Aeolian Sediments of the Mcmurdo Dry Valleys, Antarctica a Thesis Presented in Partial Fulfillment of the Requirements for the D

Home , Aeolian landform, Beacon Supergroup, Bull Pass, Kukri Hills, Lake Brownworth, Lake Chad (Antarctica)

Aeolian Sediments of the McMurdo Dry Valleys, Antarctica

A Thesis Presented in Partial Fulfillment of the Requirements for The Degree Master of Science in the Graduate School of The Ohio State University

Kelly Marie Deuerling, B.S.

Graduate Program in Geological Sciences

The Ohio State University

2010

Master‘s Examination Committee: Dr. W. Berry Lyons, Advisor Dr. Michael Barton Dr. Garry D. McKenzie

Kelly Marie Deuerling

2010

ABSTRACT

The role of dust has become a topic of increasing interest in the interface between climate and geological/ecological sciences. Dust emitted from major sources, the majority of which are desert regions in the Northern Hemisphere, is transported via suspension in global wind systems and incorporated into the biogeochemical cycles of the ecosystems where it is ultimately deposited. While emissions within the McMurdo Dry

Valleys (MDV) region of Antarctica are small compared to other source regions, the redistribution of new, reactive material by wind may be important to sustaining life in the ecosystem.

The interaction of the dry, warm foehn winds and the cool, moist coastal breezes

―recycles‖ soil particles throughout the landscape. The bulk of sediment movement occurs during foehn events in the winter that redistribute material throughout the MDV.

To understand the source and transfer of this material samples were collected early in the austral summer (November 2008) prior to the initiation of extensive ice melt from glacial and lake surfaces, aeolian landforms, and elevated sediment traps. These were preserved and processed for grain size distribution and major element composition at the sand and silt particle sizes. Major elemental oxide analysis indicated that the silt and sand size particles are of different composition: SiO2 values for silt range from 50 to 59% by weight and for sand range from 59 to 74%. When compared to the elemental oxide composition four rock types present in the MDV, the composition of the silt indicates a

ii mixing influenced mostly by the igneous rock types (Ferrar Dolerite and McMurdo

Volcanic basanite) and sand a mixing influenced largely by the sedimentary rocks

(Beacon Sandstone and the metasedimentary Basement Complex). This could imply a local source of the aeolian material that is corroborated by low CIA values at both particle sizes (44-57%) indicating low degrees of chemical weathering. In addition, comparison of 87Sr/86Sr and 143Nd/144Nd to values published for the major MDV rock types and ice core dust to values analyzed in 3 silt size glacier sample and one bulk glacier sample also indicates a local source of sediments and that it is not likely to be transferred inland.

During the melt season, the aeolian material is actively solubilized where it interacts with water, releasing solutes and vital bioavailable nutrients throughout the aquatic system. Differences in the chemistry of supra- and proglacial streams as well as lake surface waters may be derived from the deposition and dissolution of these aeolian sediments. To simulate these conditions, a two-step leaching method using deionized water to represent glacial melt in field conditions was employed and leachates analyzed for major ion and nutrient constituents. Leachates represent a small degree (<0.7%) of dissolution of major elements, and are solubilized to a greater extent from samples closer to the coast or with increased silt content. The composition of the leachates reflects the dissolution of the major salts found in the MDV. Leach 1 (cold water) indicates that Na- and Cl-bearing salt phases are dissolved to a greater extent than seen in Leach 2 (freeze- thaw). Conversely, Leach 2 compositions indicate that carbonate mineral dissolution and

Mg-bearing silicate weathering are proceeding to a greater extent than in Leach 1.

iii

Inorganic N:P ratios follow the same patterns of nutrient limitations based on the

Redfield Ratio found by Priscu (1995) in the terminal lakes of the Taylor Valley: N- limited in the Fryxell and Hoare basins (east) and P-limited in the Bonney basin (west).

This is also consistent with the age of the tills in the area, as found by Gudding (2003).

The concentration of soluble Fe in the leachates is about the same as soluble inorganic P, and thus is not a limiting nutrient in the leachates. Comparison of total dissolved N and P to their inorganic counterparts reveals increased organic nutrients in the glacier and lake leachates that may indicate the influence of biota. Nutrient fluxes based on known sediment fluxes from elevated sediment traps deployed throughout the MDV and the composition of these leachates range from 0.34-330 g a-1 for N, 0.02-8.3 g a-1 for P, and

0.03-8.6 g a-1 for Fe. These are at least two orders of magnitude less than calculated loads from streams to the lakes in the Taylor Valley and, thus, should be considered underestimations or minima.

This work provides the first investigation into the composition and source of aeolian transported materials in the MDV, as well of what is potentially solubilized from it during the austral summer melt season. In addition, it will contribute to the understanding of the interplay between aeolian and aquatic processes in the MDV and further the understanding of this unique ecosystem.

For my parents.

ACKNOWLEDGEMENTS

I would like to thank my advisor, Dr. Berry Lyons, for his guidance and support throughout the course of my two years at OSU. You have provided me with many opportunities to learn and grow as a scientist and encouraged me to take advantage of opportunities as they avail themselves. You sent me to the Ice twice, which is more than a person could hope for, and an opportunity for which I am forever grateful.

Thank you to Kathy Welch for the miles, conversations, unit conversions, and IC analyses. To Sue Welch, thank you for the nutrient analyses and rundowns, fun times, and re-introduction to swimming. Steve Goldsmith taught me everything there is to know about the XRF and helped me get started. All three of you have been a great support system.

Thank you to Hassan Basagic, Dr. Liz Bagshaw, and Dr. Martyn Tranter for letting me tag along to collect samples of the glaciers. To Rae Spain and Sandra Liu – thanks for putting up with my FNG-y ways. Sandra, I will always remember our attempt on the Matterhorn.

To Deb, JD, and Terra: thanks for the support and being great friends; Rich for the love, laughs, and support; my family for always being my biggest fans. And finally,

Annette Trierweiler: your puns, conversation, and support have meant the world to me.

VITA

November 4, 1985 Born – Tallahassee, Florida

May 2008 B.S. Geology, University of Florida

September 2008-2009 University Fellow, The Ohio State University

September 2009-2010 National Science Foundation Graduate Research Fellow,

The Ohio State University

October-December 2010 Graduate Research Associate, The Ohio State University

FIELD OF STUDY

Major Field: Geological Sciences Environmental Geochemistry

vii

TABLE OF CONTENTS

Abstract ...... ii

Acknowledgements ...... vi

Vita ...... vii

Field of Study ...... vii

List of Figures ...... x

List of Tables ...... xiii

Dust in the Global Scale ...... 1

Geochemistry of Aeolian Sediments of the McMurdo Dry Valleys, Antarctica ...... 9 Introduction ...... 9 Field Area Description ...... 15 Methods ...... 23 Results ...... 29 Discussion ...... 34 Conclusions ...... 47

Experimental Leaching of Aeolian Sediments, McMurdo Dry Valleys, Antarctica ...... 69 Introduction ...... 69 Methods ...... 75 Results ...... 78 Discussion ...... 82 Conclusions ...... 94

Conclusions and Future Work ...... 125

Appendix A: Sample site location, description, and particle distribution...... 145

viii

Appendix B: Sand XRF and LOI data in weight percent major elemental oxide...... 149

Appendix C: Silt XRF and LOI data in weight percent major elemental oxide...... 153

Appendix D: ANOVA results for solid samples...... 155

Appendix E: Leach 1 (cold water) water chemistry – major ions and nutrients...... 157

Appendix F: Leach 2 (freeze-thaw) water chemistry – major ions and nutrients...... 159

Appendix G: Estimated bulk composition of aeolian sediments based on grain size distribution and composition of the specific sites used in the experimental leaching methods...... 161

Appendix H: Surface area of lakes and glacier ablation areas of Taylor Valley used in calculation of solute and nutrient flux...... 163

LIST OF FIGURES

Figure Page

1.1 Satellite image of the geographic extent of the three major, labeled McMurdo Dry Valleys. Color indicators mark sample locations for this study...... 7

1.2 Predicted dust deposition results from a production/transport/deposition simulation by Mahowald et al. (1999) for (a) present day climate and dust sources and (b) LGM climate and dust sources...... 8

2.1 Base map of the Taylor Valley with dots indicating areas of EST transects (red) and Sensit acoustic wind eroding mass sensors (blue)...... 48

2.2 Wind directions and sources in the Taylor Dry Valley, Antarctica...... 49

2.3 A cross-section of the Taylor Dry Valley indicating location of major geologic barriers, glaciers, and lakes...... 49

2.4 Generalized stratigraphic sequence of Southern Victoria Land, Antarctica...... 50

2.5 Examples of ice surface sample sites on (a) the Taylor Glacier in Taylor Valley and (b) Lake Fryxell in Taylor Valley...... 50

2.6 Examples of aeolian landforms sampled including (a) the Packard Dunes in Victoria Valley and (b) Lake Hoare Beach in Taylor Valley...... 51

2.7 An example of an elevated sediment trap from the Lake Fryxell transects ...... 51

2.8 Grain size distributions of bulk aeolian sediments differentiated by type and location, and then plotted as percent gravel, sand, and silt...... 52

2.9 Bar graph of LOI from the sand and silt particle fraction of each location...... 53

2.10 Bar graph of the absolute concentrations of major elements in sand and silt size fractions of samples...... 54

2.11 Major elemental oxide variation diagrams with average rock values for reference...... 59

Figure Page

87 86 2.12 εNd(0) and Sr/ Sr fields of ice core dust and potential source areas superimposed with values from this study...... 60

2.13 Photographic examples of weathering and wind in the MDV...... 61

2.14 Schematic of the weathering potential of various silicate minerals...... 62

2.15 Annual mean dust deposition from (a) South America, (b) Australia, (c) South Africa, and (d) Northern Hemisphere...... 63

2.16 Map of the tills related to glaciations and their age in the Taylor Valley...... 64

3.1 Flowchart of the partitioning of available water in Antarctica and how it relates to biota...... 96

3.2 Schematic diagram of the hydrologic continuum of the McMurdo Dry Valleys...... 97

3.3 Schematic vertical cross-section of a cryoconite hole...... 97

3.4 Bar graph of total dissolved solids from each sample location for both leach methods...... 98

3.5 Taylor Valley total dissolved solids with distance inland...... 99

3.6 Piper-style cation and anion variation diagrams...... 100

3.7 Relative percentages of major cations and anions from glacier sediments for Leach 1 and Leach 2...... 101

3.8 Relative percentages of major cations and anions from EST sediments for Leach 1 and Leach 2...... 102

3.9 Relative percentages of major cations and anions from lake sediments for Leach 1 and Leach 2...... 103

3.10 Relative percentages of major cations and anions from landform sediments for Leach 1 and Leach 2...... 104

3.11 Plot of Na+ vs. Cl- for both leach methods...... 105

2+ 2+ + - 2- - 3.12 Plot of Mg +Ca +Na -Cl vs. SO4 +HCO3 for both leach methods...... 106

- 3.13 Plot of HCO3 vs. H4SiO4 for both leach methods...... 107

Figure Page

2+ 2+ 2- - 3.14 Plot of Mg +Ca vs. SO4 +HCO3 for both leach methods...... 108

2+ - 3.15 Plot of Ca vs. HCO3 for both leach methods...... 109

2+ 2+ - 3.16 Plot of Ca +Mg vs. HCO3 for both leach methods...... 110

2+ 2- 3.17 Plot of Ca vs. SO4 for both leach methods...... 111

3.18 Plots of (a) inorganic and (b) total N vs P for both leach methods...... 112

3.19 Plots of (a) inorganic and (b) total N:P ratios for both leach methods with distance inland in the Taylor Valley...... 113

3.20 Schematic of the pathways of model evaporation of natural waters...... 114

- - 3.21 Plot of NO3 vs. Cl for both leach methods with values for high and low elevation soils for reference...... 115

xii

LIST OF TABLES

Table Page

2.1 Locations and sediment fluxes of elevated sediments traps in the MDV...... 65

2.2 Nitrogen, carbon, and organic carbon data presented as weight percentages for each sample...... 66

2.3 Sr and Nd isotopic values for three silts and one bulk glacial sediment sample from this study and rocks, waters, and soils from the Dry Valleys literature. .... 67

2.4 Chemical Index of Alteration values for average location and rock type data...... 68

3.1 Equivalent ratios of major ions to chloride for both leach methods with seawater for reference...... 116

- - 3- 2+ 3.2 Absolute concentration of bioavailable nutrients (NO3 +NO2 , PO4 , and Fe ) for Leach 1...... 117

3.3 Percent dissolution of aeolian material for leachates from both experimental methods...... 118

3.4 Grain size distribution of the samples that were experimentally leached...... 119

3.5 Major salts found in the MDV...... 120

2- - 3.6 Equivalent ratio of SO4 to HCO3 for both experimental leach methods...... 121

3.7 Nutrient limitation found in Taylor Valley terminal lakes...... 121

3.8 Fluxes of solutes to the MDV ecosystem from glaciers and lakes...... 122

3.9 Fluxes of inorganic nutrients to the MDV ecosystem from glaciers and lakes. 123

3.10 Comparison of estimated nutrient fluxes from aeolian sediments compared to the concentration of nutrients in leachates...... 124

xiii

CHAPTER 1

DUST IN THE GLOBAL SCALE

Aeolian transport involves the entrainment of material that is moved and deposited by wind circulation (Ridgwell, 2002). Termed ‗dust,‘ these particles are generally less than 50 μm in diameter (7-20% less than 1 μm in diameter) and are composed of mineral substrate such as soil, rock, and salt, but may also consist of viruses, pollen, and other materials of the same size fraction. The role that dust plays in the Earth system is dictated by its physical (size, sphericity, color) and chemical characteristics that differ between source regions (Ridgwell et al., 2002; Harrison et al.,

2001; McKenna Neuman, 2004; Denman et al., 2007; Goldstein et al., 2008). Miller and

Tegen (1998) and Ridgwell (2002) assert that the uplift of dust particles occurs most efficiently in areas of low soil moisture (ergo low precipitation), sparse vegetation, and minimal cohesion between particles. All of these preconditions are found in desert regions. Harrison et al. (2001) contend that 30% of continental land area today represents potential dust source regions that contribute 1-2 Pg per year to the atmosphere.

Natural dust sources are concentrated in arid and semi-arid tropical and subtropical regions (Harrison et al., 2001), while anthropogenic dust sources are derived from land use change (e.g. deforestation, agriculture), desertification and industrial emissions in the mid-latitudes of the northern hemisphere (Denman et al., 2007).

The deflation of dust occurs when wind speed is strong enough to overcome the cohesive forces of the substrate (Ridgwell and Watson, 2002). Particle size plays a key role in this process, as larger particles (60-100 µm) in the substrate are more easily entrained than smaller particles due to their relatively smaller surface area and decreased adhesive forces (Tegen, 2003). McKenna Neuman (2004) conducted a series of mass transport experiments in a wind tunnel facility at varying temperature and humidity conditions, and found that cold air (-40º C) was able to support rates of mass transfer up to 70% greater than hot air (32º C). This is consistent with a 2 to 20 times increased dust deposition found in glacial stages in the East Antarctic (EA) ice core record (Harrison et al., 2001; Tegen, 2003; Bar-Or et al., 2008; Basile et al., 1997; Delmonte et al., 2002;

Mahowald et al., 1999) and modeled last glacial maximum (LGM) dust fluxes (Harrison et al., 2001).

Deposition occurs through two main processes: dry deposition/sedimentation and wet deposition into terrestrial and marine environments. Dry deposition is gravitationally driven and accounts for the selective settling of larger particles closer to the source region, while the finer fractions (~ <1 µm) are transported significantly farther via suspension in the prevailing winds (Ridgwell and Watson, 2002). Formenti et al. (2001),

Shinn et al. (2000), and Neuer et al. (2004) all record evidence for this intercontinental/interoceanic transfer of fine dust from Africa to the Americas. Removal of the fine fraction is generally by wet deposition, also known as particle scavenging

(Miller and Tegen, 1998).

As a whole, dust composition exhibits spatiotemporal variance. Coastal regions reflect the influence of sea salt while inland and arid regions are dominated by soil-

2 derived constituents that may vary according to source area (Schlesinger, 1997). In the

McMurdo Dry Valleys of Antarctica (MDV), the polar desert region that is the focus of this study takes place (Figure 1.1), Fortner et al. (2005) found the influence of salts to be more important on the eastern (coastal) side and calcareous dust to be more important on the western (inland) side of the Canada Glacier. Also in the MDV, an analysis of the

- chemical composition of snow pits found elevational and distance inland trends in NO3 ,

Ca2+, and Cl- content (Witherow et al., 2006). Similarly, in the desert southwestern

United States compositional trends based on major and trace element abundance, particle size distribution, and Sr and Nd isotopes provide evidence for shifting sources of dust

(Goldstein et al., 2008; Reheis et al., 2009). These differences in influence also have an impact on nutrient content and transfer within the ecosystem, as dust from different source areas will have differing composition and may contribute otherwise unavailable nutrients (Denman et al., 2007; Schlesinger, 1997). In Hawaii, an atmospheric source of cations from marine aerosol and phosphorus from dust transported from central Asia are integral to maintaining productivity of Hawaiian rainforests (Chadwick et al., 1999).

Additionally, this demonstrates the large scale teleconnection of ecoregions through atmospheric dust transfer (Peters et al., 2008).

The composition, shape, and concentration of dust in the atmosphere can have important effects on the energy budget at the Earth‘s surface (Ridgwell, 2002; Denman et al., 2007; Miller and Tegen, 1998). The color of the dust relative to the surface underneath the atmospheric column or deposition area determines the kind of radiative, and thereby temperature, feedback that occurs (Ridgwell, 2002). Dust that is lighter than its surroundings will induce a localized cooling effect, while dust darker than its

3 surroundings will induce localized heating up to 2º C (Miller and Tegen, 1998). These localized temperature changes become especially important in polar regions where dust encounters high albedo ice and snow where the localized heating effect causes melting, thus reducing the albedo of the planet as a whole (Ridgwell, 2002). Miller and Tegen

(1998) also discuss the effect of seasonality on the extent of aerosol radiative forcing. In the Northern Hemisphere, where the majority of dust sources are located, radiative forcing due to dust is greatest during the summer when strong monsoon-derived winds carry the dust east out of India and Asia and west out of Africa. This effect is more muted during the Southern Hemisphere summer due to the relative sparseness of dust source regions.

Over glacial-interglacial time periods, there is a correlation between atmospheric

CO2 content and the relative dustiness recorded in ice cores from East Antarctica

(Ridgwell and Watson, 2002). Martin (1990) hypothesized that primary production in the ocean is Fe-limited, and that the introduction of Fe via dust deposition in the oceans would then increase primary productivity in phytoplankton and increase CO2 drawdown.

In today‘s oceans there are areas of reduced primary productivity despite the availability of macronutrients known as ―high nitrate low chlorophyll‖ (HNLC) regions.

The ―Fe hypothesis‖ postulated that the 90 ppm difference in atmospheric CO2 concentration between glacial (190 ppm) and interglacial (280 ppm) time periods was due to increased CO2 drawdown due to dust-Fe addition and hence ocean fertilization

(Martin, 1990). However, laboratory experiments were unable to conclusively account for Fe as the limiting nutrient in simulated HNLC environments (Ridgwell, 2002). In response, several in situ Fe-enrichment experiments including the Southern Ocean Fe

Release Experiment (SOIREE) were carried out to test the hypothesis (Watson et al.,

2000). During the course of SOIREE, Watson et al. (1991) fertilized a patch of ocean with reactive Fe and an inert tracer (SF6) and monitored fCO2 values inside the patch.

Their data supported the hypothesis that Fe is the limiting nutrient in the HNLC system.

However, the Fe found in dust deposited on the ocean surface is largely bio-unavailable because its dissolved (bio-available) phase is not found in high concentrations. Only

~0.5% of soil-derived Fe is soluble in seawater (Ridgwell, 2002; Falkowski et al., 2000;

Mahowald et al., 2005). Evidence exists indicating that atmospheric chemical/photochemical processing increases the solubility of dust-Fe by 0.01-80%, but the mechanisms behind the increased solubility is unclear (Mahowald et al., 2005; Baker et al., 2006b). Fe solubility also increases with decreasing particle size and mineral concentration, thus the highest solubility is found at deposition sites remote from their desert sources (Baker and Jickells, 2006). This was also found to be true for P (Baker et al., 2006b).

The solubility of other dust-derived elements can also be important to ecosystem dynamics, as other elements beside Fe help control ecosystem function. A study of the solubility of Al, Mn, and P in addition to Fe over the Atlantic Ocean found Fe, Al, and P solubility in Saharan dust (1.7%, 3.0%, and 10% respectively) was less than those of other source regions (5.2%, 9.0%, and ~32%, respectively). Mn solubility varied with source region but was 55% in Saharan dust and 56% for the dataset as a whole (Baker et al., 2006a).

Current dust flux

In the current climate regime, the major sources of dust are arid and sub-arid tropical regions including North Africa (Sahara and Sahel, the largest emitter of dust), the

Arabian Peninsula, central Asia (Gobi and Taklamakan Deserts), China, Australia, and the desert regions of North and South America (Mahowald et al., 2005). Altogether, these regions make up ~30% (50 x 106 km2) of continental landmass and contribute 1-3

Pg per year of dust to the atmosphere ((Harrison et al., 2001; Denman et al., 2007).

Other authors have modeled annual dust flux from the information of modern source regions and found the Sahara-Sahel dust corridor of Africa and the Gobi/Taklamakan

Deserts of central Asia to have the highest areal deposition fluxes on the order of 100 g m-2a-1 (Ridgwell, 2002; Mahowald et al., 1999; Mahowald et al., 2005). Of the estimated

1.7 Pg a-1 of dust emission predicted by Jickells et al. (2005), two-thirds is derived from

Sahara-Sahel dust corridor (Figure 1.2a). The relatively enriched areas of dust deposition juxtaposed to major source regions lay along major wind paths (Harrison et al., 2001).

Figure 1.2a shows predicted dust deposition based on the model of Mahowald et al.

(1999) and shows these trends in downwind dust transport.

Past dust flux

Over glacial time periods, there is a correlation between glaciations and increased dust deposition recorded in ice and marine sediment cores (Ridgwell, 2002). Mahowald et al. (1999) show that an expansion of source areas along with the increased wind intensity and a decreased hydrologic cycle expected in glacial time periods would increase atmospheric dust levels 2-20 times present levels worldwide during the LGM.

Source areas would have an average areal flux in excess of 200 g m-2a-1 (Mahowald et al.,

1999). Figure 1.2b shows predicted deposition during the LGM. When compared to

Figure 1.2a, the overall increase in dust deposition in glacial times is illustrated.

Figure 1.1 Satellite image of the geographic extent of the three major, labeled McMurdo Dry Valleys. Color indicators mark sample locations for this study. Base image derived from the USGS Land Image Mosaic of Antarctica (http://lima.usgs.gov/).

Figure 1.2 Predicted dust deposition results from a production/transport/deposition simulation by Mahowald et al. (1999) for (a) present day climate and dust sources and (b) LGM climate and dust sources. From Ridgwell (2002).

CHAPTER 2

GEOCHEMISTRY OF AEOLIAN SEDIMENTS OF THE MCMURDO DRY

VALLEYS, ANTARCTICA

INTRODUCTION

Dust has become a topic of increasing interest in the interface between climate and geological/ecological sciences. In much of the literature discussed previously, ―dust‖ refers to a size fraction that would be better considered an aerosol, <5 μm. However, in the sparse literature of the aeolian fluxes of the McMurdo Dry Valleys (MDV), ―dust‖ refers to the silt and smaller particle sizes together (Lancaster, 2002). In part, this change in nomenclature is because sand size particles dominate the aeolian sediment load in the

MDV (Lancaster, 2002; Speirs et al., 2008b) and in cold regions in general (Sepälä,

2004). Sepälä (2004) indicates this increased particle size compared to the <5 μm size is due to the fact these aeolian materials represent only a first-order physical weathering of parent material and regolith. Therefore, for the purposes of this study we will refer to

―dust‖ as silt size and smaller sediment particles, <63 μm.

Investigations into the differences in aeolian sediment transport at cold temperatures have been investigated by McKenna Neuman (1990, 1993, 2004) and discussed in both laboratory controlled and field settings. The processes controlling particle transport are the same in cold regions as they are in warmer analogs, but the deflation of particles can only occur when a critical shear velocity is exceeded (Bagnold,

1941). The critical shear velocity of natural sediments, however, is probably better defined by a range of values rather than a single value (McKenna Neuman, 1993). This activation feature occurs in both hot and cold environments. However, the roughness of a surface also affects the shear velocity and transport rate, especially in unvegetated terranes, and is a three-dimensional property outlined by the equation:

where n is the number of roughness elements in a defined area S, b is average element diameter, and h is average element height (Lancaster, 2004). Thus, Lancaster (2004) and

Lancaster et al. (2010) have suggested that a more accurate description of the shear velocity that takes into account the proportion of the shear velocity absorbed by roughness elements in a given area to be given by the equation

where σ is the basal to frontal area ratio of the roughness element (i.e. a boulder), λ is the roughness density (equation 3.1), β is the ratio of element to surface drag coefficients, and m is an empirical constant. On a sand sheet in the Victoria Valley, this meant that wind shear threshold velocities in rough areas were 1.2 times that of bare sand surfaces

(Lancaster et al., 2010).

Wind speeds are often higher in cold environments, while air density varies inversely with air temperature (McKenna Neuman, 1993). The higher density of air at colder temperatures increases the drag exerted on entrained particles and led Selby et al.

(1974) to speculate that the strong, cold winds of the Victoria Valley could transport sediments more effectively than a warm desert analog. Based on the Packard Dune Field

10 in Victoria Valley as the cold weather extreme, it was estimated that at -70º C a 36 m s-1 wind could carry a 3 mm grain to 2 m, which would require a wind speed of 42 m s-1 at

0ºC and 45 m s-1 at +50º C (Selby et al., 1974).

In warm environments, aeolian deposits are well-sorted, but in cold environments deposits contain a wider range of particle sizes (McKenna Neuman, 1993). Lag surfaces play an important role in the MDV, as indicated in a study on the roughness density of an area and its effect on shear threshold values in the Victoria Valley (Lancaster et al.,

2010). As a whole, the MDV valley floors exhibit poorly sorted sandy gravel (Fountain et al., 1999b) available for aeolian transport, so this grain size and roughness effect would be of greater consequence.

Snow and ice deposited on and within aeolian deposits act to reduce particle deflation and transport through stabilization of landforms (McKenna Neuman, 2004;

McKenna Neuman, 1993). Snow acts as an effective sediment trap and a protective cover as referenced by niveo-aeolian deposits found worldwide during the winter

(McKenna Neuman, 1993). These findings are corroborated by a study in the Packard

Dune Field of Victoria Valley that showed particle fluxes increase when the dune surface is above freezing and wind speeds were relatively higher (Speirs et al., 2008a). During the shaded ―night‖ saltation was essentially shut down due to freezing of pore water that had melted during the ―day‖ from niveo-aeolian deposits, atmospheric stability near the dune surface, and reduced wind speed (Speirs et al., 2008a). The last observation can be tied back to colder temperatures during the ―night‖ that create a more stable boundary layer and decrease the thermal gradient between valley floor and glacial surface, which is responsible for the strong winds (Speirs et al., 2008a). The presence of niveo-aeolian

11 deposits have been found not to completely anchor the dune, but do impede migration rates, and therefore might lead to decreased rates of dune migration (Bourke et al., 2009).

Prevailing wind systems and their interactions are the main driver of sediment and organic matter redistribution in the MDV (Lancaster, 2002). Studies of the total flux of particulate matter have found a total flux at 1 m above land surface averaged over a five year collection period to range from 0.26 g m-2a-1 on the Commonwealth Glacier in the

Taylor Valley to 441.80 g m-2a-1 at Lake Brownworth in the Wright Valley (Table 2.1)

(Lancaster, 2002). A strong correlation was found between mean wind speed and annual flux of aeolian silt (r2=0.74), while there was no such correlation in the sand flux. There is an exponential decrease in the flux of silt and clay material with increasing elevation

(r2=0.49), which illustrates the importance of local source materials (Lancaster, 2002).

Within the Taylor Valley, Lancaster (2002) also found a strong correlation (r2=0.92) between silt flux and distance from the coast. An increased flux inland could only be qualitatively shown. Additionally, the total sediment flux varied with sediment availability as seen in increased total sediment flux at collectors located on valley floors as opposed to at elevation on glaciers.

A series of nine elevated sediment traps (EST) (~0.3 m) deployed at Lakes

Fryxell, Hoare, and West Bonney in the Taylor Valley by the McMurdo Dry Valleys

Long Term Ecological Research team (Figure 2.1) received 355.15, 470.10, and 70.72 g m-2a-1 averaged over a 9 year period between 1999 and 2008, respectively (Table 2.1)

(Šabacká et al., 2010). Sensit Eroding Mass Sensors have been installed at Lake Bonney,

Lake Hoare, and Lake Fryxell to collect wind speed, direction and particle flux every 15 minutes. Not surprisingly, these data indicate that there is an increase in particle flux

12 during katabatic wind events. Katabatic winds are strong (average 2.5-5.3 m s-1 in the

MDV), dry, topographically-driven winds from the west off the ice sheet and are discussed further below (Doran et al., 2002; Nylen and Fountain, 2004) The strength of the katabatic winds would effectively entrain particles and move them further downvalley, which accounts for the pattern of particle flux in the ESTs of Lake Hoare >

Lake Fryxell > Lake Bonney (Šabacká et al., 2010). The Canada Glacier separating Lake

Hoare from Lake Fryxell deters the transfer of sediment further downvalley by all but the strongest katabatic wind events.

In comparison, the Lancaster (2002) study outlined above followed a total particle flux pattern at 1 m of Lake Bonney > Lake Fryxell > Lake Hoare. The variation in particle fluxes is due to the sand size flux, which was found to be most important at

Lake Bonney and sites in the Wright and Victoria Valleys where sand flux exceeded 90% of total flux (Lancaster, 2002). The total sediment flux trend in the Lancaster (2002) data may also reflect climatic conditions: Lake Bonney is the driest of the Taylor Valley basins due to the desiccating effect of the katabatic winds and lack of coastal wind interaction and the low humidity indicates a greater potential for source particles

(McKenna Neuman, 1990).

Dust and ice cores

Dust from four ice cores (EPICA Dome C, ―old‖ Dome C, Dome B, and Vostok) in East Antarctica has been analyzed for both absolute concentration and sediment provenance. Each ice core records a ~20-50 fold increase in absolute dust concentration between the present interglacial (15-35 ppb) and the Last Glacial Maximum (640-875 ppb) (Delmonte et al., 2002). A more extensive comparison between the dust records of

13 the EPICA Dome C and Vostok ice cores reveals a coherent record between the two sites during the past 220,000 years with respect to dust flux, including three major dust events corresponding to glacial stages 2, 4, and 6 (Delmonte et al., 2004).

Recent research has focused on developing a reliable means of detecting provenance of dust, especially as it applies to the dust found in ice cores of East

Antarctica (Basile et al., 1997; Delmonte et al., 2002; Mahowald et al., 1999; Delmonte et al., 2004; Hesse and McTainish, 1999; Hinkley and Matsumoto, 2001; Delmonte et al.,

2008; Gaiero, 2007). The bulk of these studies have relied upon Sr and Nd isotopic signatures to identify potential source areas and reconstruct atmospheric pathways leading to deposition. However, there is a systematic isotopic fractionation with particle size noted in the studies. A recent study by Feng et al. (2009) showed that Sr isotopes are sensitive to wind sorting and Nd isotopes are sensitive to dust source region, and therefore variations in isotopic composition with grain size must be considered.

Because dust concentration is an indicator of arid conditions, wind strength, and the extent of sea ice in the Antarctic, dust in ice cores has been utilized for paleoclimate studies (Ridgwell et al., 2002; Hinkley and Matsumoto, 2001; Gaiero, 2007). In addition, the provenance of dust in the EPICA Dome C and Vostok ice cores of EA has been the

87 86 subject of debate in recent years. Through Sr/ Sr and εNd(0) isotopic studies, researchers have asserted that the dust in the ice cores isotopically resembles southern

South America (Basile et al., 1997; Delmonte et al., 2004; Gaiero, 2007). The data presented in Delmonte et al. (2004) for South American Patagonia and MDV superpose

87 86 each other and the ice core data in Sr/ Sr vs. εNd(0) plots, but only Patagonia is discussed as a potential source of dust; the MDV were precluded because of the katabatic

14 winds thought to impede the transfer of dust inland and the extent of MDV glaciation thought to exist during glacial maxima. These studies note there is a fractionation of isotope composition with particle size (Basile et al., 1997; Delmonte et al., 2004).

Study objectives

The objectives of this study are to determine the chemical composition of MDV aeolian materials and from this information ascertain their origin. The following hypotheses will be addressed through the results of these experiments:

1. The composition of the dust reflects a nearby source

2. MDV-derived dust isotopically resembles preserved dust in ice cores in East

Antarctica and may indicate an inland transfer of material.

Due to the possibility of composition varying with grain size, major elements will be investigated at both the sand and silt grain size. As mentioned previously, dust found in the Dome C and Vostok ice cores has a mode of ~2 μm. Thus, it follows that any comparison with the isotopic composition of MDV aeolian material should occur at the smallest grain size fraction available

FIELD AREA DESCRIPTION

Roughly 97% of Antarctica is covered by ice that may be greater than 3000 m in depth. The remaining 3% of land area consists of ice-free regions, the largest of which is the McMurdo Dry Valleys (MDV) in the Transantarctic Mountains that cover approximately 4800 km2 located at 76o30‘-78o30‘S, 160-164oE (Figure 1.1). They are a series of transverse valleys situated perpendicular to the Ross Sea coast of southern

Victoria Land, Antarctica that consist of a mosaic of hyper-arid landscapes made up of

15 coarse soils, glaciers, and in some cases closed-basin lakes. The region exists because the Transantarctic Mountains dam much of the eastward flow of the East Antarctic Ice

Sheet into McMurdo Sound. The MDV have been the focus of the McMurdo Long Term

Ecological Research (MCM-LTER) program supported through the National Science

Foundation since 1993. The MCM-LTER represents the coldest and driest of the LTER network‘s 26 sites. Taylor, Wright, and Victoria are the three main MDVs and consist of the majority of the ice-free land area (Marchant and Denton, 1996). Much of the work of the MCM-LTER in this and other studies has focused on the centrally located Taylor Dry

Valley located ~80 km from McMurdo Station (Figure 1.1). The valley extends ~35 km southwest from Ross Sea terminating in the Taylor Glacier, an outlet glacier of the East

Antarctic Ice Sheet.

Climate

The MDV are one of the coldest and driest ecosystems globally, and are considered a polar desert due to their extremely low mean annual temperatures and very low precipitation. Mean annual valley floor temperatures in the MDV range from -15ºC to -30ºC (Doran et al., 2002); within the Taylor Valley, this range narrows from -16ºC to

-21ºC (Fountain et al., 1999b). Based on long term data collected at 7 meteorological stations in the three main MDV, Doran et al. (2002) found the relationship of mean annual temperature between valleys to be Taylor Valley > Wright Valley > Victoria

Valley. Annual precipitation for the MDV is less than 10 cm, and probably less than 5 cm, water equivalent that falls as snow that is subsequently sublimated before extensive melting can occur (Fountain et al., 1999b; Fountain et al., 1998; Fountain et al., 2010).

However, some accumulation of snow occurs near the coast, which is also where relative humidity is generally highest (Doran et al., 2002).

An east-to-west climatic gradient exists within the MDV, and is most pronounced within the Taylor Valley. Humidity and precipitation both increase with proximity to the coast in the eastern terminus; wind speed and temperature increase toward the ice sheet in the west (Fountain et al., 1999b; Doran et al., 2002). These gradients are largely associated with the interaction of the coastal and katabatic wind systems that predominate in the MDV (Figure 2.2) (Fountain et al., 1999b). In the Taylor Valley, another control on the gradient is the Nussbaum Riegel that bisects the valley ~20 km inland (Figure 2.3).

This geomorphologic boundary further confines the effects of the two wind systems creating two climatically disparate regions (Fountain et al., 1999b).

The topographically channeled, adiabatically warmed katabatic winds that flow from the west off the ice sheet are more precisely called foehn winds (Figure 2.2) (Speirs et al., 2008a; Ayling and McGowan, 2006). These foehn winds are also dry, acting as a desiccant and deterring precipitation. Doran et al. (2008) found a strong correlation between foehn wind events and degree days above freezing (a general indicator of melt) in the MDV region. The other wind regime is the easterly coastal winds that advect into the valleys from Ross Sea. These are generally cooler and moister than the foehn winds, allowing for increased precipitation where they predominate (Fountain et al., 1999b).

Additionally, the foehn winds are generally stronger than the coastal breezes accounting for the increased wind speeds with proximity to the polar plateau documented by Doran et al. (2002).

The interaction of these wind systems results in the upvalley increase in air temperature and wind speed, and decrease in precipitation and relative humidity

(Fountain et al., 1999b; Doran et al., 2002). Areas that are shielded from the foehn winds, such as Victoria Valley, or are further downvalley consistently exhibit colder mean annual temperatures, despite the more sporadic nature of the foehn events (Doran et al., 2002). This interaction is exemplified by the comparison of hydrologic response to the ―non-flood‖ 2000-2001 and ―flood‖ 2001-2002 melt seasons wherein the flood year was associated with a 2.4ºC increase in mean summer temperature and an ten-fold increase in degree days above freezing brought about by increased foehn activity (Doran et al., 2008).

Landscape

The physical landscape of the MDV vividly displays the interaction between wind and water in an unvegetated ecosystem, as well as distinct glacial geomorphology.

Unconsolidated sandy gravel valley floors and glacial tills give way to extensive areas of exposed bedrock at valley margins and intervening mountain ranges (Fountain et al.,

1999b; Keys and Williams, 1981). Alpine glaciers enter the valley from the mountains terminating in 20 m ice cliffs that are the major source of water to the ecosystem during the summer melt season. The meltwater enters stream channels and flow for 4-12 weeks per year into the endorheic lakes in the area (Fountain et al., 1999b; McKnight et al.,

1998).

Geology

A generalized stratigraphic sequence of Taylor Valley can be found in Figure 2.4.

The basement complex consists of Precambrian metasediments (schist, hornfels, marble)

18 that are overlain by a sequence of greywackes, argillites, limestone and conglomerate known as the Ross Supergroup(Haskell et al., 1965). These were both intruded by the

Granite Harbor Intrusive Complex of granites and granodiorites(Haskell et al., 1965).

The Devonian-Triassic age Beacon Supergroup consisting of near flat-lying sandstones, siltstones, and conglomerates were then deposited unconformably following the glacial erosion of the lower units (Marchant and Denton, 1996; Haskell et al., 1965). These are intruded by the Jurassic-age Ferrar Dolerite and Cenozoic volcanics related to the

McMurdo Volcanism (Marchant and Denton, 1996; Haskell et al., 1965; Green et al.,

1988). The McMurdo Volcanism is further manifested in the Taylor Valley by the presence of approximately thirty scoraceous olivine basalt cinder cones from 550 to 1220 m above sea level (Haskell et al., 1965). Quaternary glacial, alluvial, and lacustrine sediments blanket the valley floors and are a result of the advance and retreat of the East and West Antarctic Ice Sheets and the alpine glaciers into and out of the valleys since the

Last Glacial Maximum (LGM) (Hall and Denton, 2000).

Aeolian “Landforms”

The evidence of abundant aeolian activity is observed throughout the MDV ecosystem. The most widespread aeolian deposits seen in the MDV are pebble ripples.

Sand sheets, dunes and accumulation areas in ventifacts are also present to varying degrees (Selby et al., 1974). The most well-documented aeolian landform is the dune field south of the Packard Glacier in Victoria Valley made up of transverse and whaleback dunes (Selby et al., 1974; Speirs et al., 2008a; Bourke et al., 2009; Calkin and

Rutford, 1974). An additional deposit known as Lake Hoare ―Beach‖ abuts the Canada

Glacier and is known to be of aeolian origin (Nedell et al., 1987).

Soils

The most extensive landform in the MDV is arid soils developed on glacial tills of varying age that cover approximately 95% of the unglaciated areas below 1000 m (Poage et al., 2008). Features common to MDV soils include pebbles or stone pavements, coarse and variable textures, lack of cohesion and structural development, wide variation of high salinity, neutral to alkaline pH, low organic matter content, and the presence of permafrost at variable depth but primarily at depths lower than 100 cm (Poage et al.,

2008; Campbell and Balks, 1998). The local climate, rock/regolith substrate, and age of the surface determine the degree to which any of these features are developed. In general, MDV soils are classified by the physiographic position defined with climate, weathering, and parent material subdivisions (Campbell and Claridge, 1987). These divisions decrease with moisture availability and elevation from coastal regions proceeding through valley floors, valley sides, upland valleys, and the plateau fringe

(Campbell and Balks, 1998).

Glaciers

The equilibrium line altitude (ELA) of the glaciers in the Taylor and Wright

Valleys increase with distance from the coast with an average gradient of 18 m km-1, a factor of five times greater than the gradient of the continental United States (Fountain et al., 1999a). The increase in ELA away from the coast is a function of the climate gradient resulting from the interaction of the foehn and coastal winds noted previously.

While the MDV glaciers are relatively free of debris(Fountain et al., 2006), dust blown in by the prevailing wind systems preferentially deposits onto glacial (and lake) surfaces in relatively low-lying areas and affect the overall radiation budget by decreasing the albedo

20 of the ice (Lyons et al., 2003; Bagshaw et al., 2007). During the summer melt season, the sediments absorb more heat compared to the surrounding ice and preferentially melt into the ice surface. On glaciers this process forms cylindrical structures known as cryoconite holes that deepen as a function of radiation balance until the rate of melt-in equals the rate of glacial ablation (Bagshaw et al., 2007; Wharton et al., 1985)(E. A.

Bagshaw et al., 2007; Wharton, McKay, Simmons, & Parker, 1985)(Bagshaw, et al.,

2007; Wharton, et al., 1985)(Bagshaw et al. 2007, Wharton et al. 1985) but this theoretical equilibrium depth is not recognized readily in nature. The presence of organisms in these features implies nutrient input that must come from the dissolution of the cryoconite material (Foreman et al., 2007). On Canada Glacier in the MDV electrical conductivity, an indication of the degree of cryoconite dissolution into the meltwater is greatest during the first part of the summer season indicating that there is generation of melt, an important precondition for nutrient cycling (Fountain and Tranter, 2008).

Streams

Ephemeral streams formed by glacial melt are the only source of water to the lakes of the MDV and flow intermittently during the 4-10 week summer melt season

(Fountain et al., 1999b; Conovitz et al., 1998). They similarly serve to redistribute nutrients and other solutes derived from glacial meltwater into the environment. In addition, aeolian transport of sediments into the streambed when water is not flowing

(winter) is a further potential source of solutes to the streams changing the chemical composition of the glacial meltwater (Lyons et al., 1998). However, much of the weathering of the streambed and channel occurs in the hyporheic zone (Maurice et al.,

2002). These solute-rich waters exchange with relatively undiluted streamwater allowing

21 weathered solutes and nutrients to be redistributed throughout the hydrologic continuum

(Nezat et al., 2001; Gooseff et al., 2002).

Lakes

The lakes of the MDV fit into four descriptive categories: permanently frozen to the base, seasonally frozen, permanently liquid, and permanently ice covered (Hendy,

2000). The lakes of the Taylor Valley are endorheic and covered permanently with 3-6 m of ice (Fountain et al., 1999b). The larger lakes in the valleys to the north (Wright and

Victoria) are more often found to be frozen to the base with the exception of Lake Vanda in Wright Valley which also sustains a floating perennial ice cover (Hendy, 2000). This may be a function of their location at higher altitude, as Taylor Valley as located at lower elevation as the Wright or Victoria Valleys (Spigel and Priscu, 1998).

Like the MDV glaciers, the uneven lake surfaces act as a trap for wind transported sediments (Wharton et al., 1985; Sleewaegen et al., 2002; Squyres et al., 1991; Nedell et al., 1987; Jepsen et al., 2010). Differential warming of these aeolian deposits forms vertical conduits through melting and fracture and act as the primary sedimentation mechanism in the lakes with perennial ice cover over water (Squyres et al., 1991; Nedell et al., 1987). Aeolian sediments are also found in layers within the permanent ice cover

(Squyres et al., 1991; Jepsen et al., 2010). The lakes of the Taylor Valley also exhibit strikingly different chemistry even in their relatively fresh surface waters despite their similar climate and basin geology (Lyons et al., 2000). Lake Hoare is essentially freshwater while Lake Fryxell is brackish, and Lake Bonney hypersaline (Spigel and

Priscu, 1998). These differences have been attributed to differences in past climate, age, and landscape position (Lyons et al., 1998).

METHODS

Field site identification and description

Sites of aeolian deposition were identified and sampled during November and

December 2008. The sites can be subdivided into three general descriptions: ice surface

(lake and glacier), ―landform‖, and elevated sediment traps (EST) described below. All samples are listed in Appendix A along with their coordinates, and site description.

Ice Surface Samples

Samples were collected from the surface ice of each of the lakes (Fryxell, Hoare,

East and West Lobe Bonney), Mummy Pond, and four of the glaciers (Commonwealth,

Howard, Canada, Taylor) in the Taylor Valley (Figure 2.5). Some mechanism of transport is required to transfer sediments from the landscape onto these ice surfaces; two logical possibilities are gravitational settling and wind. In selection of sites, care was taken to avoid the perimeter of the lake and glacier surfaces where gravitational settling would be more likely to occur, so it has been assumed that the materials collected were transported there by aeolian processes.

An additional concern involved the degree to which leaching had already occurred within the sample sediments. Ideally, the samples would be newly deposited and have no prior exposure to meltwater. In the field, however, these conditions were difficult to identify and subsequently the sampling protocol was altered to shallow collections of material that appeared to be dry or that deposited on snow.

Aeolian “Landforms”

The selection of aeolian landform sample sites was much more subjective than that of ice surface sample sites because a more extensive knowledge of desert/aeolian

23 geomorphology had to be acquired to interpret landforms in the field. As noted above, aeolian landforms are documented in the literature throughout the MDV in the form of ripples, sand sheets, dunes, and accumulation areas (Selby et al., 1974). These attributes and the modern desert depositional environments outlined in Boggs (2006) were sought when selecting sample sites used in this study. The most well-documented of these landforms is the desert erg that has formed in the Victoria Valley (Selby et al., 1974;

Speirs et al., 2008a; Bourke et al., 2009; Calkin and Rutford, 1974), and was sampled in addition to a single dune structure adjacent to the western face of the Canada Glacier known as Lake Hoare ―Beach‖ (Figure 2.6). Additionally, several areas of active sediment accumulation including ripple and sand sheets in Bull Pass, Taylor Valley, and

Wright Valley were also sampled (Figure 1.1).

Elevated Sediment Traps

The EST apparatus (Figure 2.7) consists of an aluminum bundt pan mounted on a

PVC pipe. The pipe was then driven into the ground and secured with rocks to a total height of ~0.3 m. Wire mesh was cut to fit the opening of the bundt pan just below the surface, clean glass marbles were placed on top of the mesh. The ESTs have been deployed at three locations on the southern side of TDV adjacent to Lakes Fryxell, Hoare and West Bonney. At each location, a transect of nine ESTs were installed perpendicular to the lake margin south toward the Kukri Hills. Accumulated aeolian sediments have been collected annually since 1999 (Šabacká et al., 2010).

Sample collection and handling

All samples were collected in new Nasco Whirl-Pak® bags with clean, plastic utensils. Care was taken to face into the prevailing wind direction during collection and

Nitrile gloves warn to minimize contamination of the samples. The bags were then sealed and stored at -2ºC at Lake Hoare field laboratory until transport back to McMurdo

Station. The latitude and longitude of each sample location was obtained using a handheld Garmin GPS that was accurate to within 10 meters and recorded along with pertinent field notes.

At Crary Laboratory in McMurdo Station, each sample was placed in a clean porcelain crucible and dried overnight at 100ºC to drive off any residual moisture. After cooling to room temperature, samples were repackaged in new Whirl-Pak® bags and stored at -2ºC in insulated cardboard containers until early February when they were transported back to The Ohio State University for further experimentation and analysis.

At The Ohio State University samples were homogenized as best possible prior to any further investigation. Fifteen samples representing glaciers, lakes, and aeolian landforms were subsampled for carbon and nitrogen analyses. These fifteen samples plus the three EST samples were used in the leaching experiments described in Chapter 3. To avoid contamination, nitrile gloves were worn at all times during sample preparation and handling.

Carbon and nitrogen analysis

Approximately 10 g of the eighteen samples were homogenized and crushed by hand in a ceramic mortar and pestle, then placed in a 20 mL plastic scintillation vial and sent to personnel at the School of Natural Resources at The Ohio State University for analysis of carbon (C), organic carbon (OC), and nitrogen (N). The dry combustion methods of Nelson and Sommers (1996) for C and OC, and Bremner (1996) for N were conducted on a CE instruments® NC2100 soil analyzer. Prior to sample analysis, the

25 instrument was calibrated to a four-point calibration curve using a 4.84% N, 70.055% C atropine standard. Samples were then dried at 60ºC to drive off residual moisture, and loaded into a tin capsule for analysis on the instrument. A standard was analyzed at the end of the run to detect any drift in the instrument. Precision is ±10% or better for all the samples discussed here. Detection limits were determined to be 0.8% for N and 0.4% for

C and OC.

Particle size analysis

Approximately 100 g of the homogenized sample was placed in a two-size (2 mm and 62.5 μm) nested sieve and mechanically shaken for 15 minutes. This process was repeated until the entire sample had been separated. Each size fraction was weighed to obtain percent gravel (≥ 2 mm), sand (0.0625-2 mm), and fines (≤ 62.5 μm) calculated from the mass data of individual particle sizes compared to the total mass of the sample.

Preparation for elemental analysis

The sand and fine size fractions of each sample was crushed in an alumina ceramic shatterbox for a minimum of 8 minutes until a grain size of less than 2 μm was achieved. They were collected in clean high density polyethylene (HDPE) sediment storage containers with lids or new Whirl-Paks® and labeled with site identifiers and size fraction. The gravel fraction was not crushed, but also collected and labeled as described above.

Loss on ignition (LOI) analysis

A total LOI method was utilized to account for organic material, carbonate material, and bound water. An aliquot of the powdered sample was re-dried overnight at

100ºC in an alumina weigh dish, and then allowed to return to room temperature in a

26 glass desiccator. The mass of a ceramic crucible (mc) was recorded, then approximately

3 g of the re-dried sample added and the new mass recorded (ms). The sample and crucible were ignited in a muffle furnace for 1 hour at 1025ºC and gradually cooled back to 100ºC over a four hour period. A final mass of sample and crucible (mf) was then determined. This method was applied to both the sand and fine fractions. Total LOI was calculated according to the equation:

Elemental oxide analysis

A 2.5 g aliquot of the dried, ignited LOI sample was combined with 10 g of

® SpectroCertified Pre-Fused Fusion Flux Lithium Tetraborate (100% Li2B4O7) in a disposable plastic beaker and homogenized with an alumina spatula. The mixture was then transferred to a platinum crucible and loaded into a Phillips® Perl‘x 3 automatic bead machine and ignited sequentially for 4 minutes at 800ºC, 4 minutes at 1100ºC, and 8 minutes at 1150ºC. The resulting molten, homogenous sample was poured onto a platinum casting dish and air dried for 3 minutes to solidify. The resulting glass bead was labeled and stored in a glass desiccator until analysis. Any bead with a visual imperfection (i.e. incomplete sample dissolution, cracks, mottled appearance) was discarded.

Due to limitations in the total mass of the fine size fraction, all of the fines samples from each feature were combined. For example, all the fines from the Canada

Glacier were combined into one bead for analysis. Even when combined many of the fines samples did not consist of the 2.5 g minimum required for the 1:4 sample to flux ratio described above. Therefore, the combined-fines samples were made into 1:10 27 sample to flux ratio beads following the same procedure above except decreasing the amount of ignited sample to 1.0 g.

The fused major element beads were analyzed in a PANalytical® MagiX Pro® x- ray fluorescence (XRF) spectrometer with a PW2540 VRC sample changer to determine elemental oxide composition as weight percent using internal calibration curves.

Samples were analyzed in triplicate and an average concentration calculated. This average was then normalized to the sum of concentrations using LOI data previously determined. Relative standard deviation were calculated for the triplicate runs and determined to be ±5 % and generally less than ±1 %. Major element data are reported as weight percent oxide in Appendices B and C for sand and silt, respectively. Percent major elemental oxide values were converted to absolute concentrations in g element kg-1 for more direct comparison between C, OC, N, and other literature values.

The United States Geological Survey (USGS) Columbia River Basalt (BCR-2) major oxide bead was used as a standard for both 1:4 and 1:10 sample to flux ratio analyses. The BCR-2 bead was run as a sample approximately every four samples and the accuracy of the measurements determined to be within 2% of the recommended values provided by the USGS.

Sr and Rb isotopic analysis

A portion of the fine size fraction of from the Canada Glacier, Commonwealth

Glacier, and Howard Glacier, as well as a bulk sample from the Taylor Glacier were manually ground as fine as possible with an agate mortar and pestle. These were placed in labeled 20 mL plastic scintillation vials and sent to personnel at the Thermal Ionization

Mass Spectrometry (TIMS) Facility at Boston University for analysis. Details of sample

28 preparation and analysis for Nd can be found in Harvey and Baxter (2009). Samples for

Sr preparation and analysis follow standard methods including Sr spec resin (Ethan

Baxter, personal communication).

Statistical analyses

Sand and silt samples were considered separately during the course of statistical analysis. Due to the mass limitations explained above resulting in a single silt-size analysis at a majority of sample locations, statistical analysis between sand and silt size fractions could only be considered from a systematic standpoint. Statistical differences between site type and basins in the Taylor Valley were assessed using a single factor

Analysis of Variance (ANOVA). All ANOVA analyses were performed using the Data

Analysis package of Microsoft Excel with an alpha value of 0.05. Where F-value is greater than the F-critical value and the p-value is sufficiently small (p<0.05), there is a statistically significant difference between the variables indicating a true phenomena instead of random variation. In the text, ―statistically significant‖ refers to an analysis with p<0.05 calculated from the ANOVA.

RESULTS

Particle size distribution

Appendix A contains location information and the total mass of sand, silt and gravel collected at each sample site. A comparison of the relative distribution of particle sizes in the bulk aeolian sediment samples was achieved by calculating percent gravel, sand, and silt and plotting these values on a ternary plot as shown in Figure 2.8. All but five samples have a composition of greater than 80% sand by weight.

Silt composition ranges from 0 to 15.8% by weight. Samples containing the greatest amount of silt were collected from glacial surfaces in the eastern Fryxell and

Hoare basins of Taylor Valley. Gravel composition ranges from 0 to 60.0% by weight.

The samples with the highest composition of gravel are lake and glacial surfaces of the

Bonney basin of Taylor Valley, in addition to Bull Pass located between Wright and

Victoria Valleys. In general, there is higher gravel composition in the Bonney basin and greater silt composition in the Fryxell basin. This could be related to the age of the respective basins; wind has been reworking and deflating silt and sand size material in the older Bonney basin for a longer period of time than the younger Fryxell basin. Lake

Hoare lies intermediary to these two. Within each basin, glacial surface samples have a greater composition of silt than lake samples, indicating an additional effect due to elevation on the bulk composition of aeolian sediments.

The samples from Hoare Beach (an aeolian landform in the Hoare basin of Taylor

Valley) along with those from the Wright Valley and the Packard Dunes of Victoria

Valley are extremely well-sorted with a composition of greater than 90% sand. As all of these areas are dunes or dune-like in structure, the extremely high sand composition could be an inherent feature reflecting the formation of the structures through the sorting and transport of particles by wind. This is further corroborated by the fact that EST samples consisted only of sand sized particles.

Carbon, organic carbon, and nitrogen

Values for N, C, and OC analysis along with calculated carbonate values are found in Table 2.2 for fifteen selected sites consisting of glacial surfaces sediments (6 sites), lake surface sediments (5 sites) and aeolian landforms (4 sites). Only OC has been

30 measured for the 3 EST sites. There are only two samples where nitrogen was detected in the sediments: Lake Fryxell and the seaward side of the Commonwealth Glacier. The

East Commonwealth Glacier sample is just above the detection limit and may be within the error of the analysis, essentially leaving Lake Fryxell as the only sample with nitrogen found in its sediments.

Carbon values range from below detection to 12.7 mg g-1, while OC ranges from below detection to 14.3 mg g-1. A general decrease in both C and OC with distance from the coast exists in both glacial samples and aeolian landforms, and may be related to relative age of the landscape. The sample from Lake Fryxell in the Fryxell basin is enriched in C and OC compared to the Hoare and Bonney basins. The latter two basins show variable composition with distance up-valley.

Loss on ignition

Appendices B and C contain LOI values for all samples of the sand and silt size fraction, respectively. Sand LOI values range from 0.16 to 4.15%; silt LOI values range from 0.95 to 7.62%. At all sample locations, the LOI of the silt fraction is greater than the LOI of the sand fraction (Figure 2.9). There is no significant difference between sample types in either size fraction. Similarly, there does not appear to be any MDV- or

Taylor Valley-wide spatial pattern in the silt LOIs. However, there is a statistically significant difference in the LOI of sands between the lake basins of Taylor Valley with

Fryxell Basin > Hoare Basin > Bonney Basin.

Major elements

Absolute concentration of the major elements in g kg-1 of both particle sizes can be found in Figure 2.10. In general, Si and Na are more concentrated in the sand size particles while Ti, Al, Fe, Mn, Mg, Ca, K and P demonstrate enrichment in the silt size.

All sand samples contain a higher concentration of Si than silt samples. There are statistically significant differences between sample type and Taylor Valley lake basin

(p=0.002) only in the sand particle size in Si, while silt samples do not demonstrate any distinct differences. Concentration of Na is more mixed between sample sizes, but at each individual sample site Na has a greater concentration in sand than in silt with the exception of Victoria Dunes and Bull Pass. Significant differences were found between sample types only in the sand fraction and between Taylor Valley lake basins only in the silt size in Na.

Significant variations in sample type were found in the sand size of Ti and Mn, and the silt size of Ca. Taylor Valley basin variation was noted in the sand and silt size for Ti and the sand size of Ca. No significant variation was noted for group type of either size fraction with regards to Fe. Three sand P concentrations (Commonwealth Glacier,

Howard Glacier, and Lake Fryxell) are greater than the lowest silt P concentration.

ANOVA analyses indicated significant variation in the sand size between sample type and in both particle sizes between the basins of Taylor Valley. Commonwealth Glacier,

Howard Glacier, and Lake Fryxell all lie in the eastern Fryxell basin of the Taylor Valley.

While the Fryxell basin has the highest average P concentration in the sand size, the

Hoare basin has the highest concentration in the silt size.

Between particle sizes, Al, Mg, and K demonstrate the most variation. At individual sites however, all three elements are more highly concentrated in silt than in sand. Al, Mg, and K all demonstrate statistically significant variation in both sample type and Taylor Valley basins in the sand particle size.

A series of major elemental oxide variation diagrams for individual sites with published values for the Lower Beacon Sandstone, Ferrar Dolerite, and the Basement

Complex directly from the Taylor Valley, McMurdo Volcanics Basanite from Hut Point consistent in composition with the basanitic cinder cones found in the Taylor Valley, and the Upper Continental Crust (UCC) for reference are presented in Figure 2.11. Rock formations of the Dry Valleys and the aeolian sediments of this study demonstrate a wide variation in composition. The geochemical signature of the aeolian sediment samples appears to be largely explained by the mixing of the four MDV rock types.

The alkali metals (Na2O+K2O) are depleted with respect to UCC in both sand and silt particle sizes; MgO, MnO, and CaO are enriched with respect to UCC in both particle sizes. UCC plots within the sand values of TiO2, Fe2O3, and Al2O3. The silt values of

TiO2 and Fe2O3 are enriched with respect to the UCC. Al2O3 values of the silt and size fractions are similar and the difference in UCC is due to the SiO2 range of particles.

Isotopic provenance

The isotopic values for 143Nd/144Nd and and 87Sr/86Sr reported by the TIMS lab at

Boston University are reported in Table 2.3 along with average values for some of the rocks and soils in the Dry Valleys region. While Sr isotope ratios are available from the literature, Nd isotopic values were only found for the McMurdo Volcanics phonolitic lavas from Mt. Erebus and the Ferrar Dolerite. Sr isotopic values from the glaciers are

33 similar to the soils from their respective basins: Commonwealth is similar to Lake Fryxell soil, Canada is similar to Canada soils, and Taylor is similar both Lake Bonney soil and water of Blood Falls (Table 2.3). Glacial silt Nd isotopic values lie intermediate to the

87 86 rock values. Figure 2.12 shows Sr/ Sr versus εNd(0) of the glacial silt isotopic values superimposed on the values provided by Delmonte et al. (2004) for the Vostok and Dome

C cores and potential source areas. Our values do not plot within the field of either ice core, but do plot within the greater ―Dry Valleys‖ field (Figure 2.12).

DISCUSSION

The idea of weathering in cold regions has classically been attributed solely to physical means with freeze-thaw being the predominant weathering mechanism (Hall et al., 2002). Very low temperature and moisture availability in the MDV, however, precludes the freeze-that process; frost action occurs solely in the relatively humid coastal zones (Campbell and Claridge, 1987; Selby, 1971). Instead, the dominant means of physical erosion throughout the majority of the MDV is the expansion of crystals and salts in rock pores and joints (Selby, 1971; Selby and Wilson, 1971). This process is most effective in the disintegration of coarsely crystalline rock, such as the dolerites and granites of the MDV (Selby and Wilson, 1971). The expansion works in much the same way frost action does to disaggregate parent material. Salt weathering is inhibited by the formation and presence of desert varnish (Selby and Wilson, 1971). Desert varnish is a thin coating that stains the outer surface of rocks in arid environment that increases their resistance to weathering by sealing off pores (Campbell and Claridge, 1987). These

34 coatings are enriched in Fe-oxides and form most extensively in the least humid areas and are a manifestation of chemical weathering (Selby, 1971; Selby and Wilson, 1971).

An additional means of rock disintegration, and equally as important, is horizontal and vertical jointing. This process is extremely slow but effective means of breaking down parent rock (Campbell and Claridge, 1987).

The wind regimes of the MDV are similarly effective means of both abrasion and erosion of bedrock (Figure 2.13). Abrasive forces are physically manifested as ventifacts and polished rock surfaces (Campbell and Claridge, 1987; Malin, 1985; Malin, 1984;

Gillies et al., 2009). Ventifacts and polishes can occur in a range of lithologies, they are best developed in fine-grained rocks that are easily eroded by windblown particles

(Gillies et al., 2009). Granite, sandstone, and other coarse-grained rocks are susceptible to cavernous weathering that is generally attributed to both salt weathering and wind abrasion/erosion by removing loosened rock fragments (Campbell and Claridge, 1987).

Erosive forces are demonstrated by the well-recognized aeolian features such as ripple and dune structures, as well as sediments that have been deposited on ice surfaces that have been discussed previously and are the focus of this study (Selby et al., 1974;

Speirs et al., 2008a; Bourke et al., 2009; Calkin and Rutford, 1974). As a fluid medium, wind can only readily transport via suspension fine sand and smaller particles. Sand size and larger particles are moved shorter distances through traction and saltation (Boggs Jr.,

2006). This explains the differences between the flux data of Lancaster (2002) at 1 m and Šabacká et al. (2010) at 0.3 m, as the latter is closer to (if not within) the saltation zone of adjacent soils. Low soil moisture and lack of overland flow outside of stream channels (Fountain et al., 1999b) allows saltation to readily occur year round when winds

35 are strong enough to allow particle deflation. In the Mojave Desert this translates to the highest 6% of wind speeds responsible for 80% of aeolian transport (Lancaster, 1985).

Similarly, wind is responsible for the relatively high degree of sorting that the majority of aeolian material shows. The samples with the greatest percentage of silt are from glaciers in the Fryxell basin, while those with the greatest gravel content are lakes from the Bonney basin and Bull Pass. These trends make sense based on landscape juxtaposition and sample elevation. Bonney basin lakes and Bull Pass samples were collected at low points in the valley adjacent to steep talus slopes forming the northern border (Bonney basin) or northeastern border (Bull Pass) of the associated valley; Fryxell basin glaciers, especially the Commonwealth Glacier with the highest relative percentage of silt, are 20+ m off above the valley floor and removed from the direct influence of gravitational settling. Unfortunately, aside from the ESTs, these fluxes cannot be quantitatively compared to the sediment fluxes of Lancaster (2002) or Šabacká et al.

(2010) because there is no annual or areal control. Alternatively, though less likely, the composition of the clasts might be sufficiently different between sites to account for silt or sand enrichment.

Aeolian sediment composition and source

The first-order weathering provided by physical weathering and gravitational settling is part of the process that reduces rocks to particles small enough to be entrained during the strong wind events of the MDV. As noted previously, wind is responsible for the redistribution of sediments and organic matter throughout the MDV (Lancaster, 2002;

Šabacká et al., 2010) and accounts for the degree of mixing seen in the samples analyzed during the course of this study (Figure 2.11). The 87Sr/86Sr of surface sediments collected

36 from the four major glaciers of the Taylor Valley indicate that they most resemble the soils and lake waters of their respective basins (Table 2.3), providing evidence for a local source of aeolian material. The aeolian material deposited (and subsequently collected) within the MDV is sourced locally and derived from two different sources. The actual source of these separate signatures is complicated by degrees of mixing of parent rock types along with physical/chemical weathering, and external atmospheric or biological deposition that has affected the composition of aeolian material.

The geochemical signature of silts and sands can be explained through the mixing of the four major rock types of the MDV: McMurdo Volcanics basanites, Ferrar Dolerite, metasedimentary Basement Complex, and the Beacon Sandstone. Silts are more mafic in composition while sands are more siliceous. The composition of the silt particle size appears to be most influenced by the volcanic rocks of the McMurdo Volcanics basanites and the Ferrar Dolerite. Relative depletion of Na2O+K2O, CaO, and Al2O3, however, cannot be explained by the mixing of these two end members. Possible explanations for this depletion are contribution of the Basement Complex or the Beacon Sandstone to the sample, and the chemical weathering of the silt fraction. Relative depletion of the labile oxides (Na+K and Ca) indicates chemical weathering is likely. Also, older landscape surfaces are associated with a depletion of potentially erodible material as wind has already deflated the easily removable particles.

Sand size particles are generally bracketed by the Ferrar Dolerite, Beacon

Sandstone and Basement Complex compositions. MgO and MnO both demonstrate relative enrichment in the sand size compared to their three end members. Unlike with the silt size particles, however, this could be explained by interaction/mixing with the

37 basanitic compositions of the McMurdo Volcanics that occur within the MDV. The sand size particles of the Victoria Dunes and Bull Pass trend toward the Beacon Sandstone composition, consistent with the increased exposure of the Beacon Sandstone in the

Victoria Valley as compared to the Taylor Valley.

An indication of the degree of chemical weathering a sediment has experienced is the Chemical Index of Alteration (CIA) that reveals the relative abundance of unweathered material. Values are presented as percentages calculated with the equation:

where CaO* is calcium from the silicate fraction only (Nesbitt and Young, 1982). Each sample analyzed for major elemental composition was ignited beforehand during the process of LOI calculation that efficiently burns off any carbonate-derived calcium.

Average CIA values for the two particle sizes at each aeolian sediment location and the four average rock types were calculated and are presented in Table 2.4. Unweathered rock samples have CIA values near 50, while highly weathered sediments have CIA values near 100% (Nesbitt and Young, 1982). Based on CIA values, both sand and silt particle sizes are relatively unweathered and are generally have lower CIA values than the average values of the rocks from which they are derived. While the CIA values of aeolian material indicate that the Ferrar Dolerite may be important in the compositional evolution of both particle sizes, the signature may be artificial and due to the leaching of the sediments prior to collection. Salt leaching would lower the overall CIA value for the aeolian material by the preferential removal of Na, K, and Ca. Additionally, the average rock types are representative of formations that extend throughout the MDV and

Transantarctic Mountains as a whole that vary in composition (Haskell et al., 1965;

Angino et al., 1962; Faure and Jones, 1973).

CIA values range from 44 to 57%, and there is no significant difference between the CIA values of the silt and sand size particles. The largest difference is ~6%, though is generally closer to 2% indicating that both silt and sand size particles have undergone similar degrees of chemical weathering (Table 2.4). Further support for differing sources of sand and silt is the significant difference between silt and sand LOI values. LOI is a calculation of volatiles released during ignition. The protocol used in the analysis of these samples is intended to completely drive off mineral-bound water, carbonate, organic carbon, and any other volatiles. Thus, higher LOI values in the silt fraction could indicate a greater amount of hydrous minerals present (gypsum, amphiboles, mica), carbonates, and/or volatiles released from igneous rocks. The prevalence of carbonate weathering throughout the hydrologic continuum (Lyons et al., 2003; Lyons et al., 1998;

Fortner et al., 2005), the formation of pedogenic carbonate (Foley et al., 2006), and the presence of various simple salts in all soils (Keys and Williams, 1981) provide evidence for the two former possibilities.

The accumulation of salts cannot be discounted as a possible source of variation in the major element data, though based on the high SiO2 content (>50% by weight) their influence would be diminished. Element patterns that correspond to the major cations in the common salts of the MDV (Na, K, Ca, and Mg) would be affected. Sodium and Ca salts are the most common found in the Taylor Valley and are inversely proportional to each other with increased Na in the east and increased Ca in the west. Magnesium salts are exclusive to mafic rocks that outcrop with greater frequency in the western portion of

39 the Taylor Valley (Keys and Williams, 1981). Many of these salts are readily soluble and are responsible for the flush of solutes seen at the beginning of the melt season (Fountain et al., 1999b; Fountain et al., 1999a). Additionally, relatively insoluble metal oxides

(especially Fe-oxide) could also affect the patterns in their respective metal oxide. The presence of salt is more easily discussed from the solute perspective presented in Chapter

3, including estimates of the solubility and saturation states of species based on experimental leachates.

Given that silt and sand particles are derived from different sources and a similar degree of chemical weathering, the difference in particle size must be due to mechanical means. The Goldich dissolution series, a means of predicting relative stability of igneous minerals at surface conditions, indicates that quartz is the mineral most resistant to weathering (Figure 2.14) (Goldich, 1938). The Beacon Sandstone at Windy Gully (south of the Taylor Valley) contains > 95% quartz clasts. A study conducted in the Victoria

Valley at ground surface and 0.75 m found deposits to be dominated by well-rounded quartz grains, especially within the saltation zone, with appreciable amounts of AlCa-,

AlK-, CaFeMg-, and FeMg-bearing silicates in sand size particles measured using a

QEMSCAN automated electron microscope (Speirs et al., 2008b). The quartz- enrichment seen in these sediments is less than that of warm deserts where dunes may exceed 90% quartz (Sepälä, 2004). Sepälä (2004) asserts that this is likely due to the fact that sediments in cold environments are often first-cycle weathering products. In the

Victoria Valley and MDV as a whole, this is certainly true and manifested in the widespread ventifaction and abrasion of bedrock (Malin, 1985; Malin, 1984; Gillies et al., 2009). The less stable mafic/volcanic silt composition is likely to be less resistant to

40 physical weathering and wind abrasion compared to the more siliceous sands, though mineralogical and/or grain characterization studies must be carried out for confirmation.

Spatial trends of aeolian sediment composition

In general, glacial surface and EST samples follow the same trends. These trends are complicated in the lake surface and aeolian landforms by other in situ processes that appear to arise from the smaller degree of removal from the physical weathering system compared to glacial surfaces and EST samples. As discussed previously with respect to grain size, the areas where glacial surface samples were collected were well-removed from valley walls and other influences. Similarly, the EST collection system and location on relatively unimpeded ground ensures that the sample is purely aeolian in nature.

Aeolian landforms and lake surfaces are not necessarily afforded these same conditions, nor is there any guarantee that the sample is purely aeolian in nature because physical weathering and gravitational settling are active in all areas of the MDV. Lakes, especially E. Lake Bonney and Mummy Pond, abut steep talus slopes at the base of valley walls; the majority of aeolian landforms are blown up against obstacles such as glaciers or valley walls. Therefore, the discussion of trends that follows will be based on the similar patterns seen in the glacial surface and EST samples and tie into lake surface and aeolian landforms when discernible.

Phosphorus demonstrates the greatest spatial variation manifested in a clear east- to-west decreasing trend in all four sample types (Figure 2.10j). Apatite

(Ca5(PO4)3(F,Cl,OH)) is a common accessory mineral of igneous rocks (Nesse, 2000), such as granodiorite, diorite and granite and has been identified in thin sections of the granite from the Taylor Valley (Haskell et al., 1965). Another less common accessory

41 mineral is monazite (((Ca,La,Th,Y)PO4) that is found in granitoids of the Transantarctic

Mountains (Simpson et al., 2002). These minerals are the lithogenic source of P to the

MDV ecosystem, but are not entirely responsible for the spatial P-trends seen in the

MDV (Gudding, 2003). Landscape age models of nutrient availability indicate larger pools of P in younger landscapes (Filippelli and Souch, 1999; Vitousek et al., 1997;

Vitousek and Farrington, 1997). The Ross Drift found in the Fryxell basin in the Taylor

Valley is younger than the Ross Drift in the western portion of the Valley and has experienced ~50,000 fewer years of chemical weathering (Figure 2.16) (Burkins et al.,

2000). The tills of the Fryxell basin are more favorable for the establishment of biological communities (Barrett et al., 2006). Therefore, increased P-content to the east in these samples is related to landscape age and lithology. The P-content of stream sediment from the Taylor Valley also reflect the same magnitude and east-to-west pattern

(Gudding, 2003; Nezat and Lyons, 2000).

Nitrogen, however, is greatly depleted compared to P in aeolian sediments and is only detectable in aeolian sediments at Lake Fryxell and the Eastern Commonwealth glacier in eastern Taylor Valley (Table 2.2). The flux of N compared to P estimated in aeolian material presented in Barrett et al. (2007) reflects the same igneous rock- enrichment of P compared to N that is gained through external means. That N is only detected in Lake Fryxell may be an indication of biological N-input to the aeolian source of sediments, as Lake Fryxell has the greatest coverage of algal mats that could be weathered by wind much like sediment.

The same inland decrease is seen in organic C, as well (Table 2.2) and can be explained in the same way as N. Calculated inorganic C, however, shows the greatest

42 enrichment in the middle portion of the Taylor Valley, especially between Lake Hoare and the east lobe of Lake Bonney. This enrichment coincides with a metamorphic complex containing ―many carbonate beds‖ present at the surface on the south side of the

Taylor Valley just west of the Howard Glacier and opposite the Suess Glacier (Angino et al., 1962) as well as an abundance of lacustrine carbonate (Hendy, 2000). Lacustrine carbonate is assumed to be precipitated following consumption of CO2 by photosynthetic organisms and is seen throughout the shallow waters of the MDV lakes (Hendy, 2000).

These deposits are located throughout the Taylor Valley and record the rise and fall of paleolakes, such as Glacial Lake Washburn that inundated the Taylor Valley between

~22,800 and 8,500 years before present (Chinn, 1993; Stuiver et al., 1981). Subsequent wind erosion of these deposits would transfer the carbonates locally and mostly within the Hoare basin, as is seen in the data.

Salt content of soil in the MDV can reach more than 40% by volume (Claridge and Campbell, 1977), which should have an appreciable effect on the compositional trends in related major element. For example, Keys and Williams (1981) found an inland increase in Ca- and Mg-containing salts and an inland decrease in Na-containing salts.

These trends are also seen in the EST and glacial surface samples, as well as generally realized in lake surface and aeolian landform samples (Figure 2.10 f-h). This does not detract from the hypothesis of separate sources for sand and silt particles because the trends are the same in both size fractions, but does again suggest that the dust being deposited in these areas is of local origin. However, the model would be made stronger by an examination of salt content at different size scales.

In general, the CIA value follows the pattern aeolian landforms < lakes ≈ ESTs < glaciers and spatially correlates with proximity to water in the MDV environment (Table

2.4). Niveo-aeolian influences aside, aeolian landforms are generally removed from the hydrologic continuum of the MDV. Lake and EST sediments are both from essentially the valley floor and are exposed to stream flow. Glacial sediments were collected within the ablation zone and accordingly may have been exposed to supraglacial stream flow despite best efforts to collect prior to the initiation of summer melt.

Iron, Mn, and Ti are susceptible to the formation of secondary oxides during chemical weathering and are often constituents of minerals that are considered to be refractory. In the MDV, secondary oxides are commonly found as coatings that aid in the resistance to wind abrasion. These stains are generally Fe-enriched (Campbell and

Claridge, 1987), which is logical because Fe concentrations at both sand and silt size are on the order of five times the concentration of Mn or Ti (Figure 2.10 b, d, e; Figure 2.11;

Appendices B and C). Refractory minerals are also among the most resistant to weathering and could show relative enrichment in aeolian sediment. On the Victoria

Dunes, lag deposits of heavy, dark, opaque minerals were noted. Calkin and Rutford

(1974) analyzed deposits via heavy mineral separation using bromoform and centrifugation and found them to consist of zircon, apatite, pyroxene, amphibole, garnet, biotite, tourmaline, and opaques. Relatively common crystalline opaques containing Fe,

Mn, or Ti include ilmenite (FeTiO3) and magnetite (FeFe2O4) that are associated with igneous and metamorphic rocks (Nesse, 2000). Additionally, all three of these trace elements are concentrated in igneous rocks, as referenced by the increased concentration in the McMurdo Volcanic basanites and Ferrar Dolerite compared to the Basement

Complex and Beacon Sandstone (Figure 2.11) (Roser and Pyne, 1981). Within the samples examined for this study, there is an inland decrease in Ti and inland increases in

Fe and Mn (Figure 2.10 b, d, e). The westward increase in iron likely reflects the weathering of more iron-rich minerals (and desert varnishes) coincident with the exposure of the Ferrar Dolerite, granitic complexes, and McMurdo Volcanic-related cinder cones further east in the Taylor Valley. Wind sorting is probably responsible for the relative decrease in the sand size compared to the silt size of these materials, as

Fe/Ti/Mn are generally part of a heavy mineral fraction.

Potential inland transfer of aeolian material

Strontium and Nd isotopic analysis of the smallest grain size readily available from glacial surface deposits indicate that the MDV do not contribute to the ice core dust record. However, the MDV, New Zealand, and South American fields overlap, so a contribution from all of these sources cannot be excluded (Figure 2.12) (Delmonte et al.,

2004). In addition to katabatic wind direction (see introduction and wind systems), prevailing conditions at present and the Last Glacial Maximum (LGM) account for the preclusion of the MDV as a potential source.

Since increased dust flux in ice cores correlates to glacial stages, but during LGM, ice-free areas would have been much more restricted than at present due to the growth of the East and West Antarctic Ice Sheets, and the development of a larger lake systems in the MDVs (Denton and Hughes, 2002; Hall and Denton, 2000). Perhaps more compelling is the lack of sulfate and carbonate salts in the ice core dust record (Delmonte et al., 2004). At present, gypsum and carbonates are found throughout the MDV in appreciable quantities (Keys and Williams, 1981; Campbell and Claridge, 1987), but are

45 not present in Antarctic ice (Delmonte et al., 2004). These observations indicate that, compared to other potential sources, the MDV probably do not contribute extensively (or at all) to the East Antarctic dust flux

Additionally, a model-based investigation of the emissions and deposition of several source areas indicate a predicted net deposition of 0.06 Tg a-1 from all of inland

Antarctica compared to a net emission of 7 Tg a-1 from the band defined by 15S-50S considered to be the source area of southern hemisphere dust (Li et al., 2008). Annual predicted dust deposition from four potential source areas are graphically presented in

Figure 2.15. South America and Australia demonstrate the greatest amount of deposition, though eastward wind circulation indicates that East Antarctica is more impacted by

South American dust. Li et al. (2008) report relative contributions from the four potential source areas for the 0.06 Tg a-1 of predicted deposition: 50% South American, 35%

Australian, 5% South African, and 10% Northern Hemisphere. If it assumed that greater than 50% of the deposition during LGM was similarly due to the South American dust source, then it follows that the isotopic signature would reflect a largely South American source, as is noted in the literature (Basile et al., 1997; Delmonte et al., 2002; Delmonte et al., 2004; Hinkley and Matsumoto, 2001; Delmonte et al., 2008; Gaiero, 2007;

Delmonte et al., 2010; Revel-Rolland et al., 2006). Plus it appears the dust being deposited today is local rather than regional in extent. It is not well-mixed, suggesting little transport between basins and that this is a ―source limited‖ system.

CONCLUSIONS

The geochemistry presented in this study reveal a complex interaction of wind with major bedrock types and external influence in the formation of salts and interface with biota. Isotopic data indicates a local source of aeolian material, which is explained by the mixing of the four source rocks to varying degrees with spatial trends relating to salt composition and basin biology. Evidence also suggests that the sources for sand and silt size particles may be distinct with a more mafic/igneous signature in silts and a siliceous/sedimentary signature in sands and an equal degree of weathering. This could also be explained by wind interaction, as deflation of silt size particles is more readily achieved and a more removed source of silt is possible. The inland transfer to East

Antarctica, however, is unlikely based on the evidence presented here. Isotopic data from aeolian silts collected on glacial surfaces indicate that they are not related to the East

Antarctic ice core records. Thus, wind interaction is a strong, though generally local to regional, force in the MDV.

The results of this study are also important from a potential weathering point of view. These aeolian sediments are weathered throughout the hydrologic continuum during the summer melt season. This is the topic of the next investigation (Chapter 3) concerning experimental leaching and release of bioavailable solutes and nutrients to the

MDV ecosystem.

Figure 2.1 Base map of the Taylor Valley with dots indicating areas of EST transects (red) and Sensit acoustic wind eroding mass sensors (blue).

Figure 2.2 Wind directions and sources in the Taylor Dry Valley, Antarctica.

Figure 2.3 A cross-section of the Taylor Dry Valley indicating location of major geologic barriers, glaciers, and lakes from Lyons et al. (2000).

Figure 2.4 Generalized stratigraphic sequence of Southern Victoria Land, Antarctica from Campbell and Claridge (1987).

Figure 2.5 Examples of ice surface sample sites on (a) the Taylor Glacier in Taylor Valley and (b) Lake Fryxell in Taylor Valley.

Figure 2.6 Examples of aeolian landforms sampled including (a) the Packard Dunes in Victoria Valley and (b) Lake Hoare Beach in Taylor Valley.

Figure 2.7 An example of an elevated sediment trap from the Lake Fryxell transects.

% gravel

0 100

10 90

20 80

30 70

40 60

50 50

60 40

70 30

80 20

90 10

100 0 % silt 0 10 20 30 40 50 60 70 80 90 100 % sand

Figure 2.8 Grain size distributions of bulk aeolian sediments differentiated by type and location, and then plotted as percent gravel, sand, and silt. Basins/locations are color- coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass, gray=Wright Valley). Surfaces are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST).

10 Glaciers - Sand Glaciers - Silt EST - Sand 8 EST - Silt Lakes - Sand Lakes - Silt Landforms - Sand 6 Landforms - Silt

LOI (wt. %) LOI (wt. 4

Defile

Taylor

Howard

Canada

Bull Pass

Nussbaum

Hoare EST Hoare

Fryxell EST Fryxell Hoare Lake

Lake Fryxell Lake

Hoare Beach Hoare

Wright Valley Wright

Mummy Pond Mummy

Victoria Dunes Victoria

Commonwealth

E. Lake Bonney E. Lake

W. Bonney EST W. Bonney

W. Lake Bonney W. Lake Location

Figure 2.9 Weight percent LOI from the sand and silt particle fraction of each location. Sample types are organized east to west. Error bars represent one standard deviation where multiple samples were collected.

400 (a)

300

)

-1

200

Si (g kg Si

100

Glaciers - Sand Glaciers - Silt 0 EST - Sand EST - Silt 16 Lakes - Sand (b) Lakes - Silt Landforms - Sand Landforms Silt 12

)

-1 8

Ti (g kg Ti

Defile

Taylor

Howard

Canada

Bull Pass

Nussbaum

Hoare EST Hoare

Fryxell EST Fryxell Hoare Lake

Lake Fryxell Lake

Hoare Beach Hoare

Wright Valley Wright

Mummy Pond Mummy

Victoria Dunes Victoria

Commonwealth

E. Lake Bonney E. Lake

W. Bonney EST W. Bonney

W. Lake Bonney W. Lake Location Continued Figure 2.10 Absolute concentrations of major elements in sand and silt size fractions. Sample types are organized east to west. Error bars represent one standard deviation where multiple samples were collected. 54

Figure 2.10 Continued 100 (c)

) 60

-1

Al (g kg Al 40

20 Glaciers - Sand Glaciers - Silt EST - Sand 0 EST - Silt 100 Lakes - Sand Lakes - Silt (d) Landforms - Sand Landforms - Silt 80

) 60

-1

Fe (g kg 40

Defile

Taylor

Howard

Canada

Bull Pass

Nussbaum

Hoare EST Hoare

Fryxell EST Fryxell Hoare Lake

Lake Fryxell Lake

Hoare Beach Hoare

Wright Valley Wright

Mummy Pond Mummy

Victoria Dunes Victoria

Commonwealth

E. Lake Bonney E. Lake

W. Bonney EST W. Bonney

W. Lake Bonney W. Lake Location Continued

Figure 2.10 Continued 2.5 (e)

2.0

) 1.5

-1

Mn (g kg 1.0

0.5

Glaciers - Sand 0.0 Glaciers - Silt 60 EST - Sand EST - Silt (f) Lakes - Sand Lakes - Silt 50 Landforms - Sand Landforms - Silt

)

-1

Mg (g kg 20

Defile

Taylor

Howard

Canada

Bull Pass

Nussbaum

Hoare EST Hoare

Fryxell EST Fryxell Hoare Lake

Lake Fryxell Lake

Hoare Beach Hoare

Wright Valley Wright

Mummy Pond Mummy

Victoria Dunes Victoria

Commonwealth

E. Lake Bonney E. Lake

W. Bonney EST W. Bonney

W. Lake Bonney W. Lake Location Continued

Figure 2.10 Continued 100 (g)

) 60

-1

Ca (g kg 40

Glaciers - Sand Glaciers - Silt 0 EST - Sand EST - Silt 30 Lakes - Sand (h) Lakes - Silt Landforms - Sand Landforms - Silt

)

-1

Na (g kg 10

Defile

Taylor

Howard

Canada

Bull Pass

Nussbaum

Hoare EST Hoare

Fryxell EST Fryxell Hoare Lake

Lake Fryxell Lake

Hoare Beach Hoare

Wright Valley Wright

Mummy Pond Mummy

Victoria Dunes Victoria

Commonwealth

E. Lake Bonney E. Lake

W. Bonney EST W. Bonney

W. Lake Bonney W. Lake Location Continued

Figure 2.10 Continued 30 (i)

)

-1

K (g kg (g K

Glaciers - Sand Glaciers - Silt EST - Sand 0 EST - Silt Lakes - Sand 3 Lakes - Silt (j) Landforms - Sand Landforms - Silt

)

-1

P (g kg (g P

Defile

Taylor

Howard

Canada

Bull Pass

Nussbaum

Hoare EST Hoare

Fryxell EST Fryxell Lake Hoare

Lake Fryxell

Hoare Beach Hoare

Wright Valley Wright

Mummy Pond Mummy

Victoria Dunes

Commonwealth

E. Lake Bonney

W. Bonney EST W. Bonney

W. Lake Bonney Location

4.0 V 12 3.5 10 3.0 V 8 2.5

(wt. %) (wt. 2.0 6

2 1.5 F 4

TiO 1.0 %) MgO (wt. F 2 0.5 B B 0.0 S 0 S 14 V 0.30 12 0.25 10 0.20 V 8 F

(wt. %) (wt. 0.15 3 6

O 2 B 0.10 F 4 %) MnO (wt.

Fe 0.05 B 2 S 0 0.00 S 12 18 V 10 16 F B 14 8 F V 12 6

(wt. %) (wt. 10

4 O

CaO %) (wt.

B l 8

A 2 6 S 0 S 4 8 1.0 B 0.8 V 6

(wt. %) (wt. V 0.6

4 %) (wt.

F 5 0.4

O + K

P 2 2 0.2

Na S F B 0 0.0 S 40 50 60 70 80 90 40 50 60 70 80 90 SiO SiO 2 (wt. %) 2 (wt. %)

Figure 2.11 Major elemental oxide variation diagrams. Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass, gray=Wright Valley). Filled characters are sands, unfilled characters are silt. B=Basement Complex, F=Ferrar Dolerite, S=Beacon Sandstone, V=McMurdo Volcanics Basanite (Roser and Pyne, 1981). Yellow star is the upper continental crust (Taylor and McLennan, 1981).

87 86 Figure 2.12 εNd(0) and Sr/ Sr fields of ice core dust and potential source areas superimposed with values from this study. Plot modified from Delmonte et al. (2004). 60

Figure 2.13 Examples of weathering and wind in the MDV: (a) cavernous weathering of a coarse grained granite, (b) ventifaction and polishing of a fine grained volcanic rock, and (c) expansion due to salt weathering (expected over frost action due to the elevation). Photo is taken toward the south from the plateau beneath the summit of the Matterhorn across the Bonney basin to the Kukri Hills. The Solas Glacier flows from the Kukri Hills in the center back. For scale, person in the background is approximately 1.5 m tall.

High Olivine Plagioclase (Ca-Rich) Pyroxene (Augite*)

Amphibole (Hornblende*) Relative Pagioclase Weathering Biotite (Na-rich) Potential

Orthoclase

Muscovite

Quartz Low *mineral directly referred to in orginal work Figure 2.14 Schematic of the weathering potential of various silicate minerals modified from Goldich (1938). Dotted line refers to the solid solution series between Ca and Na plagioclase.

Figure 2.15 Annual mean dust deposition from (a) South America, (b) Australia, (c) South Africa, and (d) Northern Hemisphere. White dots indicate source areas. Units are kg m-2 a-1.

Figure 2.16 Map of the tills related to glaciations and their age in the Taylor Valley from Burkins et al. (2000).

Table 2.1 Locations of elevated sediments traps from adapted from Lancaster (2002) and Šabacká et al. (2010) Location Total Sediment Flux Height of EST Period of (g m-2 a-1) (m) data collection Lancaster 2002 Explorers Cove 27.87 1 1997-2000 Lake Fryxell 1 1 1995-2000 Lake Hoare 0.86 1 1995-2000 Lake Bonney 110.53 1 1995-2000 Lake Brownworth 441.8 1 1997-2000 Lake Vanda 227.5 1 1997-2000 Lake Vida 39.56 1 1997-2000 Commonwealth Glacier 0.26 1 1997-2000 Canada Glacier 0.43 1 1997-2000 Howard Glacier 0.47 1 1997-2000 Taylor Glacier 3.73 1 1997-2000 Šabacká et al. 2010 Lake Fryxell 355.15 0.3 1999-2008 Lake Hoare 470.1 0.3 1999-2008 W. Lake Bonney 70.72 0.3 1999-2008

Table 2.2 Nitrogen, carbon, and organic carbon data presented as weight percentages for each sample. Samples are subdivided into sample type and organized east to west. Sample Name N C OC ΔC-OC* (mg g-1) (mg g-1) (mg g-1) (mg g-1) Aeolian Landforms Hoare Beach BDL 0.5 BDL 0.5 Defile BDL 0.9 0.5 0.4 Victoria Dunes BDL BDL BDL - Bull Pass BDL BDL BDL - W. Lake Bonney BDL 1.2 1.2 - Elevated Sediment Traps 1 Fryxell EST - - 0.2 - Hoare EST - - 0.1 - W. Bonney EST - - 0.1 - Glaciers E. Commonwealth 0.9 7.8 8.1 - W.Commonwealth BDL 5.5 5.5 - E. Canada BDL 3.9 3.0 0.9 W. Canada BDL 1.5 1.2 0.3 Howard BDL 4.6 2.6 2.0 Taylor BDL 0.5 0.4 0.1 Lakes Lake Fryxell 1.2 12.7 14.3 - Lake Hoare BDL 0.5 0.5 0.0 Mummy Pond BDL 1.5 BDL 1.5 E. Lake Bonney BDL 0.8 0.8 - BDL=below detection limit: 0.8 mg N g-1 and 0.4 mg C or OC g-1 *Assumed to be carbonate; 1Nine-year average values from Šabacká et al. (2010)

Table 2.3 Sr and Nd isotopic values for three silts and one bulk glacial sediment sample from this study and rocks, waters, and soils from the Dry Valleys literature 87 86 143 144 Sample ID Sr/ Sr Nd/ Nd εNd(0) Taylor 0.7149360 0.51216 -9.29 Howard 0.7089927 0.51236 -5.38 Canada 0.7110975 0.51220 -8.51 Commonwealth 0.7094701 0.51224 -7.72 Erebus Phonolite Lava1 0.70301-0.70309 0.51287-0.51293 4.56 - 5.74 Carbonates of Beacon 0.7260-0.7291 Ferrar Dolerite3 0.71050-0.71270 0.51212-0.51216 -10.07 - -9.29 Granite Harbour Intrusives4 0.71897-0.74237 0.51198-0.51202 -12.79 - -12.02 Lake Fryxell soil5 0.7092 Canada Glacier soil5 0.7101 Lake Bonney soil5 0.7136 Blood Falls5 0.7136 143 144 εNd(0) calculated with using N/ NdCHUR=0.512636 from Basile et al. 1999. 1Sims et al. (2008); 2Faure and Barrett (1973); 3Foland and Marsh (2005); 4Armienti et al. (1990)5Jones and Faure (1978)

Table 2.4 Chemical Index of Alteration values for average location and rock type data CIA (%) Rock Sand Silt Average Rock Type * Basement Complex 59.03 Beacon Sandstone 71.07 Ferrar Dolerite 55.68 McMurdo Volcanics 46.68 Aeolian Landforms Hoare Beach 51.99 50.03 Defile 50.02 44.26 Nussbaum 54.50 55.08 Victoria Dunes 46.61 48.11 Wright Valley 54.49 53.32 Bull Pass 48.23 50.27 Elevated Sediment Traps Fryxell EST 54.12 - Hoare EST 54.78 - W. Bonney EST 54.72 - Glaciers Commonwealth 56.90 55.74 Canada 55.10 54.03 Howard 51.85 54.23 Taylor 52.67 56.58 Lakes Lake Fryxell 53.83 53.60 Lake Hoare 54.24 53.06 Mummy Pond 56.52 - E. Lake Bonney 53.92 53.77 W. Lake Bonney 53.27 49.19 *Calculated from major element data of Roser and Pyne (1981)

CHAPTER 3

EXPERIMENTAL LEACHING OF AEOLIAN SEDIMENTS, MCMURDO DRY

VALLEYS, ANTARCTICA

INTRODUCTION

The interface between the aeolian and fluvial sediment supply and its resulting geomorphological and biological implications is a relatively new topic of interest

(Bullard and Livingstone, 2002). Since the MDV receive very little precipitation and have very low soil moisture, this interface between the aqueous and aeolian systems is more limited compared to more temperate climates. The importance of moisture is highlighted in the habitat suitability of soil organisms: greater diversity is shown in lower elevation, coastal regions where moisture is more plentiful (Courtright et al., 2001;

Kennedy, 1993). The relative reservoirs and nature of water in Antarctica and their relationship to biodiversity are presented graphically in Figure 3.1. The variation and sensitivity to temperature of the hydrologic cycle in the McMurdo Dry Valleys (MDV) is very important to the ecosystem due to the extreme seasonality and paucity of water.

During the summer melt season, ephemeral streams produced from glacial melt connect the two major sources of water (ice) in the MDV: glaciers and endorheic lakes (Figure

3.2) (Fountain et al., 1999b; Lyons et al., 1998)(Fountain et al., 1999b; Lyons et al.,

1998).

Sediment transport by strong seasonal winds (―aeolian transport‖ of Figure 3.2) is an important physical process in this landscape and it is expressed in widespread aeolian landforms and deposits in the MDV (Lancaster, 2002). Katabatic winds from the west dominate during the winter months and are probably responsible for the majority of dust and salt deposition. Easterly winds from the coast, prominent during the summer, contribute to the dust budget through the addition of salts and marine aerosols (Fountain et al., 1999b). These two wind regimes potentially ―recycle‖ chemicals by transporting valley floor sediments onto glacial and lake ice surfaces where they can then be solubilized and contribute solutes to the aquatic system (Lyons et al., 1998). These sediments change the albedo of the ice and can eventually lead to the formation of cryoconite holes on glaciers that affect the composition of glacial melt (Fountain et al.,

2004) as well as contribute to the surface topography of the permanent surface ice of the lakes (Squyres et al., 1991). Additionally, aeolian deposition in stream channels (Lyons et al., 1998) produces a new and changing source of material to be weathered during the melt season.

Melt generation and the evolution of glacial meltwater

Cryoconite holes on the Taylor Valley (a major valley in the MDV area, Figure

1.1) glaciers range in diameter from 5 to 145 cm and from 4 to 56 cm in depth (Fountain et al., 2004). Larger scale ―cryolakes‖ from 5 to 50 m in diameter are also formed through the same processes (Fountain et al., 2004; Bagshaw et al., 2010). As the dust melts into the glacier surface and creates the cryoconite, the hole that is formed is better able to trap more windblown material and, thereby, further decreasing the overall albedo of the area and deepening the depression (Figure 3.3) (Fountain et al., 2004; Fountain et

al., 2008; Tranter et al., 2004). Thus, a positive feedback mechanism is set up that will proceed until an equilibrium condition is attained when the rate of cryoconite melt-in equals the rate of surface ablation in the glacier (Wharton et al., 1985). Metabolic activity and biogenically-produced heat can account for differences in depth (McIntyre,

1984). Biogenic communities within these depressions occur on glaciers in the Taylor

Valley and worldwide and include auto- and heterotrophic organisms including algae, bacteria, rotifers, tardigrades, viruses, and insects where conditions permit that vary by location.

In the Taylor Valley cryoconite holes cover ~4-5% of the ablation zone and are ice-lidded except during especially warm summer seasons, such as the summer of 2001-

2002 (Fountain et al., 2004; Fountain et al., 2008). The ice lids are able to persist because their thickness (30-40 cm) is much greater than the low rates of ablation on the glaciers (~8 cm a-1) (Tranter et al., 2004). Liquid water is maintained during the summer months because the frozen cryoconite material still functions to lower the albedo of the ice, and melting proceeds in the subsurface (Figure 3.3). In a detailed investigation of cryoconite holes of the Canada Glacier in Taylor Valley, Fountain et al. (2008) found that temperatures in the ice-lidded cryoconite holes were commonly 5º C warmer than ambient air temperatures.

The composition of meltwater within these holes is affected both by the dissolution of the cryoconite sediment and the fauna present in the hole. Bagshaw et al.

(2007) summarized the processes controlling cryoconite solute chemical composition as

(1) degree of dissolution of cryoconite material, (2) the precipitation and dissolution of carbonate minerals, (3) summer net photosynthesis, and (4) winter net respiration.

Tranter and others (2004) found that the pH of the Taylor Valley cryoconite holes could reach values of up to 11. Dissolution of CaCO3 from the windblown dust could only account for pH values of <10.5 and that algal photosynthesis and subsequent CO2 uptake within a closed system must be occurring to further increase pH values and CaCO3 saturation in the melt water (Tranter et al., 2004).

Aside from internal biogeochemistry, there is interest in the physical connection of cryoconite holes and the impact on glacial hydrology and meltwater biogeochemistry.

Non-lidded holes, typical of more temperate glaciers, are directly connected to the supraglacial drainage system and are flushed on a seasonal basis. Lidded holes, however, can remain isolated from the supraglacial (or englacial) system for years at a time

(Fountain et al. 2004; Bagshaw et al. 2007). Based upon the late summer constant value of electrical conductivity, englacial drainage connects the cryoconite holes on the Canada

Glacier, but it is unknown how these connections develop (Fountain et al., 2004). From comparison of runoff from streams and cliff melt, it is estimated that cryoconite holes could be responsible for <13% of total volume of runoff during a melt season, provided all holes are hydraulically connected. The best estimate is that 50% of holes are connected (Fountain et al., 2004).

Cryoconite influx may have important biogeochemical consequences to the overall aquatic system of the MDVs. Glacial meltwater enters channels to form ephemeral streams flowing into the terminal lakes of the MDV (Figure 3.2). The overall chemical composition of the Commonwealth, Canada, Howard, and Taylor Glaciers of the Taylor Valley differ from one another and the reasons for these differences are not fully realized. Previous work has supported the role of aeolian material dissolution as at

least a partial explanation for these differences (Lyons et al., 2001). A later study of supra- and proglacial streams of the Canada Glacier found the geochemistry of melt was different with the eastern (coastal) side being influenced by marine salts whereas calcareous dust has a greater influence on the westward (inland) side of the Canada

Glacier (Fortner et al., 2005). There is a general spatial decrease in solute concentration on glaciers westward (inland) and with increasing elevation (Witherow et al., 2006; Keys and Williams, 1981).

Structures similar in ecological function to cryoconite holes also exist within the permanent ice cover (Priscu et al., 1998). When solar radiation in continuous during the austral summer, melt pockets form in the lake ice and are associated with aeolian-derived sediments that host microbial populations. These sediments reach an equilibrium depth between the upward motion of the ice through ablation at the surface and freezing at the bottom and downward movement of the sediments as they melt into the lake ice surface

(~2 m in Lake Fryxell) (Squyres et al., 1991; Jepsen et al., 2010; Priscu et al., 1998).

Much like the sediment in cryoconite holes, the interaction of the sediment with liquid water in these pockets solubilizes solutes and nutrients that can potentially affect the chemistry of lake surface waters if they are able to migrate through the ice cover (Jepsen et al., 2010; Priscu et al., 1998; Adams et al., 1998). Vertical cracks in the ice below these sediment covers and the thinning of the perennial ice cover are the major pathways of within-ice sediment and liquid water removal and release into the lake water (Nedell et al., 1987; Adams et al., 1998). Since liquid water content in the permanent lake ice can reach 40% (Fritsen et al., 1998), solubilization and removal of nutrients to the lake water is a potentially significant process. In addition, aeolian sediment can be exposed directly

to and solubilized by lake water when it deposited into cracks that penetrate completely through the ice surface (Squyres et al., 1991; Nedell et al., 1987). Based on textural studies, the majority of sedimentation to the lake bottoms is aeolian-derived and may be responsible for supplying nutrients to blue-green algal mats that inhabit the lake bottoms of the Taylor Valley and may be analogs to ice-covered lakes that may have existed on

Mars (Nedell et al., 1987).

Objectives of and precedence for the experimental leaching of MDV aeolian sediments

While studies have investigated the chemistry of the stream and lake waters and downstream solute patterns of the Taylor Valley aquatic systems (Green et al., 1988;

Lyons et al., 2003; Lyons et al., 1998; Lyons et al., 1998), few previous studies have dealt with the characteristics of the dissolution products of the aeolian material that may greatly influence the geochemical character of the initial melt water.

The following study was designed to investigate the importance of aeolian material collected on solute chemistry. Experiments were established to address the following hypotheses:

1. The dissolution of aeolian material is a viable source of nutrients to the

MDV ecosystem.

2. The dissolution products of aeolian material contribute quantitatively to

the hydrochemistry of glacial meltwater, proglacial streams, and terminal

lakes.

METHODS

Aliquots of the bulk, homogenized samples from 18 sites were collected for leaching experiments prior to particle size analysis as described in the methods section of

Chapter 3. Six samples were from glacial surfaces, five from lake surfaces, three from the EST arrays described previously, and four from various aeolian landforms. Sample names and identifying information is presented in Appendix A. It should be noted that

Nitrile gloves were worn during all portions of the leaching experiment and sample loading for analysis to prevent contamination.

Sediment leaching experimental design

A two-step leaching experiment was conducted on each of the 18 bulk samples to ascertain what was readily dissolved and, therefore, what would be readily available to the aquatic system during times of melt. The experiment was designed to investigate the role of meltwater and freeze-thaw mechanisms on the dissolution of material as originally described by Lyons and Welch (1997). A sediment to water ratio of 1:2 was used to roughly simulate cryoconite hole conditions on glaciers.

In addition, a method blank was performed following the same procedures as the sample leaching experiments with no sediment.

For the first leach (―Leach 1‖), 25 g of sample sediment was placed in a clean

HDPE sediment storage container and spread in an even layer. Fifty mL of 18Ω deionized water chilled to 3º C (pH assumed to be ~5.7) was carefully poured over the sediment, the container capped, and the slurry agitated for 5 minutes. The leachate was collected by filtering through a filter tower with 0.4 μm Whatman® Nuclepore track-etch membrane filter into clean 125 mL Nalgene® low density polyethylene (LDPE) bottles.

Care was taken to decant the leachate so that as little of the sediment as possible was transferred onto the filter.

The second leach (―Leach 2‖) was performed on the same sediment used in the first leach. Another 50 mL of 3º C 18Ω deionized water was added; then the container capped and placed in a -20º C freezer. After approximately 72 hours, the containers were removed and placed in a laminar flow hood to melt. Once all visible ice had melted, the leachate was again filtered into clean Nalgene® LDPE bottles as described above.

Leachates were stored at 4º C until time of analysis for major ions, nutrients, reactive silica, and manual iron as described below. Due to volume limitation and number of analyses to be conducted, pH could not be measured on a majority of the leachates and are not reported here.

Major ion analysis

+ + 2+ 2+ - - 2- Major cations (Na , K , Ca , Mg ) and anions (NO3 , Cl , SO4 ) were analyzed on a Dionex DX-120 ion chromatograph (IC) with an AS40 automated autosampler as originally described by Welch et al. (1996). At the start of each run, six incrementally increasing standards were evaluated to create a calibration curve from which sample concentrations were later calculated. Every ten samples a mid-range standard and duplicate sample was analyzed to verify reproducibility of results and account for instrument drift. Samples were diluted and reanalyzed if their concentrations were found to be greater than the most concentrated standard. Precision was determined to be ±10%

- for cations and ±5% for anions. Alkalinity (HCO3 ) was calculated through charge balance according to the equation:

The average error between calculated and measured alkalinity values in over 900 MDV stream samples from multiple summer seasons has been calculated to be ±14%, and is the best estimate for alkalinity precision (Lyons et al., 2010).

Nutrient analysis

- - 2- Samples were analyzed for NO2 +NO3 , PO4 , total nitrogen (TN), total phosphorous (TP), and SiO2 on a Skalar San++ Continuous Flow Analyzer with SA 1050

Random Access Autosampler. For each analysis, a six-point calibration curve was initially created from which sample concentrations were internally calculated using the

San++ Flow Access software for Windows. Every ten samples, a mid-range standard and blank were run to account for instrument drift. As sample concentrations were calculated while the analysis was in progress, if a sample was found to be more concentrated than the most concentrated standard it was diluted and added to the end of the run. Precision was found to be ±5%

Iron analysis

Only leachates from Leach 1 were analyzed for iron content via a Ferrozine method. All sample, standard, and solution containers were soaked in 10% HCl overnight, and then rinsed with 18Ω deionized water three times and placed in a laminar flow hood to dry. Leachate aliquots of 2 mL were transferred into clean 20mL scintillation vial along with 20 μL concentrated HNO3. The following day 0.2 mL 1.5M hydroxylamine solution, 0.2 mL 0.01M ferrozine solution, and 0.1 mL buffer solution

(5M ammonium acetate and 10M ammonium hydroxide) was added to the sample aliquot. The mixture was agitated between each addition and then capped and left to develop color for thirty minutes. This procedure reduces all the Fe in solution to Fe2+

with the hydroxylamine hydrochloride solution; the Ferrozine binds with the Fe2+; and the buffer solution buffers the pH (Stookey, 1970).

Treated samples were carefully poured into a 1 cm optical glass cuvette and placed in a Shimadzu UVmini-1240 UV-Vis Spectrophotometer set to 562 nm and absorbances were measured manually. The cuvette was thoroughly rinsed with deionized water and dried between each sample. Prior to sample analysis, a seven-point calibration curve was created using standards that had been treated to develop color as described above. A mid-range standard and blank were analyzed approximately every 8 samples to account for drift in the instrument. Precision was calculated to be less than ±5%. In addition, a method blank of 2 mL 18Ω deionized water was treated with the same method and analyzed along with samples.

RESULTS

Total dissolved solids

The concentration of total dissolved solids (TDS) reflects a decrease of major ions in solution between Leach 1 and 2 (Figure 3.4). TDS in both Leach 1 and 2 follow the trend EST > Lakes > Glaciers > Landforms. A single-factor ANOVA analysis of TDS noted a significant difference between surfaces in both Leach 1 and Leach 2. However, this difference is largely due to the TDS concentration in the ESTs. For example, the

Leach 1 Fryxell EST is 1.2 times the most concentrated Leach 1 lake sample (West Lake

Bonney). However, when the ESTs are removed from the ANOVA analysis there is no longer a significant difference noted between glaciers, lakes, and landforms in either leach method. This same pattern of significance is noted in the concentrations of all

major ions. Figure 3.5 depicts the changes in TDS concentration with distance inland in the Taylor Valley with sample type and basin designations. As with sample type, there is also no significant difference noted between lake basins, though there is a difference in the degree of variation with the widest range of TDS occurring in the Bonney basin.

Major ions

Major ion data are recorded in Appendices E and F for Leach 1 and Leach 2, respectively. Piper-style variation diagrams (Figure 3.6) show the dominance of Ca2+,

Na++K+ cations for both cold water (Leach 1) and freeze-thaw (Leach 2) leachates, though there is great variation in the dominant cation. There is an increase of Mg2+ concentration in Leach 2 over Leach 1. Anion concentrations similarly vary with no readily discernable differences between sample location and type. There is a shift away

- - from Cl in Leach 1 to HCO3 in Leach 2, though this does not necessarily indicate a shift in the dominant anion at a single site.

Figures 3.7 to 3.10 show percent cation and anion contribution from glaciers,

ESTs, lakes, and landforms, respectively, for both leach methods. Samples follow the same pattern of increase or decrease between Leach 1 and 2 within a single sample type.

However, patterns of increase or decrease are not constant between sample types.

- 2- - Potassium and HCO3 decrease inland while SO4 and NO3 increase inland in all sample types. Chloride increases to the west in glaciers, lakes, and aeolian landforms, but remains relatively constant in the ESTs. Sodium increases to the west in the lake and aeolian landform leachates, but decreases in glacial and EST leachates. This may indicate that, at least in the case of the lake surface dust samples that are all located within the Taylor Valley, some leaching of the collected sediments has already occurred.

Calcium and Mg2+ both increase inland in glacial and EST leachates and decreases in lake and aeolian landform leachates. These results suggest that, soluble salts, such as

NaCl and CaCO3, would be removed first from the leached aeolian material of lakes and aeolian landforms, resulting in the major ion patterns seen in Figures 3.9 and 3.10.

To ascertain the primary source of the solutes in the leachates, an initial comparison of major ions to chloride in μeq g-1 is presented in Table 3.1 with the seawater equivalent ratios for reference. Solute ratios greater than the seawater value indicate the influence of a source other than primary marine-derived aerosols. Leachate equivalent ratios are greater in Leach 2 than Leach 1 in all but a few cases. These

2- - 2+ - exceptions include the SO4 :Cl of all the ESTs and Mummy Pond and the Mg :Cl and

Ca2+:Cl- in the West Bonney EST.

Sodium ratios less than those of the seawater ratio only occur in Leach 1. While there is no spatial pattern to these lower than seawater Na+:Cl- ratio, it does occur in all sample types. Magnesium is the only other ion ratio that shows depletion compared to the seawater ratio. Depletion occurs only within the EST leachates: in Leach 1 at the

Fryxell and Hoare ESTs and in Leach 2 only at the Fryxell EST. While these would indicate a spatial pattern within the ESTs, the lack of a pattern in any other sample types suggests no MDV-wide spatial pattern exists. However, the influence of salts can be further implied from a plot of Na+ versus Cl- with the halite dissolution line and seawater ratio for reference (Figure 3.11). The majority of the leachates lie close to the halite dissolution line except at the lowest concentrations where an increased excess of Na+ is noted, indicating another source of Na+ besides NaCl dissolution. The majority of non- seasalt Mg2+, Ca2+, and Na+ in the leachates can be explained by the dissolution of salts

2- - containing SO4 and HCO3 (Figure 3.12). Deviance from the 1:1 line of Figure 3.12 may indicate the dissolution of K- or NO3-bearing salts and/or the weathering of the silicate portion of the aeolian sediments.

- A plot of H4SiO4 versus HCO3 content of the leachates with the silicate

- weathering line for reference (Figure 3.13) indicates an excess of HCO3 and, thereby, the probability of carbonate dissolution. This is supported by a plot of Mg2++Ca2+ versus

2- - 2+ 2+ SO4 +HCO3 (Figure 3.14) indicating that the concentration of Ca and Mg in the leachates can be explained through the dissolution of sulfate and/or carbonate minerals.

Subsequent plots of calcite dissolution (Figure 3.15) and CaMg-carbonate dissolution

(Figure 3.16) indicate leachates of both methods lie close to the line of stoichiometric carbonate mineral dissolution. When just calcite dissolution is considered (Figure 3.15),

- 2+ there is excess of HCO3 at low concentration; when Mg is considered in addition to

Ca2+ in carbonate dissolution (Figure 3.16), leachate composition shifts closer to the

2+ 2- carbonate dissolution line. An additional plot of Ca versus SO4 representative of anhydrite/gypsum dissolution (Figure 3.17) indicates and excess of Ca2+ in a majority of the leachates. Based on the relationships apparent in Figure 3.6, salt dissolution is predominates in Leach 1 and carbonate/silicate dissolution proceeds to a greater extent in

Leach 2.

Nutrients and iron

Inorganic and total nutrient species concentrations are recorded in Appendices E and F for Leach 1 and 2, respectively. Initial inspection reveals that all total nitrogen

(TN) and total phosphorus (TP) values are at least one order of magnitude greater than their dissolved inorganic counterparts.

Inorganic and total species of N and P are plotted opposite each other in Figure

3.18 with the Redfield Ratio (16N:1P) of balanced phytoplankton growth plotted for reference. The inorganic species show a dispersed pattern centered within the zone of P- limitation, i.e. N:P ratios > 16. Only the eastern basins of Taylor Valley (Fryxell and

Hoare) lie within the field of N-limitation. A distinct pattern emerges when TN and TP are considered: two modes of TP concentration exist (~0.001 and ~0.018 μmol g-1) over a similar range of TN concentrations. In contrast to the inorganic N:P values, the majority of samples demonstrating N-limitation in TN:TP are from the western (Bonney) basin of

Taylor Valley and the higher elevation Wright and Victoria Valleys. The majority of samples from the Fryxell basin are P-limited, while Hoare basin samples are both P- and

N-limited. The inorganic and total N:P values with distance inland in the Taylor Valley are plotted in Figure 3.19 to better show the relationships discussed previously. A clear east-to-west trend of N-limitation to P-limitation exists in the inorganic N:P, while a wider range of concentrations and a possible reverse-trend exist within the total nutrient values. Soluble Fe2+ concentrations from Leach 1 are reported along with inorganic N and P values in Table 3.2. Aqueous Fe follows no schematic pattern of enrichment in the leachates, but occurs in approximately the same concentration as inorganic P.

DISCUSSION

Dissolution of aeolian material

Based on the previous study of the geochemistry of aeolian sediments (Chapter

2), it can be assumed that solute release via the chemical weathering of aeolian material occurs during the melt season within the MDV. The actual degree of dissolution as well

as the amount of aeolian material that is deposited is, therefore, important to the distribution of solutes and nutrients throughout the hydrologic continuum of the MDV.

Percent aqueous elements for Na, K, Ca, Mg, P, and Fe are calculated according to the equation:

While N could potentially be calculated, it was excluded from this study because all but two sites the total N in the bulk sediments were below the detection limit (Table 2.2).

The dissolution of the elements noted above is very small compared to the total mass of the aeolian material available for the major cations and anions (Table 3.3). For example,

Na demonstrates the greatest degree of dissolution regardless of leach method though this only is 0.7% of original material. In contrast, Fe demonstrates the smallest degree of dissolution by at least 2 orders of magnitude. Not surprisingly, the dissolution of major cation phases (Na+, K+, Mg2+, Ca2+) is 2-4 orders of magnitude greater than that of the

3- 2+ more readily absorbed nutrients (PO4 , Fe ). These values are significantly less than the values reported for Fe (median 1.7%), P (median 32%), Al (median 3%) and Mn (median

55%) from mineral dust reported in the Pacific Ocean (Baker and Jickells, 2006; Baker et al., 2006a). These differences in solubility between MDV aeolian material and mineral dust elsewhere may be caused by the increased atmospheric processing of dust that has traveled extensively in the stratosphere (Mahowald et al., 2005; Baker et al., 2006b), decreased particle size of transported dust in the ocean (average ~2 μm) (Baker and

Jickells, 2006; Jickells et al., 2005), and/or the dissolution ability of DI versus seawater.

Spatial patterns of the extent of dissolution of aeolian material are inconsistent within the MDV. Glacial surface samples all decrease in solubility to the west, but 83

aeolian landforms and lake surfaces have different and alternating patterns of solubility.

This inconsistency appears to be at least qualitatively related to the grain size distribution

(Table 3.4). Samples with a high percentage of silt tend to have a higher percent ion loss.

Smaller particle sizes have a greater surface area available relative to their volume. An inverse relationship between solubility and particle size has been found for iron specifically, but it stands to reason that this relationship will extend to other elements as well (Baker and Jickells, 2006). Since the smallest particle size in this study refers to all particles <63 μm in diameter and surface area to volume ratios increase with decreasing diameter, it is plausible that the variation of solute loss in these samples may be related to a higher amount of a smaller size fraction than it was possible to measure through the sieving method that was employed in this study.

ESTs generally demonstrate the greatest degree of dissolution (Table 3.3) and the greatest TDS (Figures 3.4-3.5), possibly indicating an increased content of soluble salts compared to the other sample types. Alternatively, this may be a true signature of the salt content of MDV soils not in contact with the aqueous system.

Sodium, Ca, and K show relatively greater dissolution than the other elements.

Sodium and Ca are the two major cations the most abundant salts in the MDV environment: NaCl (halite), Na2SO4 (thenardite/mirabilite), CaSO4 (anhydrite/gypsum), and CaCO3 (Table 3.5). These high concentrations indicate that dissolution of these salts is occurring within the leachates and is discussed further below. Trends in the dissolution of salts and carbonates have been discussed previously and are known to function in the

MDV ecosystem (Lyons et al., 2003; Nezat et al., 2001; Fortner et al., 2005).

Chemical weathering of aeolian material

The associations between major chemical species (Figures 3.11-3.17) provide evidence that salt (NaCl, Na2SO4, CaSO4) and carbonate dissolution are proceeding to a greater extent compared to silicate weathering in the leachates of both experimental methods based on the resemblance of trends to their respective stoichiometric dissolution/weathering relationships. The majority of Mg2+, Ca2+, and Na+ not explained

2- by the dissolution of chloride salts are accounted for by SO4 salt and carbonate dissolution as shown in Figure 3.12. As noted above, deviance from the 1:1 relationship indicates the presence of K- or NO3-bearing salts and/or silicate weathering, as

+ - referenced by the very small concentrations of K , NO3 , and H4SiO4 compared to the other major ions that are generally an order of magnitude greater (Appendices E and F).

The occurrence of Na-bearing salts decreases with distance from the coast and occurs on all rock types (Keys and Williams, 1981), which is also be reflected in the leachates from glacial and EST samples of this study. One of the most common Na-salts in the MDV is halite originally derived from seasalt (Campbell and Claridge, 1987;

Claridge and Campbell, 1977). Mirabilite (Na2SO4 · 10H2O) also is common in some areas (Black and Bowser, 1968). At higher concentrations of Na+ and Cl- the stoichiometric dissolution relationship of NaCl is strong (Figure 3.11). However, the majority of samples from Leach 1 and all of the samples from Leach 2 have a source of

Na+ ions other than seasalt aerosol (based on element:Cl- equivalent ratios, Table 3.1) and halite dissolution (Figure 3.11). Similar associations with seasalt aerosol are seen in all major ions presented in Table 3.1. Silicate, sulfate, and carbonate dissolution are possible sources of this excess. In particular, sulfate and carbonate dissolution based the

relationship of non-chloride salts of Mg2+, Ca2+, and Na+ that account for the majority of

2- - SO4 and HCO3 in these leachates (Figure 3.12).

2+ 2- As discussed above, there is excess Ca relative to the SO4 in the leachates

2- (Figure 3.17). The source of the SO4 ion in the MDV is external/aeolian and isotopically related to seawater (Keys and Williams, 1981; Bao et al., 2000). The increased Ca2+ is attributed to carbonate mineral dissolution, which has also been documented in the MDV streams (Nezat et al., 2001). It has been suggested that the

2- - equivalent ratio of SO4 to HCO3 in proglacial streams may reflect the dissolution of gypsum relative to carbonates (Fortner et al., 2005): values greater than 0.5 reflect a greater degree of sulfate mineral dissolution compared to carbonate. Ratios increase to the west in leachates of both methods (Table 3.6) and reflect the relationship of inland gypsum increase noted by Keys and Williams (1981) in the Taylor Valley.

Carbonate dissolution proceeds according to the stoichiometric relationship:

Figures 3.15 and 3.16 indicate that carbonate dissolution is the dominant process, in addition to salt dissolution, occurring in these experimental leachates. This is corroborated by the evolution of leachate composition from Na++K+- and Ca2+-rich solution in Leach 1 to one with an increased Mg2+ content in Leach 2 (Figure 3.6). These relationships also suggest that much of the Mg2+ in solution is due to the dissolution of a

Mg-bearing carbonate. However, dolomite (CaMg(CO3)2) and magnesite (MgCO3) are not commonly found in the MDV soils suggesting that CaMg-carbonate dissolution relationship (Figure 3.16) may be circumstantial. Other potential sources of Mg2+ are the weathering of volcanic rock particles rich in Mg or Mg-bearing salts (Table 3.5). Within

the Taylor Valley Mg-bearing salts are associated with mafic igneous rocks and increase with distance inland (Keys and Williams, 1981), which is reflected in the relative ion data of glacial and EST samples (Figure 3.7 and 3.8). Mg-bearing salt dissolution or silicate weathering, therefore, would only occur where these mafic rocks closer in composition to the Ferrar Dolerite and McMurdo Volcanic basanite discussed in the previous chapter

2- have been abraded and transported (Figure 2.11). The decrease of SO4 in Leach 2

(Appendices E and F) indicates that silicate weathering is likely the source of the

2+ 2— increased Mg because the major Mg-bearing salts are also SO4 bearing based on the major salts found in MDV soils (Table 3.5) (Keys and Williams, 1981), and is discussed further below.

The presence of reactive Si in the leachates indicates that silicate weathering is occurring. Silicate weathering proceeds according to the following typical stoichiometric relationship:

However, the concentration of H4SiO4 increases from Leach 1 to Leach 2 (Figure 3.13), indicating that silicate weathering is proceeding to a greater extent after initial salt content is diluted in Leach 1. In experimental work, increasing cycles of freeze-thaw lead to greater precipitation or loss from solution of SiO2 (Dietzel, 2005), but this is not what we have observed between leachates. Conversely, H4SiO4 content increased with time over a 24 hour period in a leaching experiment on sediments from Lake Chad and the Canada Glacier (Tegt, 2002). The increased amount of H4SiO4 in the leachates of

Leach 2 (Appendices E and F) implies that silicate weathering could be the likely the source of the increased Mg2+ in Leach 2. This is supported by the Goldich dissolution

series (Figure 2.14): mafic minerals have a higher weathering potential compared to more felsic minerals (Goldich, 1938).

It has been noted that greater degrees of aluminosilicate weathering proceed and calcite precipitates at higher pH (Drever, 1997). Dissolution of multiple salts without common ions also enhances solubility. For example, the solubility of calcite is enhanced when NaCl is added to the solution (Krauskopf and Bird, 2003).

Leachates from Leach 1 have higher Na+ and Cl- concentrations than Leach 2

(Appendices E and F) and also more closely follow the CaCO3 dissolution pathway compared to Leach 2 (Figure 3.15). Similarly, Leach 2 leachates shift closer to the silicate weathering relationship compared to Leach 1 (Figure 3.13), perhaps suggesting that silicate weathering in these leachates proceeds to a greater extent in the second leach method.

Upon sublimation/evaporation, it has also been noted that the surface waters of

+ 2+ 2- - Lakes Bonney and Hoare would evolve to Na -Mg -SO4 -Cl dominated waters and in

+ - 2- contrast, the surface waters of Lake Fryxell would evolve to Na -HCO3 -CO3 (Lyons et al., 1998). Increased sulfate mineral content relative to carbonate mineral dissolution to the west in Taylor Valley, valleywide NaCl dissolution in the aeolian leachates (this study), and supraglacial stream chemistry (Lyons et al., 2003; Fortner et al., 2005) support the notion that the dissolution of aeolian material deposited on glaciers (and throughout the hydrologic continuum) may play a role in the evolution of lake surface waters, as speculated by Lyons et al. (2001).

The sequence of evaporite minerals precipitated continues in a definite sequence based on the composition of the original solution (Figure 3.20) (Boggs Jr., 2006; Drever,

1997; Smith and Drever, 1976). Calcite is the first mineral to precipitate, followed by

- 2+ gypsum or sepiolite based on HCO3 relative to Ca concentration. The last salts to be precipitated are Cl- salts where a source of Cl-, such as McMurdo Sound, is present

(Smith and Drever, 1976; Hardie and Eugster, 1940). The K and Mg chlorides and sulfates are usually the most soluble salts. In the case of the MDV, the abundance of

NO3- is unusual and they are highly soluble as well (Keys and Williams, 1981; Ericksen,

1983). Upon rewetting, the most soluble salt (generally Na+ and Cl- salts) dissolve quickly, followed by gypsum, calcite, and silicates at slower rates (Drever, 1997). In the

MDV, cryoconcentration and sublimation as the surficial aquatic environments dry causes salts to precipitate, presumably along the chemical divide model of Hardie and

Eugster (1970). During the first part of the melt season there is flux of solutes and nutrients in the aqueous system that is indicative of these highly soluble salts returning to solution (Fountain et al., 1999b; Fountain et al., 2008; Howard-Williams et al., 1998).

As melt continues, the less soluble salts should dissolve followed by silicate weathering.

Whether the evolution of leachate composition from Leach 1 to Leach 2 is a function of the different leaching methods is unknown. It is possible that the increase

2+ - Mg , HCO3 , H4SiO4 seen in Leach 2 could be due fracture of particles due to the freeze- thaw, but this could not be demonstrated. Alternatively, the increase in these ion species may be due to the removal of more soluble salts in Leach 1 leaving less soluble species to be solubilized in Leach 2. This idea of solute separation is supported by the flux of nutrients and solutes seen in glacial melt, presumably from cryoconite holes, at the beginning of the melt season (Fountain et al., 1999b). To address potential process, further experiments with different leaching protocols must be carried out. Nevertheless,

the dissolution of MDV aeolian material is a significant process that proceeds throughout the melt season.

Nutrient abundance and limitation

The concentration of nutrients in both Leach 1 and 2 are of the same magnitude, though Leach 1 generally had slightly higher concentrations than Leach 2 (Appendices E and F). The ratio of inorganic to total N and P varies throughout the MDV landscape and is greater in the aeolian landforms and EST leachates compared to the glacial and lake surface leachates. The low inorganic nutrient values compared to total N and P indicate a greater abundance of organic N and P present in these samples and that it is water soluble

(Figure 3.18 and 3.19). This is understandable within the MDV ecosystem since the margins of the major terminal lakes and low gradient streams can be inhabited by algal and moss communities (McKnight et al., 1998). Similarly, glaciers worldwide can be inhabited by diverse microorganisms that are found both on the snow and ice, as well as within cryoconite holes that include auto- and heterotrophic organisms including algae, bacteria, rotifers, tardigrades, viruses, and insects where conditions permit (Foreman et al., 2007; Takeuchi et al., 2001; Säwström et al., 2002; Hodson et al., 2007; Porazinska et al., 2004).

The sources of inorganic N and P in the MDV ecosystem are different: N is supplied through aeolian input derived from the stratosphere (Vincent and Howard-

Williams, 1994), while P is derived from weathering of P-bearing minerals in bedrock

(Gudding, 2003; Filippelli and Souch, 1999; Vitousek et al., 1997; Vitousek and

Farrington, 1997; Barrett et al., 2007). The extent to which these nutrients are available is dependent on landscape age and the lithology of tills present (Filippelli and Souch,

1999; Vitousek et al., 1997; Vitousek and Farrington, 1997). Older landscapes have

- 3- higher amounts of NO3 ; younger landscapes have a greater available pool of PO4 . In

- the MDV, NO3 is deposited from the atmosphere and increases with elevation. The climate at elevation (>800 m) in the western MDV, defined inland ―Climate Zone 3‖ by

Marchant and Denton (1996), is characterized by mean annual temperatures (MAT) <-27º

C, relative humidity (RH) <45%, and rarely receives precipitation or meltwater. In these

- conditions and in the absence of melt, NO3 is able to accumulate to greater concentrations in the soil compared to the more temperate coastal Zone 1 (<1000 m,

MAT=-17º C, RH=75%) and intermediate Zone 2 (>1000 m, MAT=-25º C, RH=10-70%)

- - (Marchant and Denton, 1996). This relationship is reflected in a graph of NO3 versus Cl concentrations with both high elevation (Zone 3) and lower elevation (Zones 1 and 2) soils plotted for reference (Figure 3.21). The leachates of aeolian material are enriched in

- NO3 relative to low elevation soils and depleted relative to high elevation soils. This strongly suggests that the aeolian materials are a mixed source of valley floor, higher elevation, and possibly a third endmember that has not been identified.

Trends of inorganic nutrient limitation in the Taylor Valley aeolian matter follow those that have been proposed by Priscu (1995) for the terminal lakes (Table 3.7). All three lakes lie within coastal climate Zone 1 (Marchant and Denton, 1996); Lake Fryxell is P-limited and located in the most recently glaciated, youngest area of the valley (Figure

2.16) (Barrett et al., 2007). Thus, more inorganic P should be leached from sediments derived from these areas and is supported by the Filippeli and Vitousek models of landscape evolution. Additionally, the chemical weathering of the McMurdo Volcanics or other igneous complexes (granite, granodiorite, etc.) containing P-bearing minerals

(apatite, monazite) in the MDV would release inorganic P into the environment, as they have an increased P-content relative to the other major rock types in the MDV (Figure

2.11) (Haskell et al., 1965; Roser and Pyne, 1981).

Iron is a micronutrient important to photosynthesis and, therefore, primary production (Martin et al., 1991). Fe2+ concentrations at each sample site are within one order of magnitude of the inorganic P concentration probably indicating that Fe is not the limiting nutrient in these environments. Variations in concentration may be due to the collection of varying amounts of material consisting of a relatively higher percentage of

2+ 3- more mafic or iron-rich material. The trend of similar Fe and PO4 concentration holds

2+ 3- with the leachates except the Defile where Fe greatly exceeds PO4 (Table 3.2). The reason for this increased solubility at the Defile is unknown. In addition sediment deposited from the MDV area onto the sea ice (Atkins and Dunbar, 2009) may stimulate growth in McMurdo Sound.

Flux of solutes and nutrients to the MDV ecosystem

Based on flux data previously presented (Table 2.1), the surface area of lakes and glacier ablation zones (Appendix H) annual fluxes of solutes and nutrients were calculated for the various sample types. These data are presented for solutes and nutrients in Tables 3.8 and 3.9, respectively. Taylor Glacier is currently not included in this flux data because the ablation zone begins near the polar plateau and the ablation zone drains into several valleys (Basagic, personal communication). Aeolian landforms were also excluded from the flux calculation because their individual area-normalized sediment fluxes and surface areas are not known. More importantly, it is unclear if this

material is solubilized in place and contributes to the fluxes to the aquatic system except in unusual conditions.

The solute and nutrients flux of glaciers and lakes are most dependent on the aeolian sediment flux, which is highest at Lake Bonney. This inland increase in sediment flux is related to the foehn winds that are strongest in the Bonney basin. Thus, solute and nutrient fluxes are greatest in the Bonney basin. Increased solute flux at Lake Fryxell relative to Lake Hoare is like related to the influence of the coastal wind system

(Lancaster, 2002). Measured glacier solute and nutrient fluxes are highest in the Fryxell basin at the Commonwealth Glacier and is related to the total surface area that 5 times

- that of the next largest glacier, the Canada Glacier. However, NO3 flux is greatest on the eastern Canada Glacier that also drains into Lake Fryxell.

Aeolian nutrient flux has been discussed previously in regards to sediment composition by Barrett et al. (2007) in the lake basins of Taylor Valley (Table 3.10).

These fluxes are based on the sediment flux data from Lancaster (2002). The bioavailable/inorganic nutrient fluxes calculated here are ~6 orders of magnitude less than the potential flux of the sediments. Moreover, the leachate inorganic nutrient values are two or more orders of magnitude less than the inorganic nutrient load from streams to lake surface water (Table 3.8)

The total flux to any of the lake basins, but especially the Bonney basin, is an underestimation because not all of the potentially contributing glaciers are included in the calculations. There are a larger number of glaciers that drain into the Bonney basin compared to the Fryxell and Hoare basins. Other geomorphological constraints, such as stream gradient (Nezat et al., 2001), affect the degree of chemical weathering and thereby

the flux of nutrients and solutes into the MDV ecosystem. Furthermore, the fluxes presented here represent only the result of our experimental design as opposed to aeolian material contact throughout the summer melt season. Streamflow is austral summer air temperature-dependent and can change ten-fold in the course of a few hours (Fountain et al., 1999b; McKnight et al., 1999), thus the actual flux of solutes and nutrients is also related to temporal climate variability as manifested in streamflow and probably changes dramatically from year to year (Foreman et al., 2004). Based on the greater P- solubilization found in the 1:50 DI extracts that were allowed to react 16-24 hours of

Gudding (2003) and the 1:2 extracts allowed to react 5 minutes of this study (Table 3.9), the actual flux due to the solubilization of aeolian material is probably significantly greater than those measured here. Nevertheless, these calculations provide an initial estimate of solute and nutrient flux due to the dissolution of aeolian sediments. The calculated fluxes indicate that a small amount of solutes and nutrients were released during the course of the two leach methods. If these reflected the extent of aeolian dissolution, then the influence of dust dissolution to the aquatic systems of the Taylor

Valley is negligible. However, the greater fluxes achieved at a lower dust to DI ratio

(Gudding, 2003) and the increase in solute dissolution found in DI extracts over time

(Tegt, 2002) indicate that the actual influence of dust dissolution could be much greater than indicated here.

CONCLUSIONS

While the degree of dissolution of the aeolian sediment into the leachates described here is relatively small, the concentration of solutes and nutrients leached from

the sediments is similar in both leach methods. This is significant because chemical weathering/dissolution within the hydrologic continuum is the only way to disperse solutes, and more importantly nutrients, to the MDV ecosystem. Total nutrient content of the leachates of both methods indicate an organic nutrient source to the glacier and lakes.

Trends in inorganic (bioavailable) nutrients of Taylor Valley leachates indicate that nutrient limitation follows the pattern of limitation in the terminal lakes and their basins of the Taylor Valley. The foehn winds from the ice sheet deposit aeolian material to ice surfaces and streams beds providing new, more reactive material for leaching by meltwater. These dissolution/chemical weathering processes proceed throughout the meltseason, redistributing solutes and nutrients, and helping support the biota of the

MDV ecosystem. This transport of solutes, via wind deposition and later solubilization, to the closed-basin lakes was a heretofore described and unquantified process.

Figure 3.1 The partitioning of available water in Antarctica and how it relates to biota from Kennedy (1993).

Figure 3.2 A schematic diagram of the hydrologic continuum of the McMurdo Dry Valleys. Aeolian transport has been added to reflect the importance of wind as a redistributor of sediment and organic matter, as well as its importance in the geochemical evolution of meltwater.

Figure 3.3 Schematic vertical cross-section of a cryoconite hole adapted from Fountain et al. (2004).

1200 Leach 1 Glaciers (107 g g-1) Leach 2 Glaciers (80 g g-1) Leach 1 EST (599 g g-1) -1 1000 Leach 2 EST (250 g g ) Leach 1 Lakes (149 g g-1) Leach 2 Lakes (83 g g-1) Leach 1 Landforms (96 g g-1) Leach 2 Landforms (76 g g-1) 800

)

-1

g g

 600

TDS ( TDS

400

200

Defile

Taylor

Howard

Bull Pass

E. Canada

Hoare EST Hoare

W. Canada

Fryxell EST Fryxell Hoare Lake

Lake Fryxell Lake

Hoare Beach Hoare

Mummy Pond Mummy

Victoria Dunes Victoria

E. Lake Bonney E. Lake

W. Bonney EST W. Bonney

W. Lake Bonney W. Lake

E. Commonwealth

W. Commonwealth Location Figure 3.4 Total dissolved solids grouped by sample type. Within groups, sites are arranged east to west with increasing perpendicular distance from coast. Average values for each leach and sample type are in parentheses.

1200

1000

800

)

-1 600

g g



400

TDS (

200

0 10 20 30 40 Distance from coast (km) Figure 3.5 Taylor Valley total dissolved solids with distance inland. Sites are arranged east to west with increasing perpendicular distance from the coast. Basins are color- coded (black=Fryxell, blue=Hoare, red=Bonney). Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1; unfilled characters are from Leach 2.

0 0 100 100 10 10 90 90 20 20 80 80 30 30 70 70 40 40 60 - 60 2- 2+ + + HCO Mg Na + K 3 SO 50 50 4 50 50

10 60 60 40 40

0 70 70 30 30 80 80 20 20 90 90 10 10 100 100 0 0 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100

2+ Ca Cl- Figure 3.6 Piper-style cation and anion variation diagrams. Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

100

Glaciers

East (Coast)------West (Ice Sheet) East (Coast)------West (Ice Sheet) 100

% Cation 40

+ Na + 20 K 2+ Mg 2+ Ca 0 100

% Anions 40 Cl- NO - 3 20 2- SO 4 HCO - 3 0

Howard Taylor Howard Taylor E. CanadaW. Canada E. CanadaW. Canada E. CommonwealthW. Commonwealth E. CommonwealthW. Commonwealth

Leach 1 Leach 2

Figure 3.7 Relative percentages of major cations and anions from glacier sediments for Leach 1 and Leach 2. Sites are arranged east to west with increasing perpendicular distance from the coast. Percentages are calculated from μmol g-1.

101

Elevated Sediment Traps

East (Coast)------West (Ice Sheet) East (Coast)------West (Ice Sheet) 100

% Cations 40

+ Na + 20 K 2+ Mg 2+ Ca 0 100

% Anions 40

Cl- NO - 20 3 SO 2- 4 HCO - 3 0 Fryxell EST Hoare EST W. Bonney EST Fryxell EST Hoare EST W. Bonney EST

Leach 1 Leach 2

Figure 3.8 Relative percentages of major cations and anions from EST sediments for Leach 1 and Leach 2. Sites are arranged east to west with increasing perpendicular distance from the coast. Percentages are calculated from μmol g-1.

102

Lakes

East (Coast)------West (Ice Sheet) East (Coast)------West (Ice Sheet) 100

% Cation 40

+ Na + 20 K 2+ Mg 2+ Ca 0 100

% Anion 40

Cl- NO - 3 20 2- SO 4 HCO - 3 0

Lake Hoare Lake Hoare Lake Fryxell Mummy Pond Lake Fryxell Mummy Pond E. Lake BonneyW. Lake Bonney E. Lake BonneyW. Lake Bonney

Leach 1 Leach 2

Figure 3.9 Relative percentages of major cations and anions from lake sediments for Leach 1 and Leach 2. Sites are arranged east to west with increasing perpendicular distance from the coast. Percentages are calculated from μmol g-1.

103

Aeolian Landforms

East (Coast)------West (Ice Sheet) East (Coast)------West (Ice Sheet) 100

% Cation 40

+ 20 Na K+ Mg2+ Ca2+ 0 100

% Anions 40 Cl- NO - 3 20 2- SO 4 HCO - 3 0

Defile Defile Bull Pass Bull Pass Hoare Beach Victoria Dunes Hoare Beach Victoria Dunes

Leach 1 Leach 2

Figure 3.10 Relative percentages of major cations and anions from landform sediments for Leach 1 and Leach 2. Sites are arranged east to west with increasing perpendicular distance from the coast. Percentages are calculated from μmol g-1.

104

)

-1

mol g



(

0.1

0.01 0.01 0.1 1 10 Cl- (mol g-1) Figure 3.11 Na+:Cl+ ratios for both leach methods. Reference lines are the salt dissolution ratio (solid) and seawater ratio (dashed). Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

105

)

-1

eq g



(

- 3 1

+HCO

0.1 0.1 1 10 Mg2++Ca2++Na+-Cl- (eq g-1) 2+ 2+ + - 2- - Figure 3.12 Mg +Ca +Na -Cl versus SO4 +HCO3 for both leach methods. Reference line is the 1:1 line. Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

106

)

-1

eq g



(

HCO

0.1

0.01 0.01 0.1 1 10

-1 H4SiO4 (mol g ) - Figure 3.13 HCO3 :H4SiO4 ratios for both leach methods. Reference line is the silicate weathering line. Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

107

)

-1

mol g



(

- 3 1

+HCO

0.1 0.1 1 10 Mg2++Ca2+ (mol g-1) 2+ 2+ 2- - Figure 3.14 Mg +Ca versus SO4 +HCO3 for both leach methods. Reference line is the silicate weathering line. Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

108

)

-1

eq g



(

HCO

0.1

0.01 0.01 0.1 1 10 Ca2+ (mol g-1) 2+ - Figure 3.15 Ca :HCO3 ratios for both leach methods. Reference line is the carbonate weathering line. Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

109

)

-1

eq g



(

HCO

0.1

0.01 0.01 0.1 1 10

Ca2++Mg2+ (mol g 2+ 2+ - Figure 3.16 (Ca +Mg ):HCO3 ratios for both leach methods. Reference line is the carbonate weathering line. Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

110

)

-1

mol g



(

0.1

0.01 0.01 0.1 1 10 Ca2+ (mol g-1) 2+ 2- Figure 3.17 Ca :SO4 ratios for both leach methods. Reference lines are the anhydrite/gypsum dissolution ratio (solid) and the seawater ratio (dashed). Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

111

1 Phosphorus Limited Nitrogen Limited

) -1 0.1

mol g



(

+ NO

- 3 0.01

(a) 0.001 0.0001 0.001 0.01 0.1 1

3- -1 PO4 (mol g ) 10 Phosphorus Limited Nitrogen Limited

)

-1

mol g 0.1



TN (

0.01

(b) 0.001 0.0001 0.001 0.01 0.1 1 10

TP (mol g-1) Figure 3.18 (a) Inorganic and (b) total N vs P for both leach methods. Reference line is the Redfield Molar Ratio of 16N:1P with indications for P- and N-limited fields. Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

112

10000

1000

)

-1

mol g

 100

(

4 Phosphorus Limited

/ PO

2 10 Nitrogen Limited

+NO

NO 1

(a) 0.1 0 10 20 30 40 10000

1000

)

-1 100

mol g  Phosphorus Limited

10 Nitrogen Limited

TN / TP (

(b) 0.1 0 10 20 30 40 Distance from coast (km) Figure 3.19 (a) Inorganic and (b) total N:P ratios for both leach methods with distance inland in the Taylor Valley. Reference line is the Redfield Molar Ratio of 16N:1P with indications for P- and N-limited fields. Basins are color-coded (black=Fryxell, blue=Hoare, red=Bonney). Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST). Filled characters are from Leach 1, unfilled characters are from Leach 2.

113

Figure 3.20 Pathways of model evaporation of natural waters modified from Drever (1997). Units are equivalents.

114

100 Low elevation soils, Barrett et al. (2006)* High elevation soils, * and Poage et al. (2008)

)

-1

mol g 0.1



(

0.01

0.001

0.0001 0.0001 0.001 0.01 0.1 1 10 100

Cl- (mol g-1)

- - Figure 3.21 NO3 :Cl ratios for both leach methods. Basins/locations are color-coded (black=Fryxell, blue=Hoare, red=Bonney, green=Victoria Dunes, purple=Bull Pass. Sample types are distinguished by shape (square=glacier, circle=landform, diamond=lake, triangle=EST).2D Graph Filled 1characters are from Leach 1, unfilled characters are from Leach 2.

Y Data Y 115 5

0 10 20 30 40 50 60 70 X Data

Table 3.1 Equivalent ratios of major ions to chloride for both leach methods with seawater for reference. Major ions to Cl- ratios (μeq g-1) + + 2+ 2+ 2- Na K Mg Ca SO4 L1 L2 L1 L2 L1 L2 L1 L2 L1 L2 Landforms Hoare Beach 2.00 6.99 1.06 4.73 3.01 14.59 8.83 31.24 0.89 1.22 Defile 0.83 1.52 0.10 0.18 0.22 0.49 0.53 0.86 0.33 0.33 Victoria Dunes 3.30 7.78 0.29 1.33 1.06 3.89 0.75 2.84 0.37 0.63 Bull Pass 2.07 4.63 0.06 0.17 0.40 1.35 0.31 0.89 0.43 0.46 Elevated Sediment Traps Fryxell EST 1.55 1.83 0.08 0.10 0.08 0.16 0.15 0.22 0.46 0.30 Hoare EST 1.27 1.83 0.07 0.12 0.16 0.22 0.32 0.35 0.50 0.38 W. Bonney EST 0.72 1.02 0.05 0.08 0.31 0.29 1.07 0.79 1.13 0.89 Glaciers E. Commonwealth 1.59 4.02 0.22 0.62 0.47 2.08 1.01 3.28 0.45 1.00

6 W. Commonwealth 1.00 1.85 0.22 0.47 0.45 1.31 1.08 2.39 0.26 0.45 E. Canada 0.88 1.80 0.29 0.66 0.52 1.91 1.60 3.82 0.38 0.86 W. Canada 0.60 1.30 0.18 0.45 0.44 1.77 0.98 2.50 0.42 1.14 Howard 3.49 6.46 2.06 4.27 3.80 11.65 23.39 49.22 0.56 0.96 Taylor 0.62 1.01 0.11 0.16 0.35 1.00 1.08 1.64 0.43 0.84 Lakes Lake Fryxell 0.47 0.97 0.16 0.40 0.49 1.69 0.90 2.28 0.19 0.28 Lake Hoare 0.42 1.00 0.16 0.42 0.49 1.79 1.14 2.72 0.15 0.45 Mummy Pond 1.07 1.64 0.16 0.21 0.39 0.51 0.85 0.93 0.90 0.50 E. Lake Bonney 3.21 5.98 0.86 1.82 2.45 9.20 2.81 8.57 0.60 0.96 W. Lake Bonney 0.80 1.22 0.06 0.10 0.22 0.38 0.69 0.73 0.74 0.82 Seawater Molar Ratio 0.86 0.02 0.19 0.04 0.10 L1 = Leach 1 (cold water), L2 = Leach 2 (freeze-thaw)

116

Table 3.2 Absolute concentration of bioavailable nutrients for Leach 1. Leach 1 Bioavailable Nutrients Absolute Concentration (nmol g-1) - - 3- 2+ NO3 +NO2 PO4 Fe Aeolian Landforms Hoare Beach 6.97 0.38 0.52 Defile 62.99 0.27 11.49 Victoria Dunes 32.97 nd nd Bull Pass 74.13 0.23 0.43 Elevated Sediment Fryxell EST 157.44 4.50 nd Hoare EST 82.39 0.93 0.06 W. Bonney EST 432.19 0.11 nd Glaciers E. Commonwealth 13.20 8.59 2.69 W. Commonwealth 14.49 1.85 1.79 E. Canada 68.23 1.09 3.87 W. Canada 48.64 0.24 0.33 Howard 22.78 0.68 1.24 Taylor 84.30 0.12 nd Lakes Lake Fryxell 35.68 3.09 3.15 Lake Hoare 23.84 0.18 0.33 Mummy Pond 18.10 0.14 0.43 E. Lake Bonney 12.60 0.40 0.43 W. Lake Bonney 163.99 0.73 0.97 - - 3- 2+ nd = not detected; NO3 +NO2 and PO4 measured by skalar, Fe measured colorimetrically

117

Table 3.3 Percent dissolution of aeolian material for leachates from both experimental methods. Take notice of the different magnitudes in the units. Percent dissolution of aeolian material calculated from μmol g-1 Na K Mg Ca P Fe (% x 10-2) (% x 10-2) (% x 10-2) (% x 10-2) (% x 10-3) (% x 10-5) L1 L2 L1 L2 L1 L2 L1 L2 L1 L2 L1 Aeolian Landforms Hoare Beach 1.6 1.2 1.7 1.6 0.4 0.4 1.8 1.4 3.0 7.0 2.4 Defile 28.6 12.4 6.6 2.8 1.3 0.7 5.2 2.0 2.2 1.4 87.9 Victoria Dunes 15.8 12.9 2.8 4.4 0.5 0.6 0.5 0.7 - 2.0 0.0 Bull Pass 15.4 9.2 0.8 0.6 0.4 0.4 0.5 0.4 3.3 3.1 3.5 Elevated Sediment Traps Fryxell EST 57.5 22.8 6.0 2.5 1.6 1.0 3.0 1.5 13.3 15.3 - Hoare EST 65.2 22.8 7.6 3.0 2.4 0.8 6.9 1.8 7.6 18.4 - W. Bonney EST 73.7 28.2 8.8 3.8 9.3 2.4 47.0 9.5 0.9 - - Glaciers

118 E. Commonwealth 11.4 6.3 2.5 1.5 1.4 1.3 3.0 2.1 24.4 11.4 25.6 W. Commonwealth 10.7 4.9 3.8 2.1 2.1 1.5 5.1 2.8 5.9 6.8 17.2

E. Canada 4.3 2.3 2.5 1.6 1.0 1.0 2.9 1.8 4.3 3.5 39.0 W. Canada 2.7 1.3 1.7 1.0 1.0 0.9 1.8 1.0 1.0 0.8 3.5 Howard 2.2 1.4 2.5 1.7 0.9 0.9 5.0 3.5 2.6 3.3 11.0 Taylor 4.1 2.3 1.1 0.5 0.5 0.5 2.7 1.4 0.8 - - Lakes Lake Fryxell 8.2 3.6 4.9 2.6 2.9 2.1 6.1 3.3 13.6 13.1 30.7 Lake Hoare 1.5 0.9 1.2 0.8 0.6 0.6 2.1 1.3 1.4 1.7 3.6 Mummy Pond 12.6 7.7 3.9 2.1 2.1 1.1 5.4 2.3 0.8 1.1 4.7 E. Lake Bonney 3.4 1.9 1.6 1.0 0.5 0.6 0.8 0.8 3.5 2.8 3.4 W. Lake Bonney 39.7 12.7 5.3 1.7 3.2 1.1 9.9 2.2 5.3 1.8 8.7 Calculated from bulk values (Appendix G) and solute/nutrient concentrations (Appendices E & F)

118

Table 3.4 Grain size distribution of the samples that were experimentally leached. Relative percentage Silt Sand Gravel (<63 μm) (0.063-2 mm) (>2 mm) Aeolian Landforms Hoare Beach - 100.0 - Defile 0.2 99.4 0.4 Victoria Dunes - 100.0 - Bull Pass 0.1 99.6 0.2 Elevated Sediment Traps Fryxell EST - 100.0 - Hoare EST - 100.0 - W. Bonney EST - 100.0 - Glaciers E. Commonwealth 15.8 83.6 0.6 W. Commonwealth 7.7 91.1 1.2 E. Canada 7.5 90.1 2.4 W. Canada 2.6 96.9 0.5 Howard 7.2 88.8 4.0 Taylor 0.1 74.6 25.4 Lakes Lake Fryxell 1.3 98.6 0.1 Lake Hoare 0.0 92.9 7.1 Mummy Pond 0.2 39.8 60.0 E. Lake Bonney 1.0 97.1 1.8 W. Lake Bonney 0.9 98.8 0.3

119

Table 3.5 Major salts found in the MDV from Keys and Williams (1981) Common salts of the MDV Name Chemical Formula

thenardite Na2SO4

gypsum CaSO4·2H2O halite NaCl

calcite CaCO3

darapskite Na3NO3SO4·H2O

soda nitre NaNO3

mirabilite Na2SO4·10H2O

bloedite Na2Mg(SO4)2·4H2O

epsomite MgSO4·7H2O

hexahydrite MgSO4·6H2O Keys and Williams (1981)

120

2- - Table 3.6 Equivalent ratio of SO4 to HCO3 for both experimental leach methods. 2- - SO4 :HCO3 Equivalent Ratio L1 L2 Aeolian Landforms Hoare Beach 0.07 0.02 Defile 1.03 0.20 Victoria Dunes 0.10 0.05 Bull Pass 0.37 0.09 Elevated Sediment Traps Fryxell EST 1.33 0.31 Hoare EST 1.77 0.35 W. Bonney EST >10 4.05 Glaciers E. Commonwealth 0.25 0.13 W. Commonwealth 0.18 0.10 E. Canada 0.23 0.14 W. Canada 0.60 0.31 Howard 0.02 0.01 Taylor 0.77 0.47 Lakes Lake Fryxell 0.22 0.09 Lake Hoare 0.15 0.10 Mummy Pond 1.75 0.29 E. Lake Bonney 0.08 0.04 W. Lake Bonney >2.5 1.73

Table 3.7 Nutrient limitation found in Taylor Valley terminal lakes. Nutrient Limitation in Taylor Valley Lakes Bioassay Bioassay at chlorophyll Chemical analysis maxima (under ice cover) (2m below ice cover) (stoichiometry) Lake Fryxell P and N N N Lake Hoare P and N N N Lake Bonney P P P (Priscu, 1995)

121

Table 3.8 Fluxes of solutes to the MDV ecosystem from glaciers and lakes based on sediment fluxes and surface areas from the noted sources. -1 Estimated yearly flux of solutes from sample sites in g a + + 2+ 2+ - 2- 4+ Na K Mg Ca Cl SO4 Si Glaciers E. Commonwealth 190 46 37 110 140 110 48 W. Commonwealth 170 67 48 170 230 92 57 E. Canada 24 14 10 40 34 22 14 W. Canada 17 9 8 26 36 27 5 Howard 17 18 12 100 6 6 12 Lakes

122 Lake Fryxell 160 96 104 280 430 120 11

Lake Hoare 10 7 7 23 28 8 4 E. Lake Bonney 2900 1400 1614 2800 1200 1100 1100 W. Lake Bonney 8600 1100 1299 6000 15000 16000 350 Calculated from nutrient concentrations (Appendices E & F), sediment fluxes from Table 2.1, and landform surface area (Appendix H)

122

Table 3.9 Fluxes of inorganic nutrients to the MDV ecosystem from glaciers and lakes based on sediment fluxes and surface areas from the noted sources.

Estimated yearly flux of in organic nutrients from sample sites in g a-1 N P Fe 2+ Glaciers E. Commonwealth - 2.0 0.77 W. Commonwealth 0.34 0.64 0.51 E. Canada 1.1 0.11 0.39 W. Canada 0.72 0.02 0.03 Howard 0.46 0.11 0.16 Lakes Lake Fryxell 19 1.3 1.2 Lake Hoare 0.34 0.02 0.03 E. Lake Bonney 18 8.3 8.6 W. Lake Bonney 330 3.3 5.8 Lake Sediments - 1:50 Extract (Gudding, 2003) Lake Fryxell 22 Lake Hoare 0.32 Lake Bonney 490

Average Total Flux From Streams 1994-2002 (Foreman et al. 2004) Lake Fryxell 2.6 x 104 9.7 x 103 E. Lake Bonney 7.3 x 103 880 W. Lake Bonney 1.3 x 105 1.1 x 104 Calculated from nutrient concentrations (Appendices E & F), sediment fluxes from Table 2.1, and landform surface area (Appendix H)

123

Table 3.10 Comparison of estimated nutrient fluxes from aeolian sediments compared to the concentration of nutrients in leachates. Nutrient Flux Sediment* Leachate (kg a-1) (g a-1) N P N P Lake Fryxell 0.2 5.7 19 1.3 Lake Hoare 0.1 0.6 0.34 0.02 Lake Bonney 35.5 181.1 348 11.6 *Barrett et al. (2007)

124

CHAPTER 4

CONCLUSIONS AND FUTURE WORK

The major conclusions of this study are as follows:

 The wind regimes of the MDV are responsible for the transport and redistribution

of soils. This effect is local in scale, but potentially important due to the

availability of nutrients and solutes that are leached from the substrate into the

aqueous system during the austral summer melt season.

 The composition of aeolian material can be explained by the mixing of the four

major rock types found in the MDV: McMurdo Volcanic basanite, Ferrar

Dolerite, Beacon Sandstone, and metasedimentary Basement Complex.

 Silt and sand particle size compositions are different. Silt size particles appear to

be more mafic in composition (SiO2: 50-59% by weight) while sand size particles

appear to be more siliceous (SiO2: 59-74% by weight). When compared to the

average composition of the four rock types, this means that silt size particles are

most influenced by a mixing of mostly McMurdo Volcanic basanites and Ferrar

Dolerite and sand size particles are a mixture of mostly the Basement Complex

and Beacon Sandstone. Higher LOI values in the silt fraction could indicate a

greater amount of hydrous minerals present (gypsum, amphiboles, mica),

carbonates, and/or volatiles released from igneous rocks compared to the lower

LOI values of sand, also indicating different degrees of mixing.

125

 The more mafic composition of the silt size fraction can be attributed to the higher

potential for weathering of mafic minerals according to the Goldich dissolution

series. While the Goldich dissolution series specifically concerns the breakdown

of particles by chemical weathering, the resistance to weathering also applies to

physical weathering – in our case, wind abrasion and erosion.

 Chemical Index of Alteration (CIA) values of both particle sizes indicates that the

aeolian material is largely unweathered. CIA values also indicate that the Ferrar

Dolerite may be important in the compositional evolution of both particle sizes,

but this may be artificial as some degree of leaching of the aeolian material may

have occurred before collection, which would remove Na, K, and Ca thereby

artificially lowering CIA values.

 Salt accumulation may be a source of variation in the composition of the aeolian

material sampled for this study. This would most greatly affect Na, Ca, and Mg

based on the most prevalent salts found in the MDV, and variations in these

elements (especially EST and glacial surface samples) generally follow the trends

seen in soil salts presented by Keys and Williams (1981).

 Silts collected from the surface of the Commonwealth, Canada, Howard, and

Taylor Glaciers have 87Sr/86Sr values ranging from 0.7089927 to 0.7149360 and

εNd(0) from -8.35 to -5.31 (Taylor Glacier not included). These samples are most

similar to the soils adjacent to the glaciers indicating a transfer of local materials.

These values are not within the range of the ice core dust from EPICA Dome C

and Vostok indicating that there is no significant amount of inland transfer of this

material.

126

 Leachates of the aeolian material indicate very small degrees of dissolution (at

most 0.7%) that are much smaller than reported values for the dissolution of

aeolian dust. Dissolution of Fe and P in these samples is at most 0.09% and

0.02%, respectively, compared to dust collected over the Pacific Ocean with

dissolution values much greater reported for Fe (1.7%) and P (32%). This

increased reactivity in other studies may be due to increased atmospheric

processing or smaller particle size of Pacific dust compared to MDV aeolian

material.

 The majority of leachates of Leach 1 (cold water) and all the leachates of Leach 2

(freeze-thaw) have ion:Cl- equivalent ratios indicating other sources of major ions

besides seasalt aerosol based on seawater ratios.

 The composition of the leachates largely represents the dissolution of the major

salt phases found in the MDV soils.

 The most soluble Na- and Cl-bearing salts are the first to dissolve in Leach 1 and

based on the chemical divide model of brine evolution presented by Hardie and

Eugster (1970). This initial reactivity is what causes the higher flux of solutes

and nutrients seen in glacial meltwater at the beginning of the summer melt

season.

-  Increased HCO3 and H4SiO4 in Leach 2 indicate that carbonate dissolution and

silicate weathering may be proceeding to a greater extent in Leach 2 compared to

Leach 1. This is also supported by the increase of Mg2+ from Leach 1 to Leach 2

indicating that a mafic silicate phase is probably being weathered.

127

 Nitrate- and K-bearing salts, while present in the MDV, are not major sources of

ions in these leachates compared to the other ion species discussed.

 Comparison of total dissolved N and P to dissolved inorganic N and P indicates

the influence of organic N and P is greatest in leachates from glacial and lake

surface samples. As lakes and glaciers are known to host biotic communities, the

organic N and P trend may be a reflection of this.

 Inorganic N:P ratios indicate that leachates from the Fryxell basin are N-limited

and those from the Bonney basin are P-limited based on the Redfield Ratio.

Leachates from the Hoare basin show either P- or N-limitation. These trends are

the same as those seen within the surface waters of the lakes (Priscu, 1995).

 Soluble Fe concentrations are the same order of magnitude as inorganic P

concentrations in the leachates indicating that Fe is likely not the limiting nutrient

in these leachates.

 The greatest solute fluxes, derived from ion concentration and known sediment

fluxes at 1 m from Lancaster (2002), are Na+ (10-8600 g a-1), Ca2+ (23-6000 g a-1),

- -1 2- -1 Cl (6-15,000 g a ), and SO4 (6-16,000 g a ) . These ions are also the major

constituents of many of the salts found widely in the MDV according to Keys and

Williams (1981).

 Nutrient fluxes based on leachate concentrations range from 0.34-330 g a-1 for N,

0.02-8.3 g a-1 for P, and 0.03-8.6 g a-1for Fe. These values are still two or more

orders of magnitude less than the calculated loads from streams to the lakes.

However, based on other experimental leaching methods that continued for longer

periods of time and generated greater flux values (Gudding, 2003; Tegt, 2002),

128

the values presented in this study probably are underestimation and perhaps even

minima.

Areas for future work

This study provides an initial investigation into the geochemistry of aeolian materials and what can be solubilized from them. Further investigations into this topic should involve the discerning of the silicate minerals and salts in the sediments themselves. While this study identified the different signatures apparent in silt and sand particles, it could not account quantitatively for the influence of salts on major element chemistry. To this end scanning electron microscope analysis of the sediments would be useful in discerning the bulk influence of salt on the sediments, as well as identification of the salt minerals present. High precision X-ray diffraction analysis might also aid in this identification.

As discussed previously, it is not known whether the dissolution/weathering relationships seen between Leach 1 (salt-dominated) and Leach 2 (increased silicate weathering) is a function of the leach method (cold water versus freeze-thaw) or the order the leaching experiments were completed in. Thus, further operational investigations into the role of freeze-thaw and cold water leaching on the composition of both the aeolian sediments and their leachates are required to clarify this relationship.

A last avenue of future research concerns the isotopic signature of MDV dust.

Aside from this work, only a few other ―sands‖ have been measured from various MDV

87 86 locales (Delmonte et al., 2004) and have produced a wide range of Sr/ Sr and εNd(0) values. To quantitatively rule out the possibility of inland transfer of material and constrain the MDV dust isotopic field, a wider scale sampling of aeolian material is

129

required. Ideally this would involve only dust of similar particle size as that seen in ice cores, which, based on these and previous MDV dust studies, may be difficult. Sampling at higher elevations in the landscape as well as during high velocity wind events would aid in this understanding.

130

REFERENCES

Adams E. E., Priscu J. C., Fritsen C. H., Smith S. R. and Brackman S. L. (1998) Permanent ice covers of the McMurdo Dry Valley lakes, Antarctica: Bubble formation and metaorphism. In Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica (ed. J. C. Priscu). American Geophysical Union, pp. 281-295.

Angino E. E., Turner M. D. and Zeller E. J. (1962) Reconnaissance geology of the lower Taylor Valley, Victoria Land, Antarctica. GSA Bulletin 73, 1553-1562.

Atkins C. B. and Dunbar G. B. (2009) Aeolian sediment flux from sea ice into Southern McMurdo Sound, Antarctica. Global Planet. Change 69, 133-141.

Ayling B. F. and McGowan H. A. (2006) Niveo-eolian sediment deposits in coastal South Victoria Land, Antarctica: Indicators of regional variability in weather and climate. Arctic Antarctic and Alpine Research 38, 313-324.

Bagnold R. A. (1941) The physics of windblown sand and desert dunes. Methuen, London.

Bagshaw E. A., Tranter M., Fountain A. G. and Welch K. A. (2007) Biogeochemical evolution of cryoconite holes on Canada Glacier, Taylor Valley, Antarctica. Journal of Geophysical Research 112,

Bagshaw E. A., Tranter M., Wadham J. L., Fountain A. G. and Basagic H. (2010) Dynamic behaviour of supraglacial lakes on cold polar glaciers: Canada Glacier, McMurdo Dry Valleys, Antarctica. Journal of Glaciology 56, 366-368.

Baker A. R., Jickells T. D., Witt M. and Linge K. L. (2006a) Trends in the solubility of iron, aluminum, manganese, and phosphorus in aerosol collected over the Atlantic Ocean. Marine Chemistry 98, 43-58.

Baker A. R., French M. and Linge K. L. (2006b) Trends in aerosol nutrient solubility along a west-east transect of the Saharan dust plume. Geophysical Research Letters 33, 43-58.

Baker A. R. and Jickells T. D. (2006) Mineral particle size as a control on aerosol iron solubility. Geophysical Research Letters 33,

131

Bao H., Campbell D. A., Bockheim J. G. and Thiemens M. H. (2000) Origins of sulphate in Antarctic dry-valley soils as deduced from anomalous 17O compositions. Nature 407, 499-502.

Bar-Or R., Erlick C. and Gildor H. (2008) The role of dust in glacial-interglacial cycles. QSR 27, 201-208.

Barrett J. E., Virginia R. A., Hopkins D. W., Aislabie J., Bargagli R., Bockheim J. G., Campbell I. B., Lyons W. B., Moorhead D. L., Nkem J. N., Sletten R. S., Steltzer H., Wall D. H. and Wallenstein M. D. (2006) Terrestrial ecosystem processes of Victoria Land, Antarctica. Soil Biology and Biochemistry 38, 3019-3034.

Barrett J. E., Virginia R. A., Lyons W. B., McKnight D. M., Priscu J. C., Doran P. T., Fountain A. G., Wall D. H. and Moorhead D. L. (2007) Biogeochemical stoichiometry of Antarctic Dry Valley ecosystems. Journal of Geophysical Research-Biogeosciences 112,

Basile I., Grousset F. E., Revel M., Petit J. R., Biscaye P. E. and Barkov N. I. (1997) Patagonian origin of glacial dust deposited in East Antarctica (Vostok and Dome C) during glacial stages 2, 4 and 6. Earth and Planetary Science Letters 146, 573-589.

Black R. F. and Bowser C. J. (1968) Salts and associated phenomena of the termini of the Hobbs and Taylor Glaciers, Victoria Land, Antarctica. International Union of Geodesy and Geophysics, Commission on Snow and Ice. Publication 79226-238.

Boggs Jr. S. (2006) Principles of Sedimentology and Stratigraphy. Prentice Hall, Upper Saddle River, NJ.

Bourke M. C., Ewing R. C., Finnegan D. and McGowan H. A. (2009) Sand dune movement in the Victoria Valley, Antarctica. Geomorphology 109, 148-160.

Bremner J. M. (1996) Nitrogen-Total. In Methods of Soil Analysis. Part 3 - Chemical Methods (ed. D. L. Sparks). Soil Science Society of America, Madison, WI. pp. 1085-1121.

Bullard J. E. and Livingstone I. (2002) Interactions between aeolian and fluvial systems in dryland environments. Area 34, 8-16.

Burkins M. B., Virginia R. A., Chamberlain C. P. and Wall D. H. (2000) Origin and distribution of soil organic matter in the Taylor Valley, Antarctica. Ecology 81, 2377-2391.

Calkin P. E. and Rutford R. H. (1974) The sand dunes of Victoria Valley, Antarctica. Geographical Review 64, 189-216.

132

Campbell I. B. and Balks M. R. (1998) The soil environment of the McMurdo Dry Valleys, Antarctica. In Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica (ed. J. C. Priscu). American Geophysical Union, pp. 297-322.

Campbell I. B. and Claridge G. G. C. (1987) Antarctica: Soils, Weathering Processes and Environment.

Chadwick O. A., Derry L. A., Vitousek P. M., Huebert B. J. and Hedin L. O. (1999) Changing sources of nutrients during four million years of ecosystem development. Nature 397, 491-497.

Chinn T. J. (1993) Physical hydrology of the Dry Valley Lakes. In Physical and biogeochemical processes in Antarctic Lakes (eds. W. J. Green and E. I. Friedmann). American Geophysical Union, pp. 297-322.

Claridge G. G. C. and Campbell I. B. (1977) The salts in Antarctic soils, their distribution and relationship to soil processes. Soil Science 123, 377-384.

Conovitz P. A., McKnight D. M., MacDonald L. H., Fountain A. G. and House H. R. (1998) Hydrologic processes influencing streamflow variation in Fryxell Basin, Antarctica. In Ecosystem Dynamics in a Polar Desert, The McMurdo Dry Valleys, Antarctica (ed. J. C. Priscu). American Geophysical Union, pp. 93-108.

Courtright E. M., Wall D. H. and Virginia R. A. (2001) Determining habitat suitability for soil invertebrates in an extreme environment: the McMurdo Dry Valleys, Antarctica. Antarctic Science 13, 9-17.

Delmonte B., Andersson P. S., Hansson M., Schöberg H., Petit J. R., Basile-Doelsch I. and Maggi V. (2008) Aeolian dust in East Antarctica (EPICA-Dome C and Vostok): provenance during glacial ages over the last 800 kyr. Geophysical Research Letters 35, doi: 10.1029/2008GRL03382.

Delmonte B., Basile-Doelsch I., Petit J. -., Maggi V., Revel-Rolland M., Michard A., Jagoutz E. and Grousset F. E. (2004) Comparing the Epica and Vostok dust records during the last 220,000 years: stratigraphical correlation and provenance in glacial periods. Earth-Science Reviews 66, 63-87.

Delmonte B., Petit J. R. and Maggi V. (2002) Glacial to Holocene implications of the new 27000-year dust record from the EPICA Dome C (East Antarctica) ice core. Climate Dynamics 18, 647-660.

Delmonte B., Andersson P. S., Schoberg H., Hansson M., Petit J. R., Delmas R., Gaiero D. M., Maggi V. and Frezzotti M. (2010) Geographic provenance of aeolian dust in East Antarctica during Pleistocene glaciations: preliminary results from Talos Dome and comparison with East Antarctic and new Andean ice core data. Quaternary Science Reviews 29, 256-264. 133

Denman K. L., Brasseur G., Chidthaisong A., Ciais P., Cox P. M., Dickinson R. E., Hauglustaine C., Heinze C., Holland E. A., Jacob D., Lohmann U., Ramachandran S., Da Silva Dias P. L., Wofsy S. C. and Zhang Z. (2007) Coupling between changes in the climate system and biogeochemistry. In Climate Change 2007: The Physical Science. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (eds. S. Solomon, D. Qin, M. Manning, et al). Cambridge University Press, Cambridge, UK. pp. 501-587.

Denton G. H. and Hughes T. J. (2002) Reconstructing the Antarctic ice sheet at the last glacial maximum. Quaternary Science Reviews 21, 365-433.

Dietzel M. (2005) Impact of cyclic freezing on precipitation of silica in Me-SiO2-H2O systems and geochemical implications for cryosoils and -sediments. Chemical Geology 216, 79-88.

Doran P. T., McKay C. P., Clow G. D., Dana G. L., Fountain A. G., Nylen T. and Lyons W. B. (2002) Valley floor climate observations from the McMurdo dry valleys, Antarctica, 1986-2000. Journal of Geophysical Research 107, 4772.

Doran P. T., Mckay C. P., Fountain A. G., Nylen T., Mcknight D. M., Jaros C. and Barrett J. E. (2008) Hydrologic response to extreme warm and cold summers in the McMurdo Dry Valleys, East Antarctica. Antarctic Science 20, 499-509.

Drever J. I. (1997) The Geochemistry of Natural Waters. Prentice Hall,

Ericksen G. E. (1983) The Chilean nitrate deposits. American Scientist 71, 366-374.

Falkowski P. G., Scholes R. J., Boyle E., Canadell J., Canfield D., Elser J., Gruber N., Hibbard K., Högberg P., Linder S., Mackenzie F. T., Moore B., Pedersen T., Rosenthal Y., Seitzinger S., Smetacek V. and Steffen W. (2000) The global carbon cycle: a test of our knowledge of Earth as a system. Science 290, 291-296.

Faure G. and Jones L. M. (1973) Isotopic compositions of strontium and geologic history of the basement rocks of Wright Valley, southern Victoria Land, Antarctica. New Zealand Journal of Geology and Geophysics 17, 611-627.

Feng J. -., Zhu L. -., Zhen X. -. and Hu Z. -. (2009) Grain size effect on Sr and Nd isotopic compositions in eolian dust: implications for tracing dust provenance and Nd model age. Geochemical Journal 43, 123-131.

Filippelli G. M. and Souch C. (1999) Effects of climate and landscape development on the terrestrial phosphorus cycle. Geology 27, 171-174.

Foley K. K., Lyons W. B., Barrett J. E. and Virginia R. A. (2006) Pedogenic carbonate distribution within glacial till in Taylor Valley, Southern Victoria Land, Antarctica. In Paleoenvironmental Record and Applications of Calcretes and Palustrine 134

Carbonates: Geological Society of America Special Paper 416 (eds. A. M. Alonso- Zarza and L. H. Tanner). pp. 89-103.

Foreman C. M., Sattler B., Mikucki J. A., Porazinska D. L. and Priscu J. C. (2007) Metabolic activity and diversity of cryoconites in the Taylor Valley, Antarctica. Journal of Geophysical Research 112,

Foreman C. M., Wolf C. F. and Priscu J. C. (2004) Impact of episodic warming events on the physical, chemical and biological relationships of lakes in the McMurdo Dry Valleys, Antarctica. Aquatic Geochemistry 10, 239-268.

Formenti P., Andreae M. O., Lange L., Roberts G., Cafmeyer J., Rajta I., Maenhaut W., Holben B. N., Artaxo P. and Lelieveld J. (2001) Saharan dust in Brazil and Suriname during the Large-Scale-Biosphere-Atmosphere Experiment in Amazonia (LBA) -- Cooperative LBA Regional Experiment (CLAIRE) in March 1998. Journal of Geophysical Research 106, 14,919-14,934.

Fortner S. K., Tranter M., Fountain A. G., Lyons W. B. and Welch K. A. (2005) The geochemistry of supraglacial streams of Canada Glacier, Taylor Valley (Antarctica), and their evolution into proglacial waters. Aquatic Geochemistry 11, 391-412.

Fountain A. G., Dana G. L., Lewis K. J., Vaughn B. H. and McKnight D. M. (1998) Glaciers of the McMurdo Dry Valleys, Southern Victoria Land, Antarctica. In Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica (ed. J. C. Priscu). American Geophysical Union, Washington DC. pp. 39-65.

Fountain A. G., Lewis K. J. and Doran P. T. (1999a) Spatial climatic variation and its control on glacier equilibrium line altitude in Taylor Valley, Antarctica. Global and Planetary Change 22, 1-10.

Fountain A. G., Lyons W. B., Burkins M. B., Dana G. L., Doran P. T., Lewis K. J., McKnight D. M., Moorhead D. L., Parsons A. N., Priscu J. C., Wall D. H., Wharton R. A. and Virginia R. A. (1999b) Physical controls on the Taylor Valley ecosystem, Antarctica. BioScience 49, 961-971.

Fountain A. G., Nylen T. H., MacClune K. L. and Dana G. L. (2006) Glacier mass balances (1993-2001), Taylor Valley, McMurdo Dry Valleys, Antarctica. Journal of Glaciology 52, 451-462.

Fountain A. G., Nylen T. H., Tranter M. and Bagshaw E. A. (2008) Temporal variations in physical and chemical features of cryoconite holes on Canada Glacier, McMurdo Dry Valleys, Antarctica. Journal of Geophysical Research 113, doi: 10.1029/2007JGR000430.

135

Fountain A. G. and Tranter M. (2008) Introduction to the special section on microcosms in ice: the biogeochemistry of cryoconite holes. Journal of Geophysical Research 113, doi: 10.1029/2008JGR000698.

Fountain A. G., Tranter M., Nylen T. H., Lewis K. J. and Mueller D. R. (2004) Evolution of cryoconite holes and their contribution to meltwater runoff from glaciers in McMurdo Dry Valleys, Antarctica. Journal of Glaciology 50, 35-45.

Fountain A. G., Nylen T. H., Monaghan A., Basagic H. J. and Bromwich D. (2010) Snow in the McMurdo Dry Valleys, Antarctica. International Journal of Climatology 30, 633-642.

Fritsen C. H., Adams E. E., McKay C. P. and Priscu J. C. (1998) Permanent ice ocvers of the McMurdo Dry Valleys lakes, Antarctica: Liquid water contents. In Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarcitca (ed. J. C. Priscu). American Geophysical Union, pp. 269-280.

Gaiero D. M. (2007) Dust provenance in Antarctic ice during glacial periods: from where in southern South America? Geophysical Research Letters 34, L17707.

Gillies J. A., Nickling W. G. and Tilson M. (2009) Ventifacts and wind-abraded rock features in the Taylor Valley, Antarctica. Geomorphology 107, 149-160.

Goldich S. S. (1938) A study in rock weathering. Journal of Geology 46, 17-58.

Goldstein H. L., Reynolds R. L., Reheis M. C., Yount J. C. and Neff J. C. (2008) Compositional trends in aeolian dust along a transect across the southwestern United States. Journal of Geophysical Research 113, doi: 10.1029/2007JF000751.

Gooseff M. N., McKnight D. M., Lyons W. B. and Blum A. E. (2002) Weathering reactions and hyporheic exchange controls on stream water chemistry in a glacial meltwater stream in the McMurdo Dry Valleys. Water Resources Research 38, 1279.

Green W. J., Angle M. P. and Chave K. E. (1988) The geochemistry of Antarctic streams and their role in the evolution of four lakes of the McMurdo Dry Valleys. Geochimica et Cosmochimica Acta 52, 1265-1274.

Gudding J. A. (2003) Phosphorus in the Taylor Valley, Antarctica: the connection between landscape age and nutrient limitation in aquatic ecosystem components. 118.

Hall B. L. and Denton G. H. (2000) Radiocarbon chronology of Ross Sea Drift, eastern Taylor Valley, Antactica: evidence for a grounded ice sheet in the Ross Sea at the Last Glacial Maximum. Geografiska Annaler 82, 305-336.

136

Hall K., Thorn C. E., Matsuoka N. and Prick A. (2002) Weathering in cold regions: some thoughts and perspectives. Progress in Physical Geography 26, 577-603.

Hardie L. A. and Eugster H. P. (1940) The evolution of closed-basin brines. Mineralogical Society of America Special Publication 3273-290.

Harrison S. P., Kohfeld K., Roelandt C. and Claquin T. (2001) The role of dust in climate changes today, at the last glacial maximum and in the future. Earth-Science Reviews 54, 43-80.

Harvey J. and Baxter E. F. (2009) An improved method for TIMS high precision neodymium isotope analysis of verysmall aliquots (1–10 ng). Chemical Geology 258, 251-257.

Haskell T. R., Kennett J. P., Prebble W. M., Smith G. and Willis I. A. G. (1965) The geology of the middle and lower Taylor Valley of South Victoria Land, Antarctica. Transactions of the Royal Society of New Zealand 2, 169-186.

Hendy C. H. (2000) Late Quaternary lakes in the McMurdo Sound region of Antarctica. Geografiska Annaler 82A, 411-432.

Hesse P. P. and McTainish G. H. (1999) Last Glacial Maximum to Early Holocene wind strength in the mid-latitudes of the Southern Hemisphere from aeolian dust to the Tasman Sea. Quaternary Research 52, 343-349.

Hinkley T. K. and Matsumoto A. (2001) Atmospheric regime of dust and salt through 75,000 years of Taylor Dome ice core: refinement by measurement of major, minor, and trace metal suites. Journal of Geophysical Research 106, 18,487-18,493.

Hodson A., Anesio A. M., Ng F., Watson R., Quirk J., Irvine-Flynn T., Dye A., Clark C., McCloy P., Kohler J. and Sattler B. (2007) A glacier respires: quantifying the distribution and respiration CO2 flux of cryoconite across an entire Arctic supraglacial ecosystem. Journal of Geophysical Research 112, doi: 10.1029/2007JG000452.

Howard-Williams C., Schwarz A., Hawes I. and Priscu J. C. (1998) Optical properties of the McMurdo Dry Valley lakes, Antarctica. In Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica (ed. J. C. Priscu). American Geophysical Union, pp. 189-205.

Jepsen S. M., Adams E. E. and Priscu J. C. (2010) Sediment Melt-Migration Dynamics in Perennial Antarctic Lake Ice. Arctic Antarctic and Alpine Research 42, 57-66.

Jickells T. D., An Z. S., Andersen K. K., Baker A. R., Bergametti G., Brooks N., Cao J. J., Boyd P. W., Duce R. A., Hunter K. A., Kawahata H., Kubilay N., laRoche J., Liss P. S., Mahowald N. M., Prospero J. M., Ridgwell A. J., Tegen I. and Torres R. 137

(2005) Global iron connections between desert dust, ocean biogeochemistry, and climate. Science 308, 67-71.

Kennedy A. D. (1993) Water as a limiting factor in the Antarctica terrestrial environment: a biogeographical synthesis. Arctic and Alpine Research 25, 308-315.

Keys J. R. and Williams K. (1981) Origin of crystalline, cold desert salts in the McMurdo region, Antarctica. Geochimica et Cosmochimica Acta 45, 2299-2309.

Krauskopf K. B. and Bird D. K. (2003) Introduction to Geochemistry. McGraw Hill,

Lancaster N. (2002) Flux of eolian sediment in the McMurdo Dry Valleys, Antarctica: a preliminary assessment. Arctic, Antarctic, and Alpine Research 34, 318-323.

Lancaster N. (1985) Variations in wind velocity and sand transport on the windward flanks of desert sand dunes. Sedimentology 32, 581-593.

Lancaster N., Nickling W. G. and Gillies J. A. (2010) Sand transport by wind on complex surfaces: Field studies in the McMurdo Dry Valleys, Antarctica. Journal of Geophysical Research 115, F03027.

Lancaster N. (2004) Relations between aerodynamic and surface roughness in a hyper- arid cold desert: Mcmurdo dry valleys, Antarctica. Earth Surface Processes and Landforms 29, 853-867.

Li F., Ginoux P. and Ramaswamy V. (2008) Distribution, transport, and deposition of mineral dust in the Southern Ocean and Antarctica: Contribution of major sources. Journal of Geophysical Research 113, D10207.

Lyons W. B., Fountain A. G., Doran P., Priscu J. C., Neumann K. and Welch K. A. (2000) Importance of landscape position and legacy: the evolution of the lakes in Taylor Valley, Antarctica. Freshwater Biology 43, 355-367.

Lyons W. B., Tyler S. W., Wharton R. A., McKnight D. M. and Vaughn B. H. (1998) A late Holocene desiccation of Lake Hoare and Lake Fryxell, McMurdo Dry Valleys, Antarctica. Antarctic Science 10, 247-256.

Lyons W. B. and Welch K. A. (1997) Lithium in waters of a polar desert. Geochimica et Cosmochimica Acta 61, 4309-4319.

Lyons W. B., Welch K. A., Fountain A. G., Dana G. L., Vaughn B. H. and McKnight D. M. (2003) Surface glaciochemistry of Taylor Valley, southern Victoria Land, Antarctica and its relationship to stream chemistry. Hydrological Processes 17, 115-130.

138

Lyons W. B., Welch K. A., Gardner C. B., Jaros C., Moorhead D. L., Knoepfle J. L. and Doran P. T. (2010) The geochemistry of Upland Ponds, Taylor Valley, Antarctica. Antarctic Science in review.

Lyons W. B., Welch K. A., Neumann K., Toxey J. K., McArthur R., Williams C., McKnight D. M. and Moorhead D. L. (1998) Geochemical linkages among glaciers, streams and lakes within the Taylor Valley, Antarctica. In Ecosystem dynamics in a polar desert: the McMurdo Dry Valleys, Antarctica (ed. J.C. Priscu ). American Geophysical Union, pp. 77-92.

Lyons W. B., Welch K. A., Priscu J. C., Labourn-Parry J., Moorhead D., McKnight D. M., Doran P. T. and Tranter M. (2001) The McMurdo Dry Valleys Long-Term Ecological Research Program: New understanding of the biogeochemistry of the Dry Valley lakes: a review. Polar Geography 25, 202-217.

Mahowald N. M., Baker A. R., Bergametti G., Brooks N., Duce R. A., Jickells T. D., Kubilay N., Prospero J. M. and Tegen I. (2005) Atmospheric global dust cycle and iron inputs to the ocean. Global Biogeochemical Cycles 19, doi: 10.1029/2004GN002402.

Mahowald N. M., Kohfeld K., Hansson M., Balkanski Y., Harrison S. P., Prentice I. C., Schulz M. and Rodhe H. (1999) Dust sources and deposition during the last glacial maximum and current climate: a comparison of model results with paleodata from ice cores and marine sediments. Journal of Geophysical Research 104, 15,895- 15,916.

Malin M. C. (1985) Rates of geomorphic modification in ice-free areas southern Victoria Land, Antarctica. Antarctic Journal of the United States Review 20, 18-21.

Malin M. C. (1984) Abrasion rate observations in Victoria Valley, Antarctica: 340-day experiment. Antarctic Journal of the United States Review 19, 14-16.

Marchant D. R. and Denton G. H. (1996) Miocene and Pliocene paleoclimate of the Dry Valleys region, Southern Victoria land: A geomorphological approach. Mar. Micropaleontology 27, 253-271.

Martin J. H. (1990) Glacial-interglacial CO2 change: the iron hypothesis. Paleoceanography 5, 1-13.

Martin J. H., Gordon R. M. and Fitzwater S. E. (1991) Iron limitation? Limnol. Oceanogr. 36, 1793-1802.

Maurice P. A., McKnight D. M., Leff L., Fulghum J. E. and Gooseff M. (2002) Direct observations of aluminosilicate weathering in the hyporheic zone of an Antarctic Dry Valley stream. Geochimica et Cosmochimica Acta 66, 1335-1347.

139

McIntyre N. F. (1984) Cryoconite hole thermodynamics. Canadian Journal of Earth Science 21, 152-156.

McKenna Neuman C. (2004) Effects of temperature and humidity upon the transport of sedimentary particles by wind. Sedimentology 51, 1-17.

McKenna Neuman C. (1993) A review of aeolian transport process in cold environments. Progress in Physical Geography 17, 137-155.

McKenna Neuman C. (1990) Role of sublimation in particle supply for aeolian transport in cold environments. Geografiska Annaler 17A, 329-335.

McKnight D. M., Alger A. S., Tate C. M., Shupe G. and Spaulding S. (1998) Longitudinal patterns in algal abundance and species distribution in meltwater streams, Taylor Valley, Southern Victoria Land, Antarctica. In Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica (ed. J. C. Priscu). American Geophysical Union, pp. 109-127.

McKnight D. M., Niyogi D. K., Alger A. S., Romblies A., Conovitz P. A. and Tate C. M. (1999) Dry Valley streams in Antarctica: Ecosystems waiting for water. BioScience 49, 985-995.

Miller R. L. and Tegen I. (1998) Climate response to soil dust aerosols. Journal of Climatology 11, 3247-3267.

Nedell S. S., Andersen D. W., Squyres S. W. and Love F. G. (1987) Sedimentation in Ice-Covered Lake Hoare, Antarctica. Sedimentology 34, 1093-1106.

Nelson D. W. and Sommers L. E. (1996) Total carbon, organic carbon, and organic matter. In Methods of Soil Analysis. Part 3 - Chemical Methods (ed. D. L. Sparks). Soil Science Society of America, Madison, WI. pp. 961-1010.

Nesbitt H. W. and Young G. M. (1982) Early Proterozoic climates and plate motions inferred from major element chemistry of lutites. Nature 299, 715-717.

Nesse W. D. (2000) Introduction to Mineralogy. Oxford University Press, New York, NY.

Neuer S., Torres-Padrón M. E., Gelado-Caballero M. D., Rueda M. J., Hernández-Brito J., Davenport R. and Wefer G. (2004) Dust deposition pulses to the eastern subtropical North Atlantic gyre: does ocean's biogeochemistry reply? Global Biogeochemical Cycles 18, doi: 10.1029/2004GB002228.

Nezat C. and Lyons W. B. (2000) Selected stream sediment XRF data.

140

Nezat C. A., Lyons W. B. and Welch K. A. (2001) Chemical weathering in streams of a polar desert (Taylor Valley, Antarctica). GSA Bulletin 113, 1401-1408.

Nylen T. H. and Fountain A. G. (2004) Climatology of katabatic winds in the McMurdo dry valleys, southern Victoria Land, Antarctica. Journal of Geophysical Research 109, D03114.

Peters D. P. C., Groffman P. M., Nadelhoffer K. J., Grimm N. B., Collins S. L., Michener W. K. and Huston M. A. (2008) Living in an increasingly connected world: a framework for continental-scale environmental science. Fronters in Ecology and the Environment. 6, doi:10.1890/070098.

Poage M. A., Barrett J. E., Virginia R. A. and Wall D. H. (2008) The influence of soil geochemistry on nematode distribution, McMurdo Dry Valleys, Antarctica. Arctic, Antarctic, and Alpine Research 40, 119-128.

Porazinska D. L., Fountain A. G., Nylen T. H., Tranter M., Virginia R. A. and Wall D. H. (2004) The biodiversity and biogeochemistry of cryoconite holes from McMurdo Dry Valleyglaciers, Antarctica. Arctic, Antarctic, and Alpine Research 36, 84-91.

Priscu J. C. (1995) Phytoplankton nutrient deficiency in lakes of the McMurdo Dry Valleys, Antarctica. Freshwater Biology 34, 227.

Priscu J. C., Fritsen C. H., Adams E. E., Giovannoni S. J., Paerl H. W., McKay C. P., Doran P. T., Gordon D. A., Lanoil B. D. and Pinckney J. L. (1998) Perennial Antarctic lake ice: An oasis for life in a polar desert. Science 280, 2095-2098.

Reheis M. C., Budahn J. R., Lamothe P. J. and Reynolds R. L. (2009) Compositions of modern dust and surface sediments in the Desert Southwest, United States. Journal of Geophysical Research 114, doi: 10.1029/2008JF001009.

Revel-Rolland M., De Deckker P., Delmonte B., Hesse P. P., Magee J. W., Basile- Doelsch I., Grousset F. and Bosch D. (2006) Eastern Australia: A possible source of dust in East Antarctica interglacial ice. Earth and Planetary Science Letters 249, 1- 13.

Ridgwell A. J. (2002) Dust in the Earth system: the biogeochemical linking of land, air and sea. Philosophical Transactions of the Royal Society of London A 360, 2905- 2924.

Ridgwell A. J., Maslin M. A. and Watson A. J. (2002) Reduced effectiveness of terrestrial carbon sequestration due to an antagonistic response of ocean productivity. Geophysical Research Letters 29, 1095.

Ridgwell A. J. and Watson A. J. (2002) Feedback between aeolian dust, climate and atmospheric CO2 in glacial time. Paleoceanography 17, 1059. 141

Roser B. P. and Pyne A. R. (1981) Wholerock geochemistry. In Antarctic Cenozoic history from the CIROS-1 drillhole, McMurdo Sound (ed. P. J. Barrett). pp. 175- 184.

Šabacká M., Priscu J. C., Basagic H., Fountain A. G. and Greenwood M. C. (2010) Aeolian transport in the McMurdo Dry Valleys, Antarctica.

Säwström C., Mumford P., Marshall W., Hodson A. and Laybourn-Parry J. (2002) The microbial communities and primary productivity of cryoconite holes in an Arctic glacier (Svalbard 79oN). Polar Biology 25, 591-596.

Schlesinger W. H. (1997) Biogeochemistry: an Analysis of Global Change. Academic Press, San Diego, CA.

Selby M. J. (1971) Slopes and their development in an ice-free, arid area of Antarctica. Geografiska Annaler 53, 235-245.

Selby M. J., Rains R. B. and Palmer R. W. (1974) Eolian deposits of the ice-free Victoria Valley, Southern Victoria Land, Antarctica. New Zealand Journal of Geology and Geophysics 17, 543-562.

Selby M. J. and Wilson A. T. (1971) The origin of the Labyrinth, Wright Valley, Antarctica. GSA Bulletin 82, 471-476.

Sepälä M. (2004) Wind as a Geomorphic Agent in Cold Environments. Cambridge University, Cambridge.

Shinn E. A., Smith G. Q., Prospero J. M., Betzer P., Hayes M. L., Garrison V. and Barber R. T. (2000) African dust and the demise of Caribbean coral reefs. Geophysical Research Letters 27, 3,029-3,032.

Simpson A. L., Read S., Cottle J. and Cooper A. (2002) Adakitic magma generation in southern Victoria Land, Antarctica. Eos, Transactions, American Geophysical Union (abstr.).

Sleewaegen S., Lorrain R., Offer Z., Azmon E., Fitzsimons S. and Souchez R. (2002) Trapping of aeolian sediments and build-up of the ice cover of a dry-based Antarctic lake. Earth Surface Processes and Landforms 27, 307-315.

Smith C. L. and Drever J. I. (1976) Controls on the chemistry of springs at Teels Marsh, Mineral County, Nevada. Geochimica et Cosmochimica Acta 40, 1081-1093.

Speirs J. C., McGowan H. A. and Neil D. T. (2008a) Meteorological controls on sand transport and dune morphology in a polar-desert: Victoria Valley, Antarctica. Earth Surface Processes and Landforms 33, 1875-1891.

142

Speirs J. C., McGowan H. A. and Neil D. T. (2008b) Polar Eolian Sand Transport: Grain Characteristics Determined by an Automated Scanning Electron Microscope (QEMSCAN (R)). Arctic Antarctic and Alpine Research 40, 731-743.

Spigel R. H. and Priscu J. C. (1998) Physical limnology of the McMurdo Dry Valleys lakes. In Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica (ed. J. C. Priscu). American Geophysical Union, pp. 153-187.

Squyres S. W., Andersen D. W., Nedell S. S. and Wharton R. A. (1991) Lake Hoare, Antarctica - Sedimentation through a Thick Perennial Ice Cover. Sedimentology 38, 363-379.

Stookey L. L. (1970) Ferrozine- A new spectrophotometric reagent for iron. Analytical Chemistry 42, 779-781.

Stuiver M., Denton G. H., Hughes T. J. and Fastook J. L. (1981) History of the marine ice sheet in West Antarctica during the last glaciation: a working hypothesis. In The Last Great Ice Sheets (eds. G. H. Denton and T. H. Hughes). Wiley-Interscience, New York, NY. pp. 319-436.

Takeuchi N., Kohshima S. and Seko K. (2001) Structure, formation, and darkening process of albedo-reducing material (cryoconite) on a Himalayan glacier: a granular algal mat growing on a glacier. Arctic, Antarctic, and Alpine Research 33, 115-122.

Tegen I. (2003) Modeling the mineral dust aerosol cycle in the climate system. Quaternary Science Reviews 22, 1834.

Tegt S. K. (2002) The chemical evolution of Canada Glacier melt: supraglacial and proglacial waters in Taylor Valley, Antarctica.

Tranter M., Fountain A. G., Fritsen C. H., Lyons W. B., Priscu J. C., Statham P. J. and Welch K. A. (2004) Extreme hydrochemical conditions in natural microcosms entombed within Antarctic ice. Hydrological Processes 18, 379-387.

Vincent W. F. and Howard-Williams C. (1994) Nitrate-rich inland waters of the Ross Ice Shelf region, Antarctica. Antarctic Science 6, 339-346.

Vitousek P. M., Chadwick O. A., Crews T. E., Fownes J. H., Hendricks D. M. and Herbert D. (1997) Soil and ecosystem development across the Hawaiian Islands. GSA Today 7, 1-8.

Vitousek P. M. and Farrington H. (1997) Nutrient limitation and soil development: Experimental test of a biogeochemical theory. Biogeochemistry 37, 63-75.

143

Watson A. J., Bakker D. C. E., Ridgwell A. J., Boyd P. W. and Law C. S. (2000) Effect of iron supply of Southern Ocean CO2 uptake and implications for glacial atmospheric CO2. Nature 407, 730-733.

Welch K. A., Lyons W. B., Graham E., Neumann J., Thomas J. M. and Mikesell D. (1996) Determination of major element chemistry in terrestrial waters from Antarctica by ion chromatography. Journal of Chromatography A 739, 256-263.

Wharton R. A., McKay C. P., Simmons G. M. and Parker B. C. (1985) Cryoconite holes on glaciers. BioScience 35, 499-503.

Witherow R. A., Lyons W. B., Bertler N. A. N., Welch K. A., Mayewski P. A., Sneed S. B., Nylen T., Handley M. J. and Fountain A. G. (2006) The aeolian flux of calcium, chloride and nitrate to the McMurdo Dry Valleys landscape: evidence from snow pit analysis. Antarctic Science 18, 497-505.

144

APPENDIX A

Sample site location, description, and particle distribution.

145

ID Location Type Basin Latitude Longitude Silt Sand Gravel (g) (g) (g) BBP W. Bonney EST EST B transect 0.00 126.82 0.00 FBP Fryxell EST EST F transect 0.00 120.27 0.00 HBP Hoare EST EST H transect 0.00 586.47 0.00 CAN1 Canada Glacier glacier F 77.62031 162.97345 3.06 175.83 2.22 CAN2 Canada Glacier glacier F 77.62732 162.97770 0.39 96.12 0.00 CAN3 Canada Glacier glacier F 77.60785 162.97917 2.92 35.25 0.93 CAN4 Canada Glacier glacier F 77.63219 162.99283 0.67 403.49 18.43 CAN5 Canada Glacier glacier H 77.62274 162.95866 2.12 224.00 1.82 CAN6 Canada Glacier glacier H 77.60905 162.93266 1.07 39.53 0.21 CAN7 Canada Glacier glacier H 77.60830 162.96166 1.20 41.39 0.00 COM1 Commonwealth Glacier glacier F 77.56452 163.30115 22.70 120.07 0.84

146 COM2 Commonwealth Glacier glacier F 77.56918 163.30712 26.94 351.33 1.29

COM3 Commonwealth Glacier glacier F 77.57289 163.28691 5.50 51.45 0.47 COM4 Commonwealth Glacier glacier F 77.57385 163.28846 15.91 188.69 2.46 HOW1 Howard Glacier glacier H 77.66400 163.07290 15.37 342.88 1.37 HOW2 Howard Glacier glacier H 77.66287 163.08154 10.57 130.43 5.83 HOW3 Howard Glacier glacier H 77.66689 163.08262 9.29 157.85 14.13 TAY1 Taylor Glacier glacier B 77.73787 162.13770 2.12 431.23 195.34 TAY2 Taylor Glacier glacier B 77.73403 162.17245 0.08 113.69 38.67 ELB1 E. Lake Bonney lake B 77.71538 162.44565 0.57 116.51 1.12 ELB2 E. Lake Bonney lake B 77.71402 162.45681 0.77 369.61 11.42 ELB3 E. Lake Bonney lake B 77.70990 162.45023 0.38 171.27 3.08 ELB4 E. Lake Bonney lake B 77.71339 162.42431 0.31 95.96 10.07 Basins: B=Bonney, BP=Bull Pass, F=Fryxell, H=Hoare, VD=Victoria Dunes, WV=Wright Valley

146

ID Location Type Basin Latitude Longitude Silt Sand Gravel (g) (g) (g) ELB5 E. Lake Bonney lake B 77.71594 162.41496 1.85 171.97 3.25 ELB6 E. Lake Bonney lake B 77.70632 162.46627 0.22 116.78 5.91 ELB7 E. Lake Bonney lake B 77.70610 162.49644 1.71 170.88 4.41 LF1 Lake Fryxell lake F 77.60733 162.12537 1.13 84.30 0.00 LF2 Lake Fryxell lake F 77.60682 163.13272 3.08 115.17 1.30 LF3 Lake Fryxell lake F 77.60199 163.23792 0.55 72.45 0.36 LF4 Lake Fryxell lake F 77.60395 163.22008 3.72 187.55 10.58 LF5 Lake Fryxell lake F 77.60702 163.16595 1.01 75.14 0.07 LF6 Lake Fryxell lake F 77.61527 163.14002 0.87 147.86 0.00 LF7 Lake Fryxell lake F 77.61982 163.09758 0.21 119.41 0.35 LH1 Lake Hoare lake H 77.62605 162.90339 0.05 95.37 1.19

14 LH2 Lake Hoare lake H 77.62956 162.90370 0.03 86.54 6.57

7 LH3 Lake Hoare lake H 77.63123 162.92107 0.30 257.81 10.93

LH4 Lake Hoare lake H 77.63559 162.85056 0.44 818.04 39.47 LH5 Lake Hoare lake H 77.64222 162.77291 0.25 111.18 1.72 LH6 Lake Hoare lake H 77.64398 162.74802 0.15 826.15 11.48 LH7 Lake Hoare lake H 77.63186 162.93888 0.03 160.48 2.46 MP Mummy Pond lake H 77.66364 162.65248 0.96 216.32 325.71 WLB1 W. Lake Bonney lake B 77.71788 162.31328 1.37 145.05 0.40 WLB2 W. Lake Bonney lake B 77.72179 162.27722 0.85 392.32 46.95 WLB3 W. Lake Bonney lake B 77.72408 162.30207 5.21 203.26 2.95 BP1 Bull Pass landform BP 77.43397 161.66226 1.23 287.55 88.54 BP2 Bull Pass landform BP 77.42597 161.67330 0.66 463.34 1.11 Basins: B=Bonney, BP=Bull Pass, F=Fryxell, H=Hoare, VD=Victoria Dunes, WV=Wright Valley

147

ID Location Type Basin Latitude Longitude Silt Sand Gravel (g) (g) (g) BP3 Bull Pass landform BP 77.44794 161.64354 1.93 404.89 102.52 DEF1 Defile landform H 77.64972 162.72581 1.07 609.02 6.08 DEF2 Defile landform H 77.64900 162.73019 0.52 565.11 0.56 DEF3 Defile landform H 77.64873 162.73241 1.09 435.51 1.61 HB1 Hoare Beach landform H 77.62391 162.91190 1.43 1271.47 15.32 HB2 Hoare Beach landform H 77.62399 162.91200 1.13 1144.09 1.32 HB3 Hoare Beach landform H 77.62398 162.91240 0.20 851.37 0.00 NB Nussbaum landform H 77.68691 162.74429 6.86 388.35 8.14 VD1 Victoria Dunes landform V 77.36923 162.22407 0.31 630.02 0.00 VD2 Victoria Dunes landform V 77.36937 162.22525 0.06 726.76 0.00 VD3 Victoria Dunes landform V 77.37073 162.21883 0.60 845.59 0.00

148 VD4 Victoria Dunes landform V 77.37027 162.21062 0.51 879.21 0.00

VD5 Victoria Dunes landform V 77.37030 162.21032 0.21 825.37 0.00 WV1 Wright Valley landform WV 77.43849 162.68156 1.05 496.42 9.90 WV2 Wright Valley landform WV 77.43018 162.70019 0.72 470.70 45.91 WV3 Wright Valley landform WV 77.43041 162.64040 0.77 480.89 0.00 WV4 Wright Valley landform WV 77.43151 162.64922 0.67 404.88 3.27 Basins: B=Bonney, BP=Bull Pass, F=Fryxell, H=Hoare, VD=Victoria Dunes, WV=Wright Valley

148

APPENDIX B

Sand XRF and LOI data in weight percent major elemental oxide.

149

Sample ID Type Basin LOI SiO2 TiO2 Al2O3 Fe 2O3 MnO MgO CaO Na2O K2O P2O5 (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) BBP EST B 0.75 63.62 0.84 13.98 5.54 0.10 4.44 5.67 3.43 2.46 0.24 FBP EST F 0.87 66.21 0.40 12.78 5.94 0.11 6.20 5.95 2.78 2.11 0.09 HBP EST H 0.68 63.91 0.43 12.86 5.84 0.11 5.88 5.61 2.68 2.33 0.09 CAN1 glacier F 0.96 63.04 0.69 13.60 6.33 0.11 5.14 6.61 2.51 2.11 0.13 CAN2 glacier F 0.67 65.01 0.58 13.00 6.48 0.12 5.80 6.56 2.50 2.14 0.10 CAN3 glacier F 1.93 62.07 0.79 15.50 6.10 0.10 4.28 6.49 2.72 2.29 0.16 CAN4 glacier F 0.77 65.82 0.52 12.55 6.01 0.11 5.97 6.18 2.72 2.14 0.11 CAN5 glacier H 1.01 62.98 0.72 13.72 6.72 0.12 5.35 6.72 2.46 2.11 0.13 CAN6 glacier H 1.23 62.69 0.78 16.34 5.61 0.09 3.91 6.81 3.16 2.18 0.16 CAN7 glacier H 1.78 62.63 0.79 14.70 6.42 0.11 4.65 7.40 2.64 2.26 0.16 COM1 glacier F 2.49 61.41 1.08 15.04 6.76 0.11 4.11 5.79 2.78 2.70 0.23

150 COM2 glacier F 2.04 61.31 0.96 14.69 6.16 0.10 4.06 5.83 2.87 2.61 0.21

COM3 glacier F 2.67 62.43 0.98 14.84 6.20 0.10 4.01 5.56 2.86 2.68 0.22 COM4 glacier F 1.74 62.26 0.98 14.74 6.25 0.11 4.08 5.70 2.87 2.67 0.21 HOW1 glacier H 2.27 58.65 0.86 12.55 7.03 0.13 5.22 8.10 2.53 1.96 0.17 HOW2 glacier H 1.70 62.92 0.86 13.84 6.76 0.13 4.84 7.12 2.70 2.18 0.18 HOW3 glacier H 2.54 61.17 0.85 13.66 6.50 0.12 4.74 7.77 2.67 2.15 0.18 TAY1 glacier B 0.44 64.95 0.55 10.77 7.90 0.14 7.50 6.56 1.77 1.65 0.08 TAY2 glacier B 0.57 63.29 0.61 12.45 7.44 0.13 6.72 5.86 2.41 2.58 0.11 ELB1 lake B 0.45 65.54 0.45 11.98 6.77 0.13 7.17 6.49 2.30 1.96 0.08 ELB2 lake B 0.58 63.42 0.46 11.21 7.16 0.13 7.74 6.62 2.13 1.81 0.07 ELB3 lake B 0.47 67.75 0.44 12.83 5.55 0.10 4.59 5.56 2.64 2.35 0.10 ELB4 lake B 0.55 65.91 0.46 13.45 5.48 0.10 4.43 5.61 2.79 2.41 0.09 ELB5 lake B 0.54 65.49 0.54 11.47 7.17 0.13 6.35 6.46 2.03 1.79 0.08 ELB6 lake B 0.61 68.25 0.41 14.18 4.63 0.08 3.45 5.05 3.16 2.72 0.10 Basins: B=Bonney, BP=Bull Pass, F=Fryxell, H=Hoare, VD=Victoria Dunes, WV=Wright Valley

150

Sample ID Type Basin LOI SiO2 TiO2 Al2O3 Fe 2O3 MnO MgO CaO Na2O K2O P2O5 (%) (%) (%) (%) (%) (%) (%) (%) (%) ELB7 lake B 0.45 64.43 0.48 12.40 6.30 0.12 6.17 6.27 2.38 2.06 0.09 LF1 lake F 4.15 59.07 0.93 13.20 6.39 0.12 5.13 7.79 2.67 2.16 0.23 LF2 lake F 1.65 59.27 1.00 13.32 6.92 0.12 5.45 6.41 2.59 2.10 0.21 LF2DUP lake F 1.77 61.49 1.04 13.82 7.14 0.12 5.59 6.56 2.63 2.15 0.21 LF3 lake F 2.84 63.45 0.81 13.08 6.26 0.12 5.09 6.36 2.61 2.21 0.19 LF4 lake F 3.37 61.22 0.85 12.92 6.35 0.12 5.19 6.64 2.58 2.17 0.20 LF5 lake F 3.64 62.42 0.72 13.19 5.93 0.11 4.84 5.85 2.54 2.17 0.16 LF6 lake F 1.11 66.46 0.62 12.79 6.42 0.12 5.92 6.43 2.51 2.08 0.12 LF7 lake F 0.68 67.08 0.52 12.66 5.60 0.10 5.33 6.07 2.63 2.19 0.11 LH1 lake H 0.51 68.04 0.40 13.44 5.02 0.09 4.81 5.86 2.89 2.38 0.09 LH2 lake H 0.94 65.65 0.38 12.33 5.70 0.11 6.04 6.29 2.54 2.10 0.09 LH3 lake H 0.54 66.76 0.44 12.96 5.60 0.10 5.44 5.76 2.67 2.18 0.09

15 LH4 lake H 0.49 67.56 0.40 13.15 5.32 0.10 5.38 5.38 2.96 2.30 0.09

1 LH5 lake H 0.42 65.58 0.42 11.58 6.46 0.12 6.95 6.01 2.45 1.92 0.08

LH6 lake H 0.56 65.39 0.40 13.39 5.16 0.09 5.24 5.25 3.11 2.32 0.10 LH7 lake H 0.51 66.04 0.41 11.99 6.02 0.11 6.37 5.93 2.50 2.04 0.08 MP lake H 0.82 65.52 0.50 14.97 5.24 0.09 4.81 5.65 3.38 2.48 0.12 WLB1 lake B 0.86 67.53 0.56 12.07 6.40 0.12 4.52 6.24 2.03 1.77 0.10 WLB2 lake B 0.58 66.30 0.48 11.21 7.16 0.13 6.00 6.29 2.15 1.93 0.09 WLB3 lake B 0.37 67.03 0.45 12.23 6.12 0.11 5.49 6.03 2.42 2.27 0.09 BP1 landform BP 0.90 65.86 0.54 9.17 7.36 0.14 6.36 6.85 1.56 1.21 0.06 BP2 landform BP 0.17 68.62 0.35 9.03 7.02 0.14 7.19 6.47 1.66 1.42 0.05 BP3 landform BP 0.72 67.04 0.57 8.89 8.33 0.16 7.61 7.49 1.35 1.07 0.06 DEF1 landform H 1.21 62.84 0.58 11.19 7.46 0.14 7.40 7.68 2.06 1.70 0.09 DEF2 landform H 0.70 63.31 0.70 11.16 8.01 0.15 7.71 7.37 2.01 1.68 0.10 Basins: B=Bonney, BP=Bull Pass, F=Fryxell, H=Hoare, VD=Victoria Dunes, WV=Wright Valley

151

Sample ID Type Basin LOI SiO2 TiO2 Al2O3 Fe 2O3 MnO MgO CaO Na2O K2O P2O5 (%) (%) (%) (%) (%) (%) (%) (%) (%) DEF3 landform H 0.56 64.17 0.54 10.99 7.50 0.14 7.95 6.92 2.16 1.74 0.08 HB1 landform H 0.69 62.23 0.57 12.44 6.82 0.12 6.90 6.60 2.61 1.90 0.11 HB2 landform H 0.60 63.68 0.52 12.51 6.81 0.12 7.23 6.44 2.64 1.87 0.11 HB3 landform H 0.36 62.82 0.55 10.75 8.14 0.15 8.80 7.21 2.07 1.56 0.09 NB landform H 0.67 65.15 0.60 13.21 6.41 0.12 5.25 6.24 2.57 2.21 0.11 VD1 landform VD 0.16 70.25 0.32 7.00 7.22 0.14 7.37 6.91 1.09 0.84 0.04 VD2 landform VD 0.58 73.65 0.27 8.95 4.29 0.08 3.72 5.20 1.47 1.28 0.05 VD3 landform VD 0.36 70.43 0.37 7.34 6.78 0.13 6.81 6.44 1.27 1.02 0.05 VD4 landform VD 0.20 66.64 0.38 6.24 9.13 0.18 9.72 8.06 0.90 0.68 0.04 VD5 landform VD 0.32 71.68 0.32 8.63 5.33 0.10 4.72 6.02 1.28 1.02 0.05 WV1 landform WV 0.60 67.60 0.33 13.33 4.30 0.08 3.56 5.22 2.94 2.47 0.10 WV2 landform WV 0.49 68.34 0.34 13.39 4.73 0.09 4.00 5.42 2.87 2.44 0.08

152 WV3 landform WV 0.47 67.31 0.43 11.89 6.39 0.12 5.89 6.22 2.48 2.00 0.08 WV4 landform WV 0.52 68.98 0.35 11.53 4.91 0.09 4.24 5.31 2.43 2.03 0.07

Basins: B=Bonney, BP=Bull Pass, F=Fryxell, H=Hoare, VD=Victoria Dunes, WV=Wright Valley

152

APPENDIX C

Silt XRF and LOI data in weight percent major elemental oxide.

153

Sample ID Type Basin LOI SiO2 TiO2 Al2O3 Fe 2O3 MnO MgO CaO Na2O K2O P2O5 (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) Taylor glacier B 3.18 54.63 1.07 13.85 9.38 0.16 6.48 6.97 1.77 1.89 0.21 E. Lake Bonney lake B 3.42 53.75 1.17 13.56 9.30 0.15 5.88 7.43 2.09 2.14 0.29 W. Lake Bonney lake B 4.01 53.81 0.97 12.98 8.65 0.15 5.91 9.39 2.00 2.02 0.20 Bull Pass landform BP 7.62 50.82 0.91 13.21 8.61 0.14 5.30 8.55 3.33 1.19 0.15 Commonwealth glacier F 3.47 54.11 1.52 13.87 8.58 0.15 5.27 6.43 2.46 2.16 0.36 Commonwealth DUP glacier F 3.47 55.52 1.57 14.30 8.83 0.15 5.39 6.59 2.52 2.21 0.38 Lake Fryxell lake F 6.06 53.40 1.52 13.41 8.75 0.17 5.41 7.31 2.24 2.05 0.39 Canada glacier F/H 3.66 55.00 1.49 13.65 9.47 0.17 5.92 7.67 2.18 2.07 0.38 Canada DUP glacier F/H 2.84 53.35 1.40 13.48 9.41 0.16 5.96 7.05 2.13 2.00 0.37 Howard glacier H 3.87 53.90 1.57 13.66 9.48 0.17 5.69 7.15 2.32 2.06 0.34

15 Lake Hoare lake H 3.94 53.53 1.38 13.22 9.41 0.16 6.05 7.61 2.18 1.90 0.35

4 Defile landform H 4.71 51.50 1.32 11.77 8.73 0.15 6.56 11.11 1.98 1.73 0.46

Hoare Beach landform H 2.09 53.71 1.64 12.42 9.98 0.18 7.19 8.69 2.04 1.67 0.48 Nussbaum landform H 1.75 57.81 1.16 14.46 8.52 0.14 5.05 6.85 2.50 2.44 0.34 Victoria Dunes landform VD 1.07 52.24 2.49 11.17 12.94 0.25 8.04 9.61 1.43 1.00 0.29 Wright Valley landform WV 0.95 59.14 1.30 13.40 8.62 0.16 5.39 7.24 2.37 2.12 0.30 Basins: B=Bonney, BP=Bull Pass, F=Fryxell, H=Hoare, VD=Victoria Dunes, WV=Wright Valley

154

APPENDIX D

ANOVA results for solid samples.

155

Element Group Type Size Fraction F F-crit p-value F>F-crit Element Group Type Size Fraction F F-crit p-value F>F-crit LOI Sample Type sand 2.71 3.89 0.107 No Mg Sample Type sand 5.76 3.16 0.005 Yes LOI Sample Type silt 0.60 3.98 0.565 No Mg Sample Type silt 0.63 3.74 0.545 No LOI Basin sand 15.74 3.20 0.000 Yes Mg Basin sand 3.52 3.20 0.038 Yes LOI Basin silt 1.00 4.26 0.404 No Mg Basin silt 1.31 3.98 0.310 No Si Sample Type sand 11.16 3.16 0.000 Yes Ca Sample Type sand 2.85 3.16 0.066 No Si Sample Type silt 0.13 3.74 0.882 No Ca Sample Type silt 4.06 3.74 0.041 Yes Si Basin sand 7.41 3.20 0.002 Yes Ca Basin sand 3.42 3.20 0.041 Yes Si Basin silt 0.06 3.98 0.945 No Ca Basin silt 1.06 3.98 0.380 No Ti Sample Type sand 16.60 3.16 <0.001 Yes Na Sample Type sand 13.62 3.16 <0.001 Yes Ti Sample Type silt 0.42 3.74 0.666 No Na Sample Type silt 0.17 3.74 0.847 No Ti Basin sand 16.97 3.20 <0.001 Yes Na Basin sand 1.87 3.20 0.165 No Ti Basin silt 10.63 3.98 0.003 Yes Na Basin silt 5.03 3.98 0.028 Yes Al Sample Type sand 23.87 3.16 <0.001 Yes K Sample Type sand 19.50 3.16 <0.001 Yes Al Sample Type silt 3.01 3.74 0.082 No K Sample Type silt 2.15 3.74 0.153 No Al Basin sand 4.45 3.20 0.017 Yes K Basin sand 3.79 3.20 0.030 Yes 156 Al Basin silt 0.91 3.98 0.430 No K Basin silt 0.58 3.98 0.578 No

Fe Sample Type sand 2.39 3.16 0.101 No P Sample Type sand 18.30 3.16 <0.001 Yes Fe Sample Type silt 0.33 3.74 0.726 No P Sample Type silt 0.19 3.74 0.830 No Fe Basin sand 0.14 3.20 0.866 No P Basin sand 23.45 3.20 <0.001 Yes Fe Basin silt 0.38 3.98 0.691 No P Basin silt 11.74 3.98 0.002 Yes Mn Sample Type sand 3.35 3.16 0.042 Yes Mn Sample Type silt 0.44 3.74 0.652 No Mn Basin sand 0.64 3.20 0.532 No Mn Basin silt 0.45 3.98 0.647 No

156

APPENDIX E

Leach 1 (cold water) water chemistry – major ions and nutrients.

157

Leach 1 - Cold Water + + 2+ 2+ - - 2- - - - 3- Na K Mg Ca Cl NO3 SO4 HCO3 * NO3 +NO2 TN PO4 TP H4SiO4 ID (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (neq g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) Aeolian Landforms Hoare Beach 103.93 55.18 78.22 229.18 51.90 23.27 23.01 645.98 8.67 22.92 0.15 0.90 25.19 Hoare Beach DUP 1 7.35 27.67 0.41 0.84 26.38 Hoare Beach DUP 2 6.61 31.54 0.49 0.75 26.01 Hoare Beach DUP 3 7.03 21.48 0.51 0.69 25.37 Hoare Beach DUP 4 8.71 31.47 0.40 0.57 24.82 Hoare Beach DUP 5 3.48 22.51 0.36 0.51 25.70 Defile 1993.27 242.98 264.66 639.09 2396.98 76.82 392.79 760.60 62.99 70.82 0.27 34.98 65.42 Victoria Dunes 555.36 49.40 89.28 62.67 168.11 31.70 31.41 637.30 32.97 46.83 nd 31.77 13.25 Bull Pass 824.50 22.77 79.71 62.57 398.84 82.96 86.65 466.32 74.13 79.03 0.23 34.77 15.99 Elevated Sediment Traps Fryxell EST 6376.13 310.63 172.93 304.63 4116.44 195.62 937.83 1413.64 157.44 187.12 4.50 50.29 46.84 Hoare EST 5839.76 337.92 362.07 728.73 4609.48 113.37 1143.11 1294.20 103.89 117.78 0.22 37.30 40.21 Hoare EST DUP 1 71.93 105.05 0.91 40.01 41.74 Hoare EST DUP 2 78.83 108.03 1.18 39.70 41.18 Hoare EST DUP 3 70.26 96.30 1.10 39.86 48.57 Hoare EST DUP 4 87.03 114.69 1.23 39.57 47.38

158 W. Bonney EST 6367.00 433.87 1355.69 4703.98 8829.15 487.34 4966.97 nd 432.19 479.32 0.11 34.92 28.02

Glaciers E. Commonwealth 999.77 140.53 148.62 316.13 627.31 0.00 141.99 1142.49 13.20 575.22 8.59 16.93 156.35 W. Commonwealth 985.03 212.03 221.75 528.40 980.45 0.00 129.91 1407.50 14.49 1620.82 1.85 4.26 166.97 E. Canada 369.74 121.98 108.50 336.52 419.71 34.84 78.88 684.18 68.23 1177.43 1.09 3.41 105.46 W. Canada 273.77 80.65 99.16 224.20 455.25 22.85 96.39 321.95 48.64 195.05 0.24 1.46 47.21 Howard 190.83 112.86 103.84 640.00 54.72 11.42 15.35 1531.13 22.78 1464.47 0.68 1.90 78.39 Taylor 316.20 57.84 91.05 277.24 513.92 84.39 109.34 283.73 84.30 124.84 0.12 36.04 31.19 Lakes Lake Fryxell 672.49 225.53 348.81 641.01 1418.22 nd 132.38 1178.59 35.68 780.14 3.09 5.10 56.07 Lake Hoare 147.63 56.51 85.78 199.17 348.63 10.28 26.83 350.94 23.84 113.90 0.18 36.55 22.46 Mummy Pond 1370.16 205.00 250.80 540.78 1278.31 46.26 575.56 658.30 18.10 35.95 0.14 36.41 34.13 E. Lake Bonney 222.84 59.95 84.98 97.60 69.39 3.57 20.71 500.19 12.60 258.65 0.40 2.20 42.41 W. Lake Bonney 2599.54 200.82 360.26 1111.31 3242.60 178.00 1201.11 nd 163.99 601.07 0.73 37.67 47.52 W. Lake Bonney DUP 3235.27 176.43 1196.92 not calc. nd = not detected; *Calculated alkalinity

158

APPENDIX F

Leach 2 (freeze-thaw) water chemistry – major ions and nutrients.

159

Leach 2 - Freeze-Thaw + + 2+ 2+ - - 2- - - - 3- Na K Mg Ca Cl NO3 SO4 HCO3 * NO3 +NO2 TN PO4 TP H4SiO4 (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (neq g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) (nmol g-1) Aeolian Landforms Hoare Beach 78.90 53.42 82.32 176.25 11.28 5.00 6.91 619.36 14.51 28.03 0.62 nd 47.08 Hoare Beach REP 1 nd nd 0.87 0.49 67.29 Hoare Beach REP 2 nd 12.02 0.94 0.65 66.23 Hoare Beach REP 3 nd nd 0.94 0.69 70.64 Hoare Beach REP 4 2.04 5.49 1.02 0.68 74.22 Hoare Beach REP 5 nd 3.29 0.93 0.50 77.44 Defile 863.17 104.91 138.90 243.98 567.51 23.85 93.15 934.08 22.44 32.91 0.17 36.70 67.75 Victoria Dunes 452.21 77.52 112.96 82.39 58.11 10.57 18.16 797.01 15.50 24.47 0.12 nd 59.15 Bull Pass 493.31 18.25 72.08 47.28 106.62 24.70 24.54 550.42 26.18 33.35 0.22 0.67 52.43 Elevated Sediment Traps Fryxell EST 2531.60 133.03 112.63 155.25 1382.68 61.40 204.47 1311.80 43.14 59.62 5.18 50.26 73.53 Hoare EST 2041.74 133.93 122.25 193.44 1114.15 27.70 212.44 1201.05 3.47 16.31 0.15 36.76 76.54 Hoare EST DUP 1106.25 27.86 212.69 Hoare EST REP 1 20.59 34.07 2.85 43.65 156.05 Hoare EST REP 2 22.00 46.12 2.80 43.59 147.87 Hoare EST REP 3 15.84 31.79 2.96 44.05 175.41 16 Hoare EST REP 4 18.90 37.49 2.46 42.47 144.23 0 W. Bonney EST 2434.42 189.02 347.30 948.74 2387.95 146.22 1063.16 524.83 121.32 151.64 nd 35.07 50.90

W. Bonney EST DUP 2386.26 145.57 1064.30 not calc. Glaciers E. Commonwealth 551.20 84.92 142.43 224.84 137.08 nd 68.58 1082.83 10.30 497.09 4.01 12.94 171.97 W. Commonwealth 453.54 116.14 160.53 292.77 244.83 4.71 54.72 1077.91 12.36 911.67 2.10 5.35 224.47 E. Canada 201.57 74.56 107.04 214.44 112.26 7.43 48.15 671.53 13.05 585.36 0.89 1.82 172.50 W. Canada 129.50 44.76 87.90 124.27 99.29 5.14 56.74 371.83 10.25 110.48 0.19 0.61 51.02 Howard 116.58 77.12 105.15 444.25 18.05 2.43 8.70 1222.30 12.32 587.57 0.86 1.88 104.74 Taylor 180.34 28.10 88.52 145.42 177.70 29.27 74.36 313.13 31.10 61.05 nd 3.96 39.01 Lakes Lake Fryxell 291.57 120.55 254.26 342.59 300.68 188.34 42.83 984.39 172.53 657.98 2.97 2.35 0.38 Lake Hoare 89.24 37.10 79.60 121.22 89.13 3.14 19.95 386.85 11.63 64.71 0.22 0.81 61.07 Mummy Pond 831.83 108.34 130.73 235.87 508.28 20.13 126.94 867.38 15.11 24.85 0.19 35.15 47.39 E. Lake Bonney 124.73 38.06 95.99 89.46 20.87 nd 9.98 461.69 8.45 113.12 0.32 1.07 65.97 W. Lake Bonney 833.21 65.43 128.13 247.44 680.90 39.27 277.99 321.54 46.71 221.63 0.24 37.37 65.95 nd = not detected; *Calculated alkalinity

160

APPENDIX G

Estimated bulk composition of aeolian sediments based on grain size distribution and

composition of the specific sites used in the experimental leaching methods.

161

Estimated bulk composition (g kg-1)

Na K Mg Ca Fe P Aeolian Landforms Hoare Beach 15 13 53 52 57 0.4 Defile 16 14 48 50 52 0.4 Victoria Dunes 8 7 44 49 51 0.2 Bull Pass 12 12 43 46 49 0.2 Elevated Sediment Traps Fryxell EST 25 20 27 41 39 1.0 Hoare EST 21 17 37 43 42 0.4 W. Bonney EST 20 19 35 40 41 0.4 Glaciers E. Commonwealth 20 22 26 42 42 1.1 W. Commonwealth 21 22 25 41 41 1.0 E. Canada 20 19 27 47 40 0.8 W. Canada 23 18 24 49 38 0.7 Howard 20 18 30 51 45 0.8 Taylor 18 21 41 42 52 0.5 Lakes Lake Fryxell 19 18 29 42 41 0.7 Lake Hoare 22 19 32 38 37 0.4 Mummy Pond 25 20 29 40 36 0.5 E. Lake Bonney 15 15 38 46 50 0.4 W. Lake Bonney 15 15 27 45 45 0.4 Calculated as a weighted average based on grain size distribution (Appendix A) and composition (Appendices B & C)

162

APPENDIX H

Surface area of lakes and glacier ablation areas of Taylor Valley used in calculation of

solute and nutrient flux.

163

Surface Area (km2) Lake Fryxell 7.1 Lake Hoare/Chad 2.1 E. Lake Bonney 3.3 W. Lake Bonney 1.0 Commonwealth1 40.0 Canada2 8.5 Howard1 5.1 1Basagic et al . personal communication 2Lewis et al. (1999) Lake data available from MCM-LTER website: http://www.mcmlter.org

164