO3-2 B-09 20.2 m deep 56.37 23.71

12.8 m F-09 48.43 28.89 24.80 37.01 4.60 16.98 intermediate

7 m E-09 32.55 46.21 45.83 12.25 shallow

129

Figure 5.12. The generalized stratigraphic sections of the shallow (left), intermediate (middle) and deep (right) cores of the Cleland lake. The linear interpolated age scales are given in thousand cal yr BP.

130

Elemental Concentrations

Principal component analysis of XRF data generated five principal components (PC) with eigen values higher than one. These five PCs explain 80.5% of variance in the XRF data. (Table

5.2)

Table 5.2. Initial eigen values and variance explained by the first five principal components of the XRF data.

Initial % of Cumulative Component Eigen Variance % values 1 7.67 33.61 33.61

2 5.46 21.04 54.65

3 1.38 13.63 68.28

4 1.26 6.57 74.84

5 1.17 5.81 80.66

131

Table 5.3. Rotated Component Matrix for the combined XRF data matrix. Significant component loadings (r≥0.7) are given in bold letters.

PC 1 PC 2 PC 3 PC 4

Fe 0.974 Zr 0.965 Ca 0.934 Cl 0.890

K 0.973 Ar 0.729 Sr 0.829 Zn 0.597

Rb 0.964 As 0.562 Mn 0.627 Ti 0.369

Si 0.947 Sr 0.458 Cr 0.452 Sr 0.076

Ti 0.907 Cr 0.444 Ar 0.392 Si 0.053

Ni 0.882 Cu 0.376 Zn 0.218 Fe 0.045

Cu 0.619 Ca 0.253 Cu 0.207 K 0.039

Cr 0.593 Mn 0.128 As 0.170 Rb 0.026

Zn 0.543 Ni 0.107 Zr 0.128 Ca 0.020

Mn 0.538 Zn 0.057 S 0.110 Zr 0.018

S 0.371 Cl 0.019 Ni 0.067 As -0.001

Ar 0.320 S -0.020 Rb 0.055 P -0.011

Al 0.293 Fe -0.034 Al 0.048 Ar -0.018

Ca 0.004 Rb -0.036 Fe 0.033 Mn -0.073

Cl 0.003 Ti -0.081 Si 0.033 Ni -0.078

As 0.000 Si -0.087 K 0.023 Cu -0.088

Zr -0.080 Al -0.094 Ti -0.026 Cr -0.111

Sr -0.146 K -0.117 Cl -0.051 Al -0.131

P -0.193 P -0.411 P -0.520 S -0.171

132

XRF-PC1 shows significant correlations with Si, K, Ti, Fe, Ni, and Rb (Figures 5.13 and

5.14). In the intermediate core, all of these elements show higher concentrations during the deglacial period from 14,300 cal yr BP to 13,000 cal yr BP). Concentrations gradually decrease after that and reach low values around 11800 cal yr BP. These values remain low throughout the

Holocene. Three XRF peaks are prominent in the tephra layers at 11500, 7699 and 470 cal yr BP.

XRF-PC 2 correlates with Mo incoherent to Mo coherent ratio (Moinc/Mocoh) and the Zr concentrations (r= 0.71 and 0.96 respectively). Moinc/Mocoh is a measre of the organic productivity, where higher values represent higher productivity and vice versa. Moinc/ Mocoh show lower levels during deglacial period (prior to 12000 cal. yr BP) and rapidly rises to higher levels. These levels remain high with minor fluctuations (4 – 5.1) throughout the rest of the record. Zr concentrations clearly show three distinct Zr levels. Lowest Zr levels (1000 to 2000 cps) are recorded during the deglacial period (prior to 11600 cal yr BP). Zr levels rapidly rise to

4000 cps by 11500 cal yr BP. Intermediate levels (3000-4000) of Zr are recorded during the early Holocene (from 11500 to 8800 cal yr BP) proceeded by a rapid rise to higher levels at 8700 cal yr BP. High Zr levels are recorded throughout the middle and late Holocene. XRF-PC 3 correlates with the Ca and Sr concentrations (Figure 5.14, 5.16). Both Ca and Sr have low concentrations in the intermediate core during the deglaciation (from 14000 to 13000 cal yr BP;

Figure 5.14). Ca concentrations start to increase around 13,000 cal yr BP and reach the highest values by12,800 cal yr BP. In contrast Sr concentrations do not start to increase untill 12400 cal yr BP and reach the highest values by 11900 cal yr BP. Both Ca and Sr levels rapidly drop to lower levels by 8650 cal yr BP. Mid-Holocene (from 8500 to 3500 cal yr BP) Ca levels remain relatively low compared to early and late Holocene. Late-Holocene (from 3500 to 170 cal yr BP)

Ca levels rapidly fluctuate between high and low values (~100,000 and ~17,000 cps,

133

respectively). In addition to these changes, five high-Ca events are centered around 5100, 3300 to 2900, 2,400 to 1900, 1200 to 1100, 600 and 200 cal yr B.P.

Similarly, in the shallow core, Ca concentrations are relatively high during early

Holocene (from 10,000 to 8100 cal yr BP). Concentrations are low during mid-Holocene (from

8000 to 5500 cal yr BP) and gradually increase to high late-Holocene levels (Figure 2.12).

During the late-Holocene, Ca concentrations fluctuate between low and high values, in a pattern similar to the intermediate core. In contrast, in the deep core, Ca concentrations are very low

(<20,000 cps). Ca concentrations are relatively high during the mid-Holocene (from 8000 to

2600 cal yr BP) and rapidly drop around 2700 cal yr BP. Late Holocene Ca levels remain low in the deep core (Figure 5.16).

134

PC 1 1

-1

1000 Ni 500

500000

200000 0

10000

4000 0

80 20000 60

K 40 10000

20 % % LOI 550C

00 2000

40 Si

201000 %CaCO3

9000

8000

7000

6000

5000

4000

3000 2000

75 1000

14000

13000

12000 11000 10000 Age (Cal yr BP) yr (Cal 50 14007.8BP max.: yr Cal 25

% % Mineral Matter 0 0 2000 4000 6000 8000 10000 12000 14000 age (cal yr BP) max.: 14047.3 cal yr BP

FigureFigure 5.13. Down core variation of percent mineral matter, concentrations of Si, K and Ti, Fe and Ni (cps), as well as XRF-PC1 in core F-09. Note that the Y-axis scale is different from one element to the other, for the clarity of visualization.

135

0 2000 4000 6000 8000 10000 12000 14000 600

400 80 60 20040

Ostracodecounts % LOI 550C 550C LOI % 00

20 %CaCO3 compcomp compcomp4 (bacillariophyceae*0.7)+(phycocyanin*-0.3) 3 25(geothite+phaeophytin-a+peridinin)comp (smectite+chlorite+phycocyanin)(pheophytin-a+allophycocyanin) 1 (sphalerite+illitePC 12345 2) 0 80 60 50000 Ca 40 20

% Mineral % Matter 00

1.5 20000

1.0

Sr (g/cc) 100000.5

Dry bulk density Drybulk 0.0 0 0 2000 4000 6000 8000 10000 12000 14000 age (cal yr BP) max.: 14047.3 cal yr BP 2

0 XRF PC 3 PC XRF

-2

1.5

0.0

Relectance2 PC -1.5 0 2000 4000 6000 8000 10000 12000 14000 age max.: 14007.8

Figure 5.14. Down core variation of reflectance PC 2, XRF-PC 3, as well as Sr and Ca

concentrations (cps), percent CaCO3 and ostracode counts in the core F-09. Note that the Y-axis scale is different from one element to the other, for the clarity of visualization.

136

Depth (cm)

Figure 5.15. Down core variation of Si, K and Ti and Fe concentrations (cps), XRF-PC 1, magnetic susceptibly (MS), and percent mineral matter (mineral m) percentage in core E-09. Grey shading represent the Mazama tephra layer. Note that the Y-axis scale is different from one element to the other, for the clarity of visualization.

137

0 2000 4000 6000 8000 10000 20

1050 50

(%) 530

OM Ostracodecounts

(%) 30 OM 10 0 10 80 80

(%) 40

CaCO3 CaCO3 0 0 60

15000060 M (%) 30

100000

M (%) Mineral 30 Ca 0 50000 Mineral 0 0 40000 age 1800 Sr 200001200

age 1800 600 0 0 Al 1200 90 6002 60

00 30 Al 90 0 XRFPC3 -260 0 1000 2000 3000 4000 5000 6000 7000 8000 9832 Age max.: 9831.7 30 0.50 0 1000 2000 3000 4000 5000 6000 7000 8000 9832 Ref. PCRef.2 -1.5 Age max.: 9831.7 0 2000 4000 6000 8000 10000

Age (cal yr BP) max.: 9995.5 cal yr BP

Figure 5.16. Down core variation of percent CaCO3, Sr and Ca concentrations (cps), XRF-PC 3, reflectance PC 2 (Ref PC 2), and ostracode counts in the core E-09. Note that the Y-axis scale is different from one element to the other, for the clarity of visualization.

138

Discussion

Reconstruction of Paleohydrological Conditions

The preservation of lake-carbonate depends on internal lake processes such as circulation patterns, lake-basin geometry, water chemistry, stratification, seasonality, and temperature changes. The cross plot between organic matter content versus the carbonate content in the three sediment cores of Cleland Lake clearly demonstrate the relationship between the water depth and the carbonate and organic content of the sediment (figure 5.11). Organic matter content is highest in the deep B-09 core (40% to 80%), and the intermediate depth F-09 core, has higher organic content than the shallow E-09 core. In contrast, the shallow core has the highest levels of carbonate (50% to 75%), and the deep core has the lowest levels. This data demonstrates that water depth is inversely correlated with the carbonate content, and is positively correlated with organic matter content. This data can be further used to reconstruct the paleolake levels of

Cleland Lake during the Holocene. XRF data further strengthen this interpretation of the LOI data.

Percent mineral matter content is highest (75% to 85%) during the deglacial period

(between 14000 and 11700 cal yr BP). In contrast, percent carbonate and organic matter are lowest prior to 11700 cal yr BP, indicating lowest lake productivity during this period.

Concentrations of detrital elements (Si, K, Ti, Fe, Ni, and Rb; XRF PC 1) follow the same trend, suggesting higher mineral input to the lake during the deglacial period (Figure 5.13). This may indicate higher inflow of dissolved mineral matter associated with glacial melt water. Higher proportion of mineral matter compared to organic matter, and higher amounts of detrital elements suggest more erosion in the landscape not stabilized by the vegetation (Haworth and

Lund, 1984). The presence of unlaminated, silty sand further supports this interpretation. Higher

139

mineral matter causes turbidity in lake water and masks the precipitation of calcium carbonates.

Concentrations of the detrital minerals remain low (~25%) after 11500 cal yr BP. Low concentrations of detrital material support the fact that the lake has remained a closed basin lake after 11500 cal yr BP, without a significant runoff input to the lake .

Fine sand transitions to alternating dark and pale laminae between 11700 and 10,800 cal yr BP (from 298 cm to 275 cm). Sediment is characterized by high carbonate (from 40 to 45%) but relatively low organic matter (from 20% to 25%) between 11800 and 8600 cal yr BP (225 to

195 cm). The presence of relatively high carbonate compared to organic matter and weakly laminated to massive marl-like sediments suggest a prevailing lake low stand with extreme evaporation (Lowe et al., 1997). This period of high carbonate is also associated with an abrupt increase in Ca and Sr concentrations, further strengthening an interpretation of prevailing low- lake levels (Figure 5.14). Simultaneous increases in dinoflagellate algae and cyanobacteria abundance (explained in Chapter 2), also support a prevailing lake low-stand. The presence of an anomolously high numbers of the ostracode species, Limnocythere sp cf friabilis, which had been also identified in shallow lakes of Canada further supports the interpretation of a P

8700 cal yr BP in the shallow and the intermediate cores of Cleland Lake but is not present at other depths. This anomalous presence further supports the idea of non-linear response of the climate system to the radiative forcing explained in chapter 2. The lack of species diversity further supports indicate that the lake was not a preferable habitat for other species intolerant for dry conditions.

Sediment composition changes to weakly lamented sediment between 9900 and 8700 cal yr BP. Simultaneous with this change, diatom abundance inferred from reflectance PC 4

140

increases, while dinoflagellate algae levels inferred from PC 2 decreases, suggesting higher lake levels. However, the absence of lamination suggests the lake was not permanently stratified or relatively shallow. The drop in calcium carbonate percentage and the XRF-derived Ca counts are not significant. This further supports the idea that this represents a brief wet period, but the lake does not reach high enough water level to maintain permanent anoxia.

Contemporaneous with the shallow water facies of the intermediate core, the shallow core consists of weakly laminated, organic matter-rich, carbonate mud (between 10,000 to 9500 cal year BP). This facies is followed by a massive carbonate mud between 9500 and 8150 cal yr

BP. Absence of laminae in the shallow core during this period supports the absence of permanent stratification and, hence, a shallow lake (Lowe et al., 1997). Higher detrital mineral and organic matter concentrations further suggest that the shoreline was closer to the shallow core site

(Pompiani et al., 2012). Low inorganic carbonate content in the shallow core during this period could be due to the inhibition of carbonate precipitation by higher residual mineral matter

(turbidity) in the littoral zone (Platt and Wright 1991; Benson et al., 2002) and shoreline reworking during a rapid lake level drop (Pompiani et al., 2012). Collectively, evidence from the shallow and intermediate cores suggests the lake level was relatively low between 11800 and

8700±200 cal yr BP.

XRF-derived Ca counts and CaCO3 percentages of the intermediate core rapidly drop from high, early Holocene values to lowest levels between 8700 and 8600 cal yr BP.

Simultaneous with this drop, lake sedimentary facies change from poorly laminated, carbonate- rich sediment to well-laminated, organic-rich sediment. These chemical and sedimentary changes can be interpreted as a rapid onset of a lake high stand. A rapid decrease in dinoflagellate algae abundance derived from reflectance PC 2 also supports a contemporaneous transition from dry to

141

wet conditions (Figure 5.14). In the shallow core, the sediment transitions to poorly laminated, organic-rich, carbonate mud around 7700 cal yr BP, indicating an increase in lake level. The carbonate percentage gradually increases, while organic carbon and residual mineral matter gradually decrease after 7700 cal year BP. However the presence of rusty-brown, ferric hydroxide deposition in the shallow core between 7800 and 7600 cal yr BP suggests oxygenated- bottom water conditions and prevailing low-water levels at the shallow core (Gorham and

Swaine, 1965). Sediment composition transitions to well-laminated carbonate mud after 7200 cal yr BP, suggesting a return to wet conditions and a deep lake (Platt and Wright, 1991; Lowe et al.

1997; Pompiani et al., 2012). Collective evidence from the intermediate and shallow cores suggest that the lake level became high enough to maintain permanent anoxia in the intermediate core by 8500 cal yr BP. However the lake levels were not as high as the present level until 7200 cal yr BP. Considering the position of the hypolimnion, we can infer that the lake level was at least 2 m lower than the present lake level between 8700 and 7200 cal yr BP.

A carbonate-mud and silty-clay clay layer between 8150 and 7700 cal years BP (from

228 to 219 cm) in the shallow core is characterized by a rapid increase in detrital mineral concentrations (Ti, Fe, Zn, and K), magnetic susceptibility, and mineral matter percentage (from

LOI data). Chemical and physical properties of this layer match that of the Mazama tephra layer, in the intermediate core (Figures 5.9 and 5.10). In the intermediate core, this tephra layer is a 4 cm thick, reddish brown-sand layer. However in the shallow core, this region consists of alternating clay and carbonate layers. This suggests reworking of the existing sediment

(carbonate layer, 287- 228 cm) during rapid deposition of the Mazama tephra around 7700 cal yr

BP.

142

Well-laminated, organic-matter facies in the intermediate and the deep cores continued untill 3600 cal yr BP suggesting a stable deep lake (figure 5.8) Laminae vary from dark-drown to light-brown and white suggest, seasonal biologically-mediated carbonate deposition. Carbonate content, remain high (30 to 35%) during these presumed whiting events. Organic matter content varies between 60 to 40% in this section. Presence of well laminated carbonate mud in the shallow core further suggests the development of anoxic conditions around that core site as well.

Well-laminated facies in the intermediate core change to weakly laminated, light-colored sediment between 3600 and 3200 cal yr BP (from 141 cm to 152 cm). This period is simultaneous with high calcium carbonate percentage and Ca concentrations, as well as low organic matter. Simultaneous with this transition, sediment composition changes to well- laminated carbonate mud in the deep core. Therefore, this period can be considered a relatively low-lake stand period or a period of higher temperature, which induces the physical precipitation of carbonates. A transition to a high stand is characterized by well-laminated, organic matter-rich sediment deposited between 3100 and 2600 cal yr BP (from 124 to 141 cm). Alternating carbonate-rich and organic matter-rich layers suggest higher seasonal climatic fluctuations.

During the late Holocene period, after 3000 cal yr BP, percent carbonate, percent organic matter content, and Ca concentrations show regular high frequency variations. Periods of high carbonate

(between 2300 and 2200 cal yr BP, between 2100 and 1800 cal yr BP, between 1300 and 1200 cal yr BP, between 700 and 600 cal yr BP) corroborate with high Ca concentrations, while low- carbonate events have low Ca concentrations but high organic matter contents. This suggests that the above specified high carbonate periods are associated with relatively low lake-water levels that enhance the precipitation of carbonate. During the late Holocene (between 5000 to 2000 cal

143

yr BP) well-laminated, organic mud in the deep core alternate with thin marl layers following the alternating pattern of high and low carbonate in the intermediate core.

Wavelet Analysis

Ca and Sr counts follow the same trend as the reflectance PC 2 and LOI derived CaCO3 percentages during the early Holocene (Figure 5.14). In contrast, during late-Holocene, Ca and

Sr concentrations show higher variations compared to the reflectance PC 2. This can be explained by comparing the wavelet power spectra and the differences in forcing mechanisms between PC 2 and carbonate percentage (Figures 2.4 and 5.4). PC 2 is mainly controlled by the solar insolation, hence it has a low-frequency signal. In contrast, carbonate deposition shows a much high frequency signal with two prominent periodicities; one with relatively low-frequency,

800 year cycles during the mid-Holocene (from 6000 to 4000 cal yr BP) and one with high- frequency, 500-year variability during the late Holocene (from 4000 to 170 cal yr BP).

Resolution of the data set is insufficient to resolve high frequency signals shorter than 100 years.

Evidence of Holocene climate variability in frequencies of 400-500 years and 900 to 1100 years had been reported across North America (Bond et al., 2001; Springer et al., 2008; Viau et al.,

2012). Five-hundred and eight-hundred year periodicities could be associated with the third and second harmonics (480 and 725 years respectively, Springer et al., 2008]) of the North Atlantic ocean IRD periodicity (1450±500 years; [Bond et al., 2001]). Droughts occurring in similar periodicities have been recorded from the eastern United States. It has been suggested that these events are associated with ocean-atmosphere tele-connections in response to solar irradiance changes (Springer et al., 2008).

144

Regional Comparison

Sedimentological evidence from the deep, shallow, and the intermediate cores of the

Cleland Lake suggest a decline in lake levels during the early Holocene. As explained in the

Chapter two, presence of an early-Holocene, dry period is well-documented from paleohydrological reconstructions across south-eastern British Columbia (Mathews and Heusser,

1981; Bartlein et al., 1998; Chase et al., 2008). Conditions shifted to higher lake-levels with higher effective moisture after 8500 cal yr BP. However, variable climatic conditions with both dry and wet intervals were recorded across British Columbia during the early part of the mid

Holocene (7600 to 5700 cal yr BP). Dog Lake, British Columbia as well as Mahoney Lake,

British Columbia had experienced variable lake levels (Lowe et al.,1997; Hallet and Hills, 2006).

Multi-proxy climate reconstruction from Big Lake suggests arid climatic conditions with higher

P>E conditions compared to early Holocene (Bennett et al., 2001). The shallowing of Dog Lake is suggested by the presence of a 3 cm thick clay layer deposited around 5700 cal yr BP (Hallet and Hills, 2006). Simultaneous with this event, a 9 cm thick silty-clay layer exists in the Cleland

Lake shallow core (Figure 5.5 [between 6200 and 5010 cal yr BP), with a sharp discontinuity at

5010 cal yr BP. Similar to the clay layer in the Dog lake, this silty-clay layer is associated with a rapid drop in organic and inorganic carbon. However, sedimentological records from the deep and intermediate cores do not show evidence of a rapid lake-level drop during this period, which places constraints on the absolute drop in lake level to between 2 and 5 m.

Increase in diatom-inferred lake levels from Felker Lake, British Columbia, between

6800 and 2230 cal yr BP (with a record high stand at 5860 cal yr BP), support a wet mid-

Holocene (Galloway et al., 2011). Increases in pollen and spore accumulation rates of Betula and

Salix, (which are currently restricted to swamp-like moist habitats), and Picea and Abies,

145

(currently restricted to wettest parts of the present Interior Douglas-fir forest zone; IDF), suggest a moist climate during mid-Holocene. Presence of a cold, moist climate is further supported by the glacial advancement in the Canadian Cordillera between 7360 and 6450 cal yr BP.

Additional evidence of a mid-Holocene moist period is supported by high lake levels in

Mahoney Lake, British Columbia (Lowe et al., 1997) and Dog Lake, British Columbia, as well as a vegetation transition to wet-closed forests (Hallet and Hills, 2006). Shifts in moisture regimes with frequent droughts and high fire frequency are reported from Dog Lake (Hallet and

Hills, 2006) and Big Lake, British Columbia (Grimm et al., 1992), as well as several other lakes in western Canada between 2400 to 1200 cal yr BP (Laird et al., 2003; Benett et al., 2001;

Cumming et al., 2002). Evidence of high fire frequency is also reported from the Coast and

Cascade Mountains of British Columbia, between 2400 and 1200 cal yr BP (Hallet and Hills,

2006 and references there in). This period of high fire frequency with frequent summer droughts is named as the Fraser Valley Fire period (Hallet and Hills, 2006 and references there in). In

Cleland Lake, presence of well-laminated carbonate mud and peaks in percent calcium carbonate

(from 2500 to 2400 cal yr BP, from 2300 to 2200 cal yr BP, from 2100 to 1800 cal yr BP, from

1300 to 1200 cal yr BP, and from 700 to 600 cal yr BP) suggest low lake levels simultaneous with the regional dry periods. The late-Holocene is identified as a period of considerable variation in pollen accumulation rates and lake levels in Felker Lake. Diatom-based lake levels from Felker Lake report three extremely low lake levels and high salinity events simultaneous with the latter three events of high carbonate deposition in Cleland Lake (from 1910 to 1800, from 1030 to 690 cal yr BP; Galloway et al., 2011). Similar fluctuations in lake levels associated with regional glacial advances are recorded from the Eleanor Lake during late Holocene (2400 cal yr BP, 1400 cal yr BP, and 800 cal yr BP; Gavin et al., 2011). In contrast to the lake-level rise

146

in Cleland Lake, Eleanor Lake, British Columbia records a rapid rise in biogenic silica at 2600 cal yr BP which is interpreted as a period of low lake level (Gavin et al., 2011). Similar to rapid, centennial lake-level fluctuations in Cleland Lake, six short periods of low effective precipitation, alternating with periods of high effective-precipitation have been recorded from

Mahoney Lake during the past 1500 years (Lowe et al.,1997).

Conclusions

Alternation between laminated sediment and non-or weakly laminated sediment as well as, organic mud and carbonate mud, in Cleland Lake during the past 14000 years suggest considerable variation in lake depths over time. Comparison of carbonate and organic matter percentage from three different depths (shallow to deep) demonstrate a positive linear relationship between lake depth and organic matter content. It also demonstrates an inverse relationship between the lake depths and carbonate percentages that can be used to reconstruct the paleolake levels. Elemental concentrations obtained from XRF analysis were used to support these reconstructions. Presence of a silty-clay layer, with higher detrital minerals and mineral matter percentages suggests a higher inflow of dissolved mineral matter associated with glacial melt water from 14000 to 11700 cal yr BP. Laminated to massive, carbonate mud, rich in Ca, Sr, and calcium carbonate (LOI derived) represent a shallow lake between 11800 and 8600 cal yr

BP. In the shallow core, this arid period is represented by weakly laminated to massive, organic matter and detrital matter-rich, carbonate mud, suggesting a shoreline reworking during a rapid lake level drop. A climate transition to wet conditions is suggested by rapid drops in percent

CaCO3 and Ca levels and facies changes to well-laminated, organic-matter rich sediment between 8700 and 8600 cal yr BP. The presence of poorly laminated, oxidized sediment in the

147

shallow core suggests that the period from 8700 to 7600 cal yr BP was a period of lake level transition predominated by a shallow lake. The presence of well-laminated, organic mud supports a stable, deep lake during mid-Holocene (between 7600 and 3600 cal yr BP). Changes in percent carbonate, organic matter, Ca concentrations, and sediment composition reflect regular century scale fluctuations during late Holocene (between 3600 and 170 cal yr BP). Low-lake level periods (from 3600 to 3200 cal yr BP, from 2300 to 2200 cal yr BP, from 2100 to 1800 cal yr BP, from 1300 to 1200 cal yr BP, and from 700 to 600 cal yr BP) alternate with periods of high lake levels. These centennial fluctuations are probably associated with ocean-atmosphere tele-connection to the changes in solar irradiance. In contrast, early-Holocene lake level changes are associated with long-term climate forcing, like solar insolation, and revealed a much longer periodicity of around 5000 years.

148

CHAPTER 6

SYNTHESIS OF RESULTS AND FUTURE DIRECTIONS

Introduction

Accurate modeling of future climatic changes requires adequate paleoclimatic data at sufficient temporal and spatial resolutions to establish the long-term climate baseline. The short instrumental climate record that spans only 100 to 150 years limits the accuracy of the climate models. This problem gives rise to the need to study paleoclimatic proxies from multiple archives. Laminated lake sediment is an excellent archive for studying past climatic changes.

This research focused on developing a new visible derivative spectroscopy based paleoclimatic data set for the Pacific Northwest. Most of the available paleoclimatic reconstructions from the region are based on pollen abundance, chironimid abundance or tree ring data. The climate system of the Pacific Northwest is affected by multiple ocean basins (Pacific, Atlantic, and

Arctic) as well as atmospheric circulation. Despite the growing number of pollen and tree-ring based paleoclimatic reconstructions, the complex climate system of the Pacific Northwest is poorly understood. Therefore another objective of this project was to understand the frequency, distribution and driving forces of pluvial and drought variability in the Pacific Northwest. In order to address these research questions, three major research hypotheses were established.

Each of them are evaluated in the preceding sections.

149

Reconstruction of Paleolake Lake Phytoplankton Abundance

H1.a -Visible derivative spectroscopy can be used as a proxy for lake phytoplankton abundance

H1.b. Phytoplanktom abundance at Cleland Lake should represent regional paleoclimatic changes

Cleland Lake is a small closed basin lake that is hydrologically and geochemically sensitive to precipitation/evaporation (P/E) balance. Visible derivative spectroscopy (VDS) was used to define sediment pigment content based on the shape of the diffuse spectral reflectance

(DSR) of components extracted from the derivative spectrum for each sample. 1751 reflectance measurements were made at a resolution of 0.5 cm, on sediment cores obtained from three different depths (shallow, intermediate and deep) of the lake. Principal component analysis of these 1751 data points derived five components that explained 97% of the variance. Down core variation of the first five principal components of VDS data explains the sediment composition and paleolake productivity during the past 14000 years. PC 1 correlates with illite and sphalerite indicating the lake sedimentary composition and fluvial input. PC 2 consist of dinoflagellate algae pigments and is a proxy for lake dinoflagellate algae abundance. PC 3 to 5 consist of different proportions of photosynthetic pigments, each of which include a component of cyanobacteria. Three different cyanobacteria communities described by PCs 3 to 5 were named as cyanobacteria community 1, 2 and 3. In addition to cyanobacteria pigments, PC 3 also correlates with smectite; an iron rich clay mineral. PC 4 also correlates with diatoms allowing us to use this PC as a proxy for diatom abundance as well.

Phytoplankton composition in temperate lakes tends to follow a succession pattern similar to that of plants on the landscape. Such a trend was observed in the evolution of phytoplankton communities at Cleland Lake. The lake productivity was low and mostly limited

150

to cyanobacteria community 1 during the late Pleistocene deglaciation (prior to 11600 cal yr BP).

Dinoflagellate algae and cyanobacteria community 2 became dominant, while diatom levels were low during the early Holocene (11600 to 8600 cal yr BP).

Diatom and cyanobacteria community 3 gradually increased after 9000 cal yr BP. A rapid decline in diatoms, cyanobacteria and dinoflagellate algae was recorded from 8500 to 7700 cal yr

BP. Lake primary productivity rapidly increased after the deposition of Mazama tephra around

7700 cal yr BP. This includes a rapid rise in diatoms, dinoflagellate algae as well as all three cyanobacteria communities (Chapter 4). Dinoflagellate algae, diatoms and cyanobacteria community 3 remain high until 6800 cal yr BP and gradually decline thereafter. Diatoms, cyanobacteria 1 and 2 dominate the late Holocene (6000 to 170 cal yr BP) Cleland lake phytoplankton composition.

It was identified that the lake primary productivity was driven by long-term climatic forcing, specifically northern hemisphere summer insolation at a scale of several thousands of years. In addition to the long term signal, rapid environmental events such as the great Mazama volcanic eruption and 8.2ka global cooling event affected lake productivity and sediment composition at scales of several hundred years. In addition, diatom abundance covaried with illite confirming the assumption of increased diatom abundance with increased nutrients input to the lake as discussed in chapter 2. Periods of enhanced diatom productivity were recorded during

Middle and Late Holocene (4400 cal yr BP; 3550 cal yr BP; 3150 cal yr BP; 2700 cal yr BP;

1150 cal yr BP; 950 cal yr BP and 600 cal yr BP) simultaneous with neoglacial advances in

Western Canada. Hence these events could be a response to the cold, humid regional climate during those evens. Such information is important to understand the response of aquatic ecosystems and evaluate water quality in response to future climatic changes as well as rapid

151

environmental disasters like volcanic eruptions. For instance, Cleland Lake was dominated by blue-green and dinoflagellate algae during the early Holocene warm period, hence we can expect aquatic ecosystems in similar future anthropogenic conditions to be dominated by such communities with associated degradation of water quality.

Our study suggests that, visible derivative spectroscopy can be applied to qualitatively and quantitatively reconstruct the paleolake productivity based on the abundance of fossil pigments preserved in sediment (Chapter 2). Accuracy of the method will be further evaluated in the future by absolute methods of pigment quantification like high performance liquid chromatography, after extracting pigments from the sediment. However these methods are much expensive, time consuming and require a proper training. Therefore the VDS method can be substituted as an initial method of quantitative and qualitative analysis of paleolake productivity.

Collaborations had also been made to quantify the absolute diatom abundance in the sediment.

Several attempts were made to count the dinoflagellate algae abundance in Cleland lake sediment samples, in collaboration with Brock University, Canada. However due to the sensitivity of dinoflagellates to the sediment storage and transportation methods, successful results were not achieved from these attempts.

Paleohdrologic Conditions and Lake Level variability

H2: Lake level changes depicted from multiple proxies should correlate with existing paleoclimatic reconstructions of the Pacific Northwest.

The phytoplankton composition of temperate lakes is strongly influenced by water chemistry and lake level. Biological communities in drought sensitive lakes are therefore a reflection of water balance changes, such that algal pigment analysis of sediment from these

152

lakes can be used to investigate past hydroclimate variability. Comparison of reconstructed paleohydrological conditions at Cleland Lake corroborates with the available paleohydrological reconstructions from the region confirming the accuracy of the method (Chapter 1, Chapters 2 and 5 discussions). Here we used PC 2 and PC 4 (dinoflagellate and diatom abundances, respectively) to reconstruct the paleohydrological conditions at Cleland Lake during the past

14000 years (Chapter 2). These results were further supported by lake sedimentological, and elemental compositions (Chapters 4 and 5). High-resolution (up to decadal scale) isotope data from the deep core suppport the interpretation of paleohydrological conditions during the past

7500 years based on the VDS data.

The presence of higher detrital minerals and mineral matter percentages suggests a higher inflow of dissolved mineral matter associated with glacial melt water from 14000 to 11700 cal yr

BP. Higher abundance of dinoflagellates between 11600 and 8600 cal yr BP, indicate warm, low lake-level conditions. Low diatom and high cyanobacteria concentrations between 11500 and

9500 cal yr BP further indicate a warm, shallow lake at this time. This interpretation is further strengthened by illite input to the lake as represented by VDS PC 1, and laminated to massive, carbonate mud, rich in Ca, Sr, and calcium carbonate (LOI derived). Dinoflagellate algae pigments decreased rapidly after 8600 cal yr BP, suggesting a transition to wet conditions and higher lake levels. A climate transition to wet conditions is also supported by rapid drops in percent CaCO3 and Ca levels and facies changes to well-laminated, organic-matter rich sediment between 8700 and 8600 cal yr BP. Lake levels increased after 8600 cal yr BP and remained relatively high until present, implying wetter conditions more consistent with the modern climate of the region. Higher diatom abundance and the presence of well-laminated, organic mud further support a stable, deep lake during mid-Holocene (between 7600 and 3600 cal yr BP). The

153

observed long-term linear trend in diatom abundance represents increasing wetness in the region from Early to late Holocene.

The Late Holocene was characterized by high-frequency moisture variability with three rapid increases in diatom abundances centered at 2600 cal yr BP, 1200 cal yr BP, 1000 cal yr BP and 650 cal yr BP and three periods (centered at 2500 cal yr BP, 2100 cal yr BP and 1400 cal yr

BP) of predominantly low diatom abundance (Chapter 3). The amplitude of mid and late

Holocene lake level changes are relatively low compared to that observed during the early

18 Holocene. Increases in diatom abundances were associated with depleted δ Ocarb values, suggesting a groundwater-associated inflow of nutrients to the lake. The largest observed 18O

(hydrologic) depletion occur centered on 2550 cal yr BP indicating a rapid Late Holocene rise in lake levels. Contemporaneous with this depletion, nutrient (P, S) levels increase, suggesting a higher rate of groundwater inflow and mineral delivery. Transition from a high Mn (with low Cu and Ni) concentrations prior 2600 cal yr BP to a low Mn (with high Cu and Ni) levels indicate a transition from a Mid-Holocene (5000 to 2500) relatively shallow lake to a well stratified deep lake during Late Holocene (2600 to 400 cal yr BP). The lake level rise indicated by increases in illite and diatom levels from 1000 to 750 cal yr BP coincide with the Medieval Climatic

Anomaly. This increase is further supported by depletion in 18O values, higher percentage of organic matter and low calcium carbonate percentages.

Forcing Mechanisms of Pluvial and Drought Variability and the Cyclicity (Chapters 2,3 and 5)

H3: Long-term variations of lake levels reconstructed from multiple proxies should represent variations of Holocene ocean-atmosphere circulations.

154

The abrupt shift to low lake levels and relatively dry conditions that characterized the late

Pleistocene-Holocene transition (and the corresponding rapid return to wetter conditions thereafter) were likely driven by non-linear climate responses to orbital forcing, and a resulting amplification of the subtropical, North Pacific high-pressure system, which reduced moist westerly air flow from the Pacific Basin (Figure 6.1). In contrast to these long term variations driven by insolation, changes in percent carbonate, organic matter, Ca concentrations, and sediment composition reflect regular, five hundred year cycles during late Holocene (between

3600 and 170 cal yr BP). Low-lake level periods (from 3600 to 3200 cal yr BP, from 2300 to

2200 cal yr BP, from 2100 to 1800 cal yr BP, from 1300 to 1200 cal yr BP, and from 700 to 600 cal yr BP) alternate with periods of high lake levels. These centennial fluctuations are probably associated with an ocean-atmosphere teleconnection to changes in solar irradiance (Springer et al., 2008). The rapid rise in lake levels around 2600 cal yr BP is simultaneous with the solar irradiance minimum centered around 2,700 cal yr BP and could be linked to the sea surface temperature (SST) increase in Eastern Tropical Pacific Ocean and resulting El-Niño like conditions. Increase in lake level, illite and diatom levels from 1000 to 750 cal yr BP is associated with the prevailing warmer central and western Pacific ocean SST, and associated La

Niña-like conditions and a warm (positive) phase of Pacific decadal oscillation (PDO) (Cobb et al., 2003; Mann et al., 2009; Steinman et al., 2014).

Collectively this evidence suggests that process affecting the SST of the Pacific ocean such as insolation changes on order of several thousand years, solar irradiance changes on the order of several hundred years, and meltwater discharges from the melting of polar ice affected the regional climate and lake productivity of the Pacific Northwest during the past 14000 years.

In the modern climate system, SST is affected by numerous anthropogenic forces like green

155

house warming, as well as melt water fluxes from melting glaciers and polar ice caps. Hence past

analogues similar to this study is crucial to understand the effects of such phenomenon and

successfully model future climatic changes.

Figure 6.12 Schematic diagram showing the locations of the North Pacific high pressure system (H) and the Alutian low pressure system (L) and the direction of moist air masses (black arrows) during normal pressure conditions (top figure) and during strengthened high pressure conditions (bottom figure).Red dot show the approximate location of the study area.

156

REFERENCES

Abbott, M. B., and Stafford Jr, T. W. (1996). Radiocarbon geochemistry of modern and ancient

arctic lake systems, Baffin Island, Canada. Quaternary Research, 45(3), 300-311.

Alley, N. (1976). The palynology and palaeoclimatic significance of a dated core of Holocene

peat, Okanagan Valley, southern British Columbia. Canadian Journal of Earth Sciences,

13(8), 1131-1144.

Alley, R. B., Marotzke, J., Nordhaus, W., Overpeck, J., Peteet, D., Pielke, R., Talley, L. (2003).

Abrupt climate change. Science, 299(5615), 2005-2010.

Anderson, L., Abbott, M. B., Finney, B. P., and Burns, S. J. (2005). Regional atmospheric

circulation change in the north Pacific during the Holocene inferred from lacustrine

carbonate oxygen isotopes, Yukon territory, Canada. Quaternary Research, 64(1), 21-35.

Ayris, P. M., and Delmelle, P. (2012). The immediate environmental effects of tephra emission.

Bulletin of Volcanology, 74(9), 1905-1936.

Barker, P., Roberts, N., Lamb, H., Van der Kaars, S., and Benkaddour, A. (1994). Interpretation

of Holocene lake-level change from diatom assemblages in lake sidi ali, middle atlas,

Morocco. Journal of Paleolimnology, 12(3), 223-234.

Bartlein, P. J., Anderson, K. H., Anderson, P., Edwards, M., Mock, C., Thompson, R. S.,

Whitlock, C. (1998). Paleoclimate simulations for north America over the past 21,000 years

features of the simulated climate and comparisons with paleoenvironmental data.

Quaternary Science Reviews, 17(6-7), 549-585.

157

Battarbee, R. W. (2000). Palaeolimnological approaches to climate change, with special regard

to the biological record. Quaternary Science Reviews, 19(1), 107-124.

Bennett, J. R., Cumming, B. F., Leavitt, P. R., Chiu, M., Smol, J. P., and Szeicz, J. (2001).

Diatom, pollen, and chemical evidence of postglacial climatic change at big lake, south-

central British Columbia, Canada. Quaternary Research, 55(3), 332-343.

Benson, L., Kashgarian, M., Rye, R., Lund, S., Paillet, F., Smoot, J., Lindström, S. (2002).

Holocene multidecadal and multicentennial droughts affecting northern California and

Nevada. Quaternary Science Reviews, 21(4-6), 659-682.

Berger, A., and Loutre, M. (1991). Insolation values for the climate of the last 10 million years.

Quaternary Science Reviews, 10(4), 297-317.

Blaauw, M., 2010. Methods and code for classical age-modeling of radiocarbon sequences.

Quaternary Geochronology5: 512-518.

Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M. N., Showers, W., Bonani, G. (2001).

Persistent solar influence on north Atlantic climate during the Holocene. Science, 294

(5549), 2130-2136.

Boothl, D. B., Troostl, K. G., Clague, J. J., and Waitt, R. B. (2003). The cordilleran ice sheet.

The Quaternary Period in the United States, 1, 17.

Bouchard, F., Pienitz, R., Ortiz, J. D., Francus, P., and Laurion, I. (2013). Palaeolimnological

conditions inferred from fossil diatom assemblages and derivative spectral properties of

sediments in thermokarst ponds of subarctic Quebec, Canada. Boreas, 42 (3), 575-595.

Bradbury, J. P., & Dieterich-Rurup, K. (1993). Holocene diatom paleolimnology of Elk Lake,

Minnesota. Geological Society of America Special Papers, 276, 215-238.

158

Brugam, R. B., McKeever, K., and Kolesa, L. (1998). A diatom-inferred water depth

reconstruction for an upper peninsula, Michigan, lake. Journal of Paleolimnology, 20(3),

267-276.

Chase, M., Bleskie, C., Walker, I. R., Gavin, D. G., and Hu, F. S. (2008). Midge-inferred

Holocene summer temperatures in southeastern British Columbia, Canada.

Palaeogeography, Palaeoclimatology, Palaeoecology, 257(1), 244-259.

Clague, J. J., and James, T. S. (2002). History and isostatic effects of the last ice sheet in

southern British Columbia. Quaternary Science Reviews, 21(1-3), 71-87.

doi:10.1016/S0277-3791(01)00070-1

Clegg, B. F., Kelly, R., Clarke, G. H., Walker, I. R., and Hu, F. S. (2011). Nonlinear response of

summer temperature to Holocene insolation forcing in Alaska. Proceedings of the National

Academy of Sciences, 108(48), 19299-19304.

Cobb, K. M., Charles, C. D., Cheng, H., and Edwards, R. L. (2003). El Nino/Southern oscillation

and tropical Pacific climate during the last millennium. Nature, 424(6946), 271-276.

Cook, E. R., Woodhouse, C. A., Eakin, C. M., Meko, D. M., and Stahle, D. W. (2004). Long-

term aridity changes in the western United States. Science, 306(5698), 1015.

Craig, H. (1961). Isotopic variations in meteoric waters. Science (New York, N.Y.), 133(3465),

1702-1703.

Cumming, B. F., Laird, K. R., Bennett, J. R., Smol, J. P., and Salomon, A. K. (2002). Persistent

millennial-scale shifts in moisture regimes in western Canada during the past six millennia.

Proceedings of the National Academy of Sciences of the United States of America, 99(25),

16117-16121.

159

Das, B., Vinebrooke, R. D., Sanchez-Azofeifa, A., Rivard, B., and Wolfe, A. P. (2005). Inferring

sedimentary chlorophyll concentrations with reflectance spectroscopy: A novel approach to

reconstructing historical changes in the trophic status of mountain lakes. Canadian Journal

of Fisheries and Aquatic Sciences, 62(5), 1067-1078.

Dean, W. E. (1993). Physical properties, mineralogy, and geochemistry of Holocene varved

sediments from Elk Lake, Minnesota. Geological Society of America Special Papers, 276,

135-158.

Dean, W. E., and Gorham, E. (1998). Magnitude and significance of carbon burial in lakes,

reservoirs, and peatlands. Geology, 26(6), 535-538.

Delorme, L. (1971). Freshwater ostracodes of Canada. part V. families Limnocytheridae,

Loxoconchidae. Canadian Journal of Zoology, 49(1), 43-64. deMenocal, P., Ortiz, J., Guilderson, T., Adkins, J., Sarnthein, M., Baker, L., and Yarusinsky, M.

(2000). Abrupt onset and termination of the African humid period:: Rapid climate responses

to gradual insolation forcing. Quaternary Science Reviews, 19(1), 347-361.

Duston. (1986). Water chemistry and sedimentological observations in Littlefield Lake,

Michigan: Implications for lacustrine marl deposition. Environmental Geology and Water

Sciences, 8(4), 229.

Erik T., B. (2011). Lake Malawi's response to “megadrought” terminations: Sedimentary records

of flooding, weathering and erosion. Palaeogeography, Palaeoclimatology, Palaeoecology,

303(1–4), 120-125.

Eberl, D.D., (2003). User guide to RockJock-a program for determining quantitative mineralogy

from X-ray diffraction data. USGS Open File Report OF 03-78. 40 pp.

160

Eberl, D.D. (2004). Quantitative mineralogy of the Yukon River system: variations with

reach and season, and determining sediment provenance. US Geological Survey, Boulder,

Colorado,. 89, 1784–1794.

Fabbro, L. D., and Duivenvoorden, L. J. (2000). A two-part model linking multidimensional

environmental gradients and seasonal succession of phytoplankton assemblages.

Hydrobiologia, 438(1), 13-24.

Flores, L.N. , Barone, R. (1998). Phytoplankton dynamics in two reservoirs with different trophic

state (lake Rosamarina and lake Arancio, Sicily, Italy). Hydrobiologia, 369, 163-178.

Folkoff, M. E., and Meentemeyer, V. (1987). Climatic control of the geography of clay minerals

genesis. Annals of the Association of American Geographers, 77(4), 635-650.

Fritz, S. C., (2008). Deciphering climatic history from lake sediments. Journal of

Paleolimnology, 39(1), 5-16.

Galloway, J. M., Lenny, A. M., and Cumming, B. F. (2011). Hydrological change in the central

interior of British Columbia, Canada: Diatom and pollen evidence of millennial-to-

centennial scale change over the Holocene. Journal of Paleolimnology, 45(2), 183-197.

Gantt, E. (1975). Phycobilisomes: Light-harvesting pigment complexes. Bioscience, 25 (12),781-

788.

Gavin, D. G., Henderson, A. C. G., Westover, K. S., Fritz, S. C., Walker, I. R., Leng, M. J., and

Hu, F. S. (2011). Abrupt Holocene climate change and potential response to solar forcing in

western Canada. Quaternary Science Reviews, 30(9), 1243-1255.

Gorham, E., and Swaine, D. J. (1965). The influence of oxidizing and reducing conditions upon

the distribution of some elements in lake sediments. Limnology and Oceanography, 268-

279.

161

Graham, L., and Wilcox, L. (2000). Algae.–640 pp.

Grimm, E. C., and Jacobson Jr, G. L. (1992). Fossil-pollen evidence for abrupt climate changes

during the past 18 000 years in eastern north America. Climate Dynamics, 6(3-4), 179-184.

Grover, J. P., and Chrzanowski, T. H. (2006). Seasonal dynamics of phytoplankton in two warm

temperate reservoirs: Association of taxonomic composition with temperature. Journal of

Plankton Research, 28(1), 1-17.

Hallett, D. J., and Hills, L. V. (2006). Holocene vegetation dynamics, fire history, lake level and

climate change in the Kootenay Valley, southeastern British Columbia, Canada. Journal of

Paleolimnology, 35(2), 351-371.

Hallett, D., Hills, L., and Clague, J. (1997). New accelerator mass spectrometry radiocarbon ages

for the mazama tephra layer from Kootenay National Park, British Columbia, Canada.

Canadian Journal of Earth Sciences, 34(9), 1202-1209.

Hammarlund, D., Björck, S., Buchardt, B., Israelson, C., and Thomsen, C. T. (2003). Rapid

hydrological changes during the holocene revealed by stable isotope records of lacustrine

carbonates from lake Igelsjön, southern Sweden. Quaternary Science Reviews, 22(2-4), 353-

370.

Haworth, E. Y., and Lund, J. W. (1984). Lake sediments and environmental history. Studies in

palaeolimnology and palaeoecology in honour of Winifred Tutin. Leicester University

Press.

Heiri, O., Lotter, A. F., and Lemcke, G. (2001). Loss on ignition as a method for estimating

organic and carbonate content in sediments: Reproducibility and comparability of results.

Journal of Paleolimnology, 25(1), 101-110.

162

Hickman, M., and Reasoner, M. A. (1994). Diatom responses to late quaternary vegetation and

climate change, and to deposition of two tephras in an alpine and a sub-alpine lake in Yoho

National Park, British Columbia. Journal of Paleolimnology, 11(2), 173-188.

Holzhauser, H., Magny, M., and Zumbuühl, H. J. (2005). Glacier and lake-level variations in

west-central Europe over the last 3500 years. The Holocene, 15(6), 789-801.

Hong, S., Candelone, J., Patterson, C. C., and Boutron, C. F. (1996). History of ancient copper

smelting pollution during roman and medieval times recorded in Greenland ice. Science,

272(5259), 246-249.

Ji, J., Shen, J., Balsam, W., Chen, J., Liu, L., and Liu, X. (2005). Asian monsoon oscillations in

the northeastern Qinghai-Tibet plateau since the late glacial as interpreted from visible

reflectance of Qinghai lake sediments. Earth and Planetary Science Letters, 233(1-2), 61-70.

Kalff, J. (2002). Limnology: Inland water ecosystems, Prentice Hall New Jersey, 592 p.

Kittleman, L. R. (1973). Mineralogy, correlation, and grain-size distributions of mazama tephra

and other postglacial pyroclastic layers, pacific northwest. Geological Society of America

Bulletin, 84(9), 2957-2980.

Knapp, P. A., Soulé, P. T., and Grissino-Mayer, H. D. (2004). Occurrence of sustained droughts

in the interior Pacific Northwest (AD 1733-1980) inferred from tree-ring data. Journal of

Climate, 17(1), 140-150.

Kruk, C., Mazzeo, N., Lacerot, G., and Reynolds, C. (2002). Classification schemes for

phytoplankton: A local validation of a functional approach to the analysis of species

temporal replacement. Journal of Plankton Research, 24(9), 901-912.

Laird, K. R., Cumming, B. F., Wunsam, S., Rusak, J. A., Oglesby, R. J., Fritz, S. C., and Leavitt,

P. R. (2003). Lake sediments record large-scale shifts in moisture regimes across the

163

northern prairies of north america during the past two millennia. Proceedings of the National

Academy of Sciences of the United States of America, 100(5), 2483-2488.

doi:10.1073/pnas.0530193100 [doi]

Leng, M. J., and Marshall, J. D. (2004). Palaeoclimate interpretation of stable isotope data from

lake sediment archives. Quaternary Science Reviews, 23(7-8), 811-831.

Lenz, C., Behrends, T., Jilbert, T., Silveira, M., and Slomp, C. P. (2014). Redox-dependent

changes in manganese speciation in Baltic Sea sediments from the Holocene thermal

maximum: An EXAFS, XANES and LA-ICP-MS study. Chemical Geology, 370, 49-57.

Liu, Z., Colin, C., Huang, W., Le, K. P., Tong, S., Chen, Z., and Trentesaux, A. (2007). Climatic

and tectonic controls on weathering in south China and Indochina Peninsula: Clay

mineralogical and geochemical investigations from the pearl, red, and Mekong drainage

basins. Geochemistry, Geophysics, Geosystems, 8(5)

Lowe, D. J., Green, J. D., Northcote, T. G., and Hall, K. J. (1997). Holocene fluctuations of a

meromictic lake in southern British Columbia. Quaternary Research, 48(1), 100-113.

Mann, M. E., Cane, M. A., Zebiak, S. E., and Clement, A. (2005). Volcanic and solar forcing of

the tropical Pacific over the past 1000 years. Journal of Climate, 18(3), 447-456.

Marchitto, T. M., Muscheler, R., Ortiz, J. D., Carriquiry, J. D., and van Geen, A. (2010).

Dynamical response of the tropical pacific ocean to solar forcing during the early Holocene.

Science (New York, N.Y.), 330(6009), 1378-1381.

Mathewes, R. W., and Heusser, L. E. (1981). A 12 000 year palynological record of temperature

and precipitation trends in southwestern British Columbia. Canadian Journal of Botany,

59(5), 707-710.

164

Mayewski, P. A., Rohling, E. E., Curt Stager, J., Karlén, W., Maasch, K. A., David Meeker, L.,

Steig, E. J. (2004). Holocene climate variability. Quaternary Research, 62(3), 243-255.

McCabe, G. J., Palecki, M. A., and Betancourt, J. L. (2004). Pacific and Atlantic Ocean

influences on multidecadal drought frequency in the United States. Proceedings of the

National Academy of Sciences, 101(12), 4136.

Menounos, B., Osborn, G., Clague, J. J., and Luckman, B. H. (2009). Latest pleistocene and

Holocene glacier fluctuations in western Canada. Quaternary Science Reviews, 28(21),

2049-2074.

Murphy. (2006). Carbonate deposition and facies distribution in a central Michigan marl lake.

Sedimentology, 27(2), 123.

Naeher, S., Gilli, A., North, R. P., Hamann, Y., and Schubert, C. J. (2013). Tracing bottom water

oxygenation with sedimentary Mn/Fe ratios in Lake Zurich, Switzerland. Chemical

Geology, 352, 125-133.

Nara, F., Tani, Y., Soma, Y., Soma, M., Naraoka, H., Watanabe, T., Nakamura, T. (2005).

Response of phytoplankton productivity to climate change recorded by sedimentary

photosynthetic pigments in lake Hovsgol (Mongolia) for the last 23,000 years. Quaternary

International, 136(1), 71-81.

Naselli Flores, L., and Barone, R. (1998). Phytoplankton dynamics in two reservoirs with

different trophic state (Lake Rosamarina and Lake Arancio, Sicily, Italy). Hydrobiologia,

369, 163-178.

Nelson, D. B., Abbott, M. B., Steinman, B., Polissar, P. J., Stansell, N. D., Ortiz, J. D., Riedel, J.

(2011). Drought variability in the Pacific Northwest from a 6000-yr lake sediment record.

Proceedings of the National Academy of Sciences, 108(10), 3870-3875.

165

Nelson, T. O., Coleman, L. J., Green, D. A., and Gupta, R. P. (2009). The dry carbonate process:

Carbon dioxide recovery from power plant flue gas. Energy Procedia, 1(1), 1305-1311.

Nõges, T., Nõges, P., and Laugaste, R. (2003). Water level as the mediator between climate

change and phytoplankton composition in a large shallow temperate lake. Hydrobiologia,

506(1-3), 257-263.

Ortiz, J., Mix, A., Harris, S., and O’Connell, S. (1999). Diffuse spectral reflectance as a proxy

for percent carbonate content in north Atlantic sediments. Paleoceanography, 14(2), 171-

186.

Ortiz, J. D., Polyak, L., Grebmeier, J. M., Darby, D., Eberl, D. D., Naidu, S., and Nof, D. (2009).

Provenance of Holocene sediment on the Chukchi-Alaskan margin based on combined

diffuse spectral reflectance and quantitative X-ray diffraction analysis. Global and Planetary

Change, 68(1-2), 73-84.

Ortiz, J. D. (2011). Application of Visible/near infrared derivative spectroscopy to Arctic

paleoceanography. IOP Conference Series: Earth and Environmental Science, 14(1) 012011.

O'sullivan, P. (1983). Annually-laminated lake sediments and the study of Quaternary

environmental changes—a review. Quaternary Science Reviews, 1(4), 245-313.

Overpeck, J. T., and Cole, J. E. (2006). Abrupt change in earth's climate system. Annual Review

of Environment and Resources, 31(1), 1.

Palmer, S., Walker, I., Heinrichs, M., and Scudder, G. (2002). Postglacial midge community

change and Holocene palaeotemperature reconstructions near treeline, southern British

Columbia (Canada). Journal of Paleolimnology, 28(4), 469-490.

166

Pellatt, M. G., Mathewes, R. W., and Clague, J. J. (2002). Implications of a late-glacial pollen

record for the glacial and climatic history of the Fraser lowland, British Columbia.

Palaeogeography, Palaeoclimatology, Palaeoecology, 180(1), 147-157.

Pellatt, M., Hebda, R., and Mathewes, R. (2001). High-resolution Holocene vegetation history

and climate from hole 1034B, ODP leg 169S, Saanich inlet, Canada. Marine Geology,

174(1), 211-222.

Peteet, D., Beck, W., Ortiz, J., O’Connell, S., Kurdyla, D., and Mann, D. (2003). Rapid

vegetational and sediment change from Rano Aroi crater, Easter Island. Easter island (pp.

81-92) Springer.

Platt, N., and Wright, V. (1991). Lacustrine carbonates: Facies models, facies distributions and

hydrocarbon aspects. Lacustrine Facies Analysis, , 57–74.

Pompeani, D., Steinman, B., and Abbott, M. (2012). A sedimentary and geochemical record of

water-level changes from Rantin Lake, Yukon, Canada. Journal of Paleolimnology, 48(1),

147-158.

Reimer, P. J., Baillie, M. G., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Edwards, R. L.

(2009). IntCal09 and Marine09 radiocarbon age calibration curves, 0-50,000 years cal BP.

Renssen, H., Seppä, H., Heiri, O., Roche, D., Goosse, H., and Fichefet, T. (2009). The spatial

and temporal complexity of the Holocene thermal maximum. Nature Geoscience, 2(6), 411-

414.

Rhoton, F., Smeck, N., and Wilding, L. (1979). Preferential clay mineral erosion from

watersheds in the Maumee River basin. Journal of Environmental Quality, 8(4), 547-550.

Ricketts, R., and Johnson, T. (1996). Early holocene changes in lake level and productivity in

Lake Malawi as interpreted from oxygen and carbon isotopic measurements of authigenic

167

carbonates. The Limnology, Climatology and Paleoclimatology of the East African Lakes, ,

475–93.

Robertson, P. K., Lawton, L. A., and Cornish, B. J. (1999). The involvement of phycocyanin

pigment in the photodecomposition of the cyanobacterial toxin, microcystin-LR. Journal of

Porphyrins and Phthalocyanines, 3(07), 544-551.

Rosenberg, S. M., Walker, I. R., Mathewes, R. W., and Hallett, D. J. (2004). Midge-inferred

Holocene climate history of two subalpine lakes in southern British Columbia, Canada. The

Holocene, 14(2), 258-271.

Ruffell, A., McKinley, J. M., and Worden, R. H. (2002). Comparison of clay mineral

stratigraphy to other proxy palaeoclimate indicators in the mesozoic of NW europe.

Philosophical Transactions of the Royal Society of London. Series A: Mathematical,

Physical and Engineering Sciences, 360(1793), 675-693.

Sanger, J. E., and Crowl, G. (1979). Fossil pigments as a guide to the paleolimnology of Browns

Lake, Ohio. Quaternary Research, 11(3), 342-352.

Sanger, J. E., and Gorham, E. (1972). Stratigraphy of fossil pigments as a guide to the postglacial

history of Kirchner Marsh, Minnesota. Limnology Oceanography, 17(6), 840-854.

Schagerl, M., and Donabaum, K. (2003). Patterns of major photosynthetic pigments in

freshwater algae.-1. cyanoprokaryota, rhodophyta and cryptophyta. Annales De Limnologie,

, 39(1) 35-47.

Schagerl, M., Pichler, C., and Donabaum, K. (2003). Patterns of major photosynthetic pigments

in freshwater algae.-2. dinophyta, euglenophyta, chlorophyceae and charales. Annales De

Limnologie, , 39(1) 49-62.

168

Schnurrenberger, D., Russell, J., and Kelts, K. (2003). Classification of lacustrine sediments

based on sedimentary components. Journal of Paleolimnology, 29(2; 2), 141-154.

Shapley, M. D., Ito, E., and Donovan, J. J. (2008). Isotopic evolution and climate paleorecords:

Modeling boundary effects in groundwater-dominated lakes. Journal of Paleolimnology,

39(1), 17-33.

Shinker, J. J., and Bartlein, P. J. (2010). Spatial variations of effective moisture in the western

United States. Geophysical Research Letters, 37(2)

Springer, G. S., Rowe, H. D., Hardt, B., Edwards, R. L., and Cheng, H. (2008). Solar forcing of

Holocene droughts in a stalagmite record from west Virginia in east-central north America.

Geophysical Research Letters, 35(17), L17703.

Stahl, K., Moore, R. D., and Mckendry, I. G. (2006). The role of synoptic‐scale circulation in the

linkage between large‐scale ocean–atmosphere indices and winter surface climate in British

Columbia, Canada. International Journal of Climatology, 26(4), 541-560.

Steinman, B. A., Rosenmeier, M. F., Abbott, M. B., and Bain, D. J. (2010) a. The isotopic and

hydrologic response of small, closed-basin lakes to climate forcing from predictive models:

Application to paleoclimate studies in the upper Columbia river basin. Limnology and

Oceanography, 55(6), 2131-2145.

Steinman, B. A., Rosenmeier, M. F., Abbott, M. B., (2010) b. The isotopic and hydrologic

response of small, closed-basin lakes to climate forcing from predictive models: Simulations

of stochastic and mean-state precipitation variations. Limnology and Oceanography, 55(6),

2246-2261.

169

Steinman, B. A., Abbott, M. B., Mann, M. E., Stansell, N. D., Finney, B. P., (2012). 1500 year

quantitative reconstruction of winter precipitation in the Pacific Northwest. P. Natl. Acad.

Sci. USA 109, 11619–11623.

Steinman, B. A., Abbott, M. B., (2013). Isotopic and hydrologic responses of small, closed lakes

to climate variability: Hydroclimate reconstructions from lake sediment oxygen isotope

records and mass balance models. Geochimica et Cosmochimica Acta, 105 (2013) 342–359.

Steinman, B. A., M. B. Abbott, M. E. Mann, J.D. Ortiz, S. Feng, D. P. Pompeani, N. D. Stansell,

L. Anderson, B. P. Finney, B. W. Bird (2014). Ocean-atmosphere forcing of centennial

hydroclimate variability in the Pacific Northwest. Geophysical Research Letters.

Stone, J. R., Fritz, S. C., (2006). Multidecadal drought and Holocene climate instability in the

rocky mountains. Geology, 34(5), 409-412.

Strahler, A. N. (1969). Introduction to physical geography.John Wiley and Sons Inc, 457 p.

Sze, P. (1993). A biology of the algae Wm. C. Brown Publishers.

Talbot, M. (1990). A review of the palaeohydrological interpretation of carbon and oxygen

isotopic ratios in primary lacustrine carbonates. Chemical Geology: Isotope Geoscience

Section, 80(4), 261-279.

Telford, R. J., Barker, P., Metcalfe, S., and Newton, A. (2004). Lacustrine responses to tephra

deposition: Examples from Mexico. Quaternary Science Reviews, 23(23), 2337-2353.

Toepel, J., Langner, U., and Wilhelm, C. (2005). Combination of flow cytometry and single cell

absorption spectroscopy to study the phytoplankton structure and to calculate the chlrophyll-

a specific absorption coefficients at the taxon level1. Journal of Phycology, 41(6), 1099-

1109.

170

Tribovillard, N., Algeo, T. J., Lyons, T., and Riboulleau, A. (2006). Trace metals as paleoredox

and paleoproductivity proxies: An update. Chemical Geology, 232(1), 12-32.

Tucker, M. E., Wright, V. P., and Dickson, J. A. (1990). Carbonate sedimentology. Oxford

[England] ; Boston : Blackwell Scientific Publications ; 482p.

Viau, A. E., Gajewski, K., Fines, P., Atkinson, D. E., and Sawada, M. C. (2002). Widespread

evidence of 1500 yr climate variability in north America during the past 14 000 yr. Geology,

30(5), 455.

Walker, I. R., and Pellatt, M. G. (2003). Climate change in coastal British Columbia-a

paleoenvironmental perspective. Canadian Water Resources Journal, 28(4), 531-566.

Whitlock, C., Dean, W. E., Fritz, S. C., Stevens, L. R., Stone, J. R., Power, M. J., Bracht-Flyr, B.

B. (2012). Holocene seasonal variability inferred from multiple proxy records from crevice

lake, yellowstone national park, USA. Palaeogeography, Palaeoclimatology, Palaeoecology,

331, 90-103.

Wolfe, A. P., Vinebrooke, R. D., Michelutti, N., Rivard, B., and Das, B. (2006). Experimental

calibration of lake-sediment spectral reflectance to chlorophyll a concentrations:

Methodology and paleolimnological validation. Journal of Paleolimnology, 36(1), 91-100.

171

APPENDICES

APPENDIX 1

Reflectance data and results from principal component analysis of visible derivative spectroscopic data

0.2 A 0.15

0.1 C 09- 1 0.05 C 09-2 C 09-3 0

C 09-4 % reflectance reflectance % -0.05 B 09-1 B 09-2 -0.1 B 09-3 -0.15 B 09-4 400 450 500 550 600 650 700 Wavelength (nm)

0.1 B 0.08 0.06 0.04 (C 09-1)/2 0.02 (C 09-2)/2 0 (C 09-3)/2

% reflectance reflectance % -0.02 (C 09-4)/2 B 09-1 -0.04 B 09-2 -0.06 B 09-3 -0.08 B 09-4 400 450 500 550 600 650 700 Wavelength (nm) Appendix 1.1. A) Comparison of the reflectance of four dry sediment samples (C/09) with the reflectance of four wet sediment samples from the deep core B09. B) Adjusted reflectance of the

173

dry sediments (C09-1, C09-2, C09-3, C09-4) by dividing from a factor of two. C09-1, C09-2, C09-3, and C09-4cm represent powdered dry sediments obtained from 16.5, 17.5, and 18.5 cm below the sediment water surface respectively. B09-1, B09-2, B09-3 and B09-4 represent wet sediment samples obtained from depths of 41 cm, 41.5 cm, 42 cm and 42.5 cm respectively from the sediment water surface.

174

Appendix 1.2. A) Spectral plot of percent reflectance of the core F09. Legend to the right shows the range of reflectance values recoded from all the samples between the wavelengths 400 to 700 nm. B) Contour plot showing the first derivatives of the reflectance measurement of the core F- 09. Legend to the right shows the range of values recoded from all the samples between the wavelengths 400 to 700 nm (0.21 to -0.71). Higher derivative values between the depths of 230 to 290 cm (8,800 to 10,400 cal yr BP) and 520 to 570 nm correspond to the reflectance peak of

175

peridinin in PC 2 and represent early Holocene peak in dinoflagaellate algae. Lower derivative values between 650 and 660 nm and higher values between 680 to 700 nm are characteristic to minimum and peak reflectance values characteristic to bacillariophyceae pigments, corresponding to diatom abundance after 230 cm (8,800 Cal yr BP).

176

APPENDIX 2

Core Logs

Legend

well laminated, carbonate mud

weakly laminated or massive, carbonate mud

well laminated organic mud

weakly laminated organic mud

silt or clay

volcanic tephra

177

Appendix 2.1. B09- Core log 0.3 to 1.2 m

39-44 cm: 10YR 3/2, dark brown and light brown, finely laminated, organic mud 44-44.5 cm: volcanic tephra 44.5-47.5: greenish brown, massive carbonate mud

47.5- 60 cm 10YR .5/3, dark olive green and 10YR 4/2 light olive green, finely laminated organic mud

60-64 cm: 10YR 3/5 olive green, massive, organic mud (gyttja)

64-66 cm: 10YR 3/2 dark brown and light brown, finely laminated organic mud 66-69 cm: 7.5YR 3/3 olive green, massive, organic mud (gyttja)

69-89 cm: 7.5YR 3/4 dark olive green and 10YR 5/1 white, well laminated organic mud (65 %),

89-133cm: 7.5YR 3/4 Dark drown, 7.5YR 2.5/1 olive green and 10YR 5/1 grey, very finely laminated organic (55%) mud

178

B09- Core log 1.3 to 2.5 m

133-136 cm : 10YR 5/1 white and greenish white, well laminated carbonate mud 136-147 cm: 7.5YR 5/2 dark drown and 7.5YR 3/2 black finely laminated organic mud (gyttja) 144-147 cm: 10YR 4/2 greyish white, well laminated carbonate mud

147- 165 cm: 7.5YR 5/2 dark drown and 7.5YR 3/2 black finely laminated organic mud (gyttja)

165-168 cm: 10YR 3/3 whitish grey, well laminated, carbonate mud 168-175 cm: 7.5YR 5/2 dark drown and 10YR 4/3 light brown finely laminated organic mud 175-177 cm: 10YR 4/2 greyish white, well laminated carbonate (45%) mud

177-190 cm: 7.5YR 2.5/2 brownish green and 7.5YR 3/3 olive green finely laminated organic (65%) mud

190-212 cm: 10YR 3/3 greenish brown and 10YR 2.5/2 reddish brown well laminated organic (60%) mud

212 to 250 cm: 10YR 2.5/2 reddish brown and 7.5YR 2.5/1 olive green well laminated organic (60%) mud

179

Appendix 2.2 F-09 Core log 30-124.5 cm

30-66 cm: 7.5YR 5/2 grey, faintly laminated, calcareous (40%) mud

66-67 cm: Mount St. Helens ash

67-84 cm: 7.5YR 5/2 grey and dark brown, 7.5YR 3/6 weakly laminated, organic rich (40%) carbonate mud

84-110.5 cm: 7.5YR 4/4 dark brown to 7.5YR 5/2 grey, finely laminated, carbonate rich (35%) organic mud

110.5-119cm: brown to 7.5YR 3/3 olive green, well laminated, organic rich carbonate mud

119-124.5 cm: 10YR 3/2 greenish brown to 10YR 5/2 white, well laminated. carbonate mud with bitoturbrated contacts

180

F-09 core log 125-230 cm

124.5-141 cm: 7.5YR 2.5/2 Well laminated blackish brown to 7.5YR 3/5 brown organic mud

141-152 cm: 10YR 4/1 light brown to 10YR 4/6 greyish brown laminated carbonate rich organic mud

152-162 cm: 7.5YR 4/4 dark brown and 10YR 4/2 light brown, well laminated (3-5 mm)organic mud

162-194 cm: 7.5YR 4/4 dark brown, 10YR 4/2 light brown and 10YR 5/2 white finely laminated carbonate rich (30%) organic mud

194-209 cm: 7.5YR 4/4 Dark brown, 10YR 4/2 light brown to 10YR 5/2 white, finely laminated (2-3mm) (reworked), carbonate rich (25%) organic mud (20%)

209-214 cm: brown, fine sand (Mazama tephra)

214-227 cm: 7.5YR 4/4 Dark brown to 10YR 4/2 light brown well laminated organic mud

181

F09-core log 230- 360 cm

227-243 cm: 7.5YR 4/4 Dark brown and 10YR 4/2 light brown, to 7.5YR 8/2 white laminated, carbonate mud

243-256 cm: 7.5YR 8/2 light brown to white, faintly laminated, carbonate mud

256-275 cm: 10YR 7/1 greyish white, massive, carbonate (50%) mud

275-290 cm: 10YR 6/2 brown to grayish white, poorly to well laminated, carbonate mud

290-291 cm: 10YR 5/6 yellowish brown, fine sand (tephra)

291-298 cm: Brown and 10YR 6/2 grayish white, laminated, carbonate mud

298-312 cm: 10YR 4/6 grey, massive clayey silt

312-330.5 cm: 1oYr 4/1 rusty brown to grey, clayey silt

330.5-360 cm: 10YR 5/1 dark grey, massive, silty clay

182

Appendix 2.3 E-09 Core Log: 30-157.5 cm

30-44 cm: 7.5YR 5/4 light brown and 10YR 7/1 brownish-white poorly laminated, carbonate mud

44-89 cm: 7.5YR 5/4 brown, 10YR 7/1 whitish-brown and 10YR 4/6 grey, well to weakly laminated carbonate mud

89-96 cm: 10YR 4/6 greenish-brown and 10YR 7/1 brown, well laminated, organic matter rich carbonate mud 96-111.5 cm: 10YR 8/4 pinkish-brown to grey, weekly laminated, carbonate mud

111.5-126.5 cm: 10YR 8/4 rosybrown well laminated, carbonate rich, organic mud

126.5-148 cm: laminated, organic rich, carbonate mud with disturbed/reworked contacts

148-157.5 cm: 10YR 6/2 grayish-white, massive, silty clay (unconfirmity at 148 cm)

183

E-09 Core Log 157.5-292 cm

157.5-166.5 cm: 10YR 8/2 pinkish grey and 10YR 6/2 white, weekly laminated, carbonate mud

166.5-175.5 cm: yellowish brown, poorly laminated to massive, carbonate mud 175.5-182 cm: 7.5YR 5/2 black, 7.5YR 4/4 dark brown , and 7.5YR 5/4 light brown, well- laminated, carbonate rich, organic mud

182-203 cm: 10YR 4/6 grey and 10YR 8/2 pinkish-grey, well laminated, carbonate mud, with Fe stained fractures

203-219 cm: 10YR 8/pink to 10YR 6/2 white, poorly laminated to massive, carbonate mud

219-228 cm: 10YR 4/6 grey, laminated, carbonate mud and silty clay

228-239 cm: 10YR 6/2 grayish-white, massive, carbonate mud

239-287 cm: 10YR 6/2 grayish-white, massive carbonate mud , speckled with sand size marl

287-292cm: 10YR 4/6 Grey to brown poorly laminated organic mater rich carbonate mud

184

APPENDIX 3

Elemental concentrations from scanning XRF

Si P S Ca Ti Cr Mn Fe Ni Cu Zn

0 100 200 0 50 100 100 400 0 110000 0 500 1000 100 200 300 400 0 3000 6000 5000 40000 0 200 500 0 400 800 0 200 500 0

1000

2000

3000 (cal yr BP) yr (cal

Age 4000

5000

6000

7000

PANGAEA/PanPlot max.: 14146.3 cal yr BP Appendix 3.1. Variation of major and trace metal concentrations in the Cleland Lake intermediate core. Note that the x axis scales are different from one metal to the other for the clarity of visualization. Concentrations are in counts per seconds.

185

Si P S Ca Ti Cr Mn Fe Ni Cu Zn

0 300 0 50 100 50 200 400 0 200000 0 350 700 0 400 0 1000 4000 0 30000 75000 0 200 400 0 400 800 0 500 1000 0

1000

2000

3000

(Calyr BP) (Calyr Age

4000

5000

6000

7000

Appendixmax.: 9996 Calyr BP3.2. Variation of major and trace metal concentrations in the Cleland Lake shallow core. Note that the x axis scales are different from one metal to the other for the clarity of visualization. Concentrations are in counts per seconds.

186