Advancing leaf wax paleohydrology: From plant source to sediment sink

A dissertation submitted to the Graduate School of the University of Cincinnati in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

in the Department of Geology of the College of Arts and Sciences

by

Erika Jacob Freimuth B.S., Cornell University, 2009

April 20th, 2018

Dissertation Committee

Dr. Aaron F. Diefendorf, Chair Dr. Broxton Bird Dr. Brooke Crowley Dr. Thomas V. Lowell Dr. Dylan Ward Abstract

Plant wax hydrogen isotopes (δDwax) are important archives of water isotopes in the geologic past. Plant waxes are common in lacustrine sediments, providing widely distributed terrestrial records of hydrologic change in locations where other water isotope proxies (e.g., ice sheets, cave deposits) do not exist.

However, application of δDwax to reconstruct past precipitation hydrogen isotopes (δDp) is subject to uncertainties resulting from modification of original water isotope signals in plants and sediments. This dissertation addresses several of these uncertainties through a series of studies that identify controls on this proxy system at increasing spatial scales, from individual plants to a single bog catchment to a regional survey of multiple lake catchments.

Chapter 2 examines the seasonal timing of leaf wax production for two commonly used paleohydrology proxies, n-alkanes and n-alkanoic acids. The goals of this project are to constrain the drivers of seasonal

δDwax variability among tree species and to determine whether leaf waxes in temperate forests record seasonally-biased precipitation. This study monitors the δD of environmental and plant waters and resulting leaf waxes as they evolve over a growing season from bud break to leaf fall at Brown’s Lake

Bog (BLB), Ohio, USA. This is the first study to track seasonal changes in both n-alkanes and n-alkanoic acids, and offers novel insight into systematic differences in water isotope fractionation and the timing, duration and amount of wax production between compound classes.

Chapter 3 compares the abundance, molecular distribution and isotopic composition (δD and δ13C) of n- alkanes and n-alkanoic acids in bog sediments with all major plant species growing in the catchment of

BLB. This project aims to identify factors that influence the integration of leaf waxes from source (plants) to sink (sediments) in a single catchment. Results of this study offer insight into sediment bias toward particular plant sources and a framework for estimating apparent fractionation (εapp) between water and wax δD in temperate forested settings.

i

Chapter 4 expands from a single catchment to asses δDwax variability at the sediment level for multiple lakes across the Adirondack Mountains, NY, USA. We evaluate plant waxes in surface sediments from 12

Adirondack lakes that have similar δDp but a range of different catchment properties. The goals of this study are to 1) quantify the variability in sediment δDwax in the absence of major differences in δDp among sites within the same region; 2) identify possible catchment-level drivers of this variability (e.g., vegetation cover, fluvial complexity, depositional setting); and 3) compare the molecular and δD composition of two plant wax compounds (n-alkanes and n-alkanoic acids) to assess potential differences in their sources. Results from this project underscore the influence that site-specific factors can have on

δDwax records.

This dissertation identifies factors and processes that influence the biological production, spatial integration and regional variation in δDwax signals from forested temperate settings. The results reveal promising directions for future proxy research, and ultimately will contribute to reducing uncertainties inherent in proxy-based observations of hydrologic responses to climate perturbations in the geologic past.

ii

iii Acknowledgements

The following sources provided funding to support this research: The National Science Foundation, The American Chemical Society, The American Philosophical Society, University of Cincinnati (UC) Sigma Xi, Geological Society of America, the UC Research Council, and the UC Department of Geology.

Access to field sites for sample collection was provided by the Ohio Nature Conservancy, the New York State Department of Environmental Conservation, the Adirondack Mountain Club and the Newcomb Campus of the State University of New York College of Environmental Science and Forestry.

This dissertation is the result of the dedicated efforts of many people. Foremost among these is my advisor, Dr. Aaron Diefendorf, whose guidance has been invaluable for every aspect of this research and for my training and development as a scientist. Since my first weeks in graduate school, Dr. Diefendorf and Dr. Thomas Lowell have involved me in the enterprising process of developing new research ideas – an ongoing education for which I am especially grateful.

I would like to acknowledge my committee members, Dr. Broxton Bird, Dr. Brooke Crowley, Dr. Lowell and Dr. Dylan Ward for generously sharing their time and expertise to strengthen my dissertation research and guide my professional development. I also thank the faculty and staff of the UC Geology Department for their advice, feedback and support.

A special thanks to my fellow graduate students, especially to Dr. Yeon Jee Suh, Ben Bates and Kristen Schlanser for their support in the lab and their tolerance for lengthy leaf wax discussions. Many thanks to undergraduates Kelly Grogan, Anna Schartman, Katie McNulty, Daniel Frost and Douglas Sberna for the assistance, insight and enthusiasm they brought to the lab and the field. Dr. Greg Wiles and Nicholas Wiesenberg provided essential support for several years of sample collection at Brown’s Lake Bog, OH.

As always, my friends and family have been inexhaustible sources of inspiration and wisdom—especially Kristen Meisner who supplied a week’s worth of levity every Saturday and Caleb Marhoover who filled our home with art and aspiration. Finally, I would like to thank Mia Jacob and Paul Freimuth for being my finest examples. I would never have chased this if they had not shown me how.

iv Table of Contents

ABSTRACT …………………………………………………………………………………………….. i

Acknowledgements ……………………………………………………………………………………... iv

Chapter 1: Introduction ………………………………………………………………………………. 1

References ……………………………………………………………………………………………… 5

Figure …………………………………………………………………………………………………… 7

Chapter 2: Seasonality of n-alkyl lipid production and δD composition in a temperate forest and implications for molecular paleohydrology

ABSTRACT …………………………………………………………………………………………… 8

1. Introduction …………………………………………………………………………………………. 10

2. Materials and Methods ………………………………………………………………………………. 11

2.1. Site description and environmental waters ……………………………………………….. 12

2.2. Plant sampling ……………………………………………………………………………. 12

2.3. Leaf and xylem water extraction and hydrogen isotope analysis ………………………… 13

2.4. Lipid extraction …………………………………………………………………………… 14

2.5. Lipid identification and quantification …………………………………………………… 15

2.6. Lipid hydrogen isotope analysis …………………………………………………………. 16

2.7. Bulk carbon isotope analysis …………………………………………………………….. 16

3. Results ………………………………………………………………………………………………. 17

3.1. Hydrogen isotope composition of environmental waters at Brown’s Lake Bog ………… 17

3.2. Leaf mass per area during leaf development …………………………………………….. 18

3.3.1. Leaf n-alkane abundance and chain length distribution ……………………………….. 18

3.3.2. Leaf n-alkanoic acid abundance and chain length distribution ………………………… 20

3.4. Plant source waters: xylem and leaf water δD …………………………………………… 21

v 3.5.1. n-Alkane hydrogen isotope composition ………………………………………………. 22

3.5.2. Hydrogen isotope fractionation in n-alkanes …………………………………………… 23

3.6.1. n-Alkanoic acid hydrogen isotope composition ………………………………………... 24

3.6.2. Hydrogen isotope fractionation in n-alkanoic acids ……………………………………. 25

3.7. Bulk foliar δ13C values …………………………………………………………………… 26

4. Discussion …………………………………………………………………………………………… 26

4.1.1. Timing of n-alkane synthesis …………………………………………………………… 26

4.1.2. Controls on δD of n-C29 alkane …………………………………………………………. 29

4.2.1. Timing of n-alkanoic acid synthesis ……………………………………………………. 32

4.2.2. Controls on δD of n-C28 alkanoic acid ………………………………………………….. 32

4.3. Implications for interpreting δDwax from sediments ……………………………………… 34

4.3.1. Interspecies δDwax variability within a growth form …………………………... 34

4.3.2. n-Alkanes and n-alkanoic acids record different source water information …… 35

4.3.3. Estimating εapp values and uncertainty ………………………………………… 36

5. Conclusions …………………………………………………………………………………………. 37

Acknowledgments ……………………………………………………………………………………... 39

References ……………………………………………………………………………………………… 40

Figures …………………………………………………………………………………………………. 45

Tables ………………………………………………………………………………………………….. 52

Chapter 3: Sedimentary plant waxes in a temperate bog are biased toward woody vegetation

ABSTRACT ……………………………………………………………………………………………. 56

1. Introduction ………………………………………………………………………………………….. 58

1.1. Drivers of εapp variability in living biomass …………………………………………….… 59

1.2. Leaf wax transfer from plant to sediment ………………………………………………… 60

2. Materials and Methods ………………………………………………………………………………. 61

vi 2.1. Site description and sediment collection ………………………………………………….. 61

2.2. Plant sampling for lipid and stem water analysis ……………………………………….… 62

2.3. Monitoring wet deposition of particulate n-alkanes …………………………………….... 63

2.4. Xylem water extraction and δD analysis ……………………………………………….… 63

2.5. Lipid extraction, quantification and stable isotope analysis ……………………………… 64

2.6. Bulk foliar carbon isotope analysis …………………………………………………….…. 66

3. Results …………………………………………………………………………………………….…. 67

3.1. Environmental waters at BLB ………………………………………………………….…. 67

3.2. Leaf wax molecular and isotopic composition in plants ……………………………….…. 68

3.3. Particulate wax flux and composition ………………………………………………….…. 69

3.4. Bog sediment chronology and bulk properties …………………………………………… 70

3.5. Plant wax molecular and isotopic composition in bog sediments ………………………... 71

4. Discussion …………………………………………………………………………………………… 71

4.1. Drivers of δDwax differences among plant species and growth forms …………………..… 71

4.2. Sediment bias toward n-alkane sources ……………………………………………….….. 74

4.3 Sediment bias toward n-alkanoic acid sources ………………………………………….… 78

4.4. Molecular and isotopic differences between n-alkanes and n-alkanoic acids ………….… 79

5. Conclusions …………………………………………………………………………………..……… 81

Acknowledgements …………………………………………………………………………………..… 82

References ………………………………………………………………………………………...……. 83

Figures ………………………………………………………………………………………………….. 87

Tables …………………………………………………………………………………………….…….. 95

Chapter 4: Comparative fidelity of n-alkanes and n-alkanoic acids to regional-scale precipitation

δD and catchment-scale vegetation: a survey of lake sediments in the Adirondack Mountains, NY

(USA)

vii ABSTRACT ……………………………………………………………………………………………. 98

1. Introduction ………………………………………………………………………………………… 100

2. Materials and methods …………………………………………………………………………….... 102

2.1. Sampling locations and sediment collection ……………………………………………... 102

2.2. Bulk sediment characterization ……………………………………………………….….. 104

2.3. Lipid extraction, separation and quantification ………………………………………...… 105

2.4. δD analysis of n-alkanes and n-alkanoic acids ………………………………………...… 106

2.5. Water sampling and isotope analysis ………………………………………………….…. 107

3. Results …………………………………………………………………………………….………… 109

3.1. Bulk sediment characteristics ………………………………………………….………… 109

3.2. Adirondack isotope hydrology …………………………...……………………….……… 110

3.2.1. Surface and meteoric water isotopes ………………………………...... ……… 110

3.2.2. Plant water isotopes …………...……………………………………….……… 111

3.3. Sedimentary plant lipid abundance and distribution ………………………………..……. 113

3.4. Sedimentary plant lipid δD and εapp composition …………………………………..….… 114

4. Discussion ………………………………………………………………………………………...… 116

4.1. Drivers of plant lipid distributions and δD among lakes …………….………………...… 116

4.2. Do n-alkanes and n-alkanoic acids record the same water signal? ………………….…… 118

4.3. εapp in the Adirondacks compared with other biomes ………………………………….… 119

5. Conclusions …………………………………………………………………………………...…….. 121

Acknowledgements ……………………………………………………………………………………. 122

References ……………………………………………………………………………………..………. 123

Figures ………………………………………………………………………………………….……… 127

Tables ………………………………………………………………………………………….………. 140

Chapter 5: Conclusion ……………………………………………………………………….………. 145

viii Chapter 1

Introduction

Atmospheric greenhouse gas concentrations are linked to surface temperature, evaporation, atmospheric moisture and ultimately increased precipitation rates in a positive feedback cycle (Trenberth, 1999; Milly et al., 2002). Future climate change is therefore expected to alter the planet’s hydrologic cycle with far reaching implications for human and natural systems (Milly et al., 2002; Palmer and Räisänen, 2002;

Milly et al., 2005). Anticipating future climate and hydrologic cycle changes is critical for global public health as well as economic and geopolitical stability (IPCC, 2014). Climate and hydrological models used to predict future changes are often evaluated by comparing model hindcasts with independent proxy data from the geological record (e.g., Ward et al., 2007). Therefore, validation of climate model performance is linked to the accuracy of information derived from paleoclimate proxies, including molecular biomarkers. These molecules can pass information about an organism’s growth environment to the geological record, preserving evidence of past climate and ecological change. However, deriving primary climate signals encoded in these secondary markers can lead to large uncertainties. This dissertation addresses several major sources of uncertainty inherent in the use of plant waxes to reconstruct past hydrologic conditions.

The hydrogen isotopic composition of leaf waxes (dDwax) primarily reflects that of plant source water

(i.e., precipitation; δDp). Therefore, dDwax preserved in lake sediments is increasingly used to unlock information about the paleohydrology of continental environments. Such records are applicable to a range of inquiry, including constraining the onset and duration of past drying events (e.g., Rach et al., 2014) and reconstructing changes in moisture balance (e.g., Jacob et al., 2007) or the strength of monsoon systems

(e.g., Bird et al., 2014). However, quantitative reconstruction of δDp is hindered by limited understanding

1 of controls on dDwax during leaf wax production and during transfer of waxes from plant to sediment (Fig.

1). Specifically, constraining the factors that influence apparent fractionation (εapp; the net hydrogen isotopic offset between plant source water and leaf wax), both in modern biomass and in sedimentary leaf waxes is key to advancing quantitative paleohydrology.

Chapter 2 is devoted to understanding the drivers of δDwax and εapp values in modern trees (Fig. 1). εapp values are the net result of many constituent factors and processes that modify plant source water δD prior to and during lipid synthesis (Sachse et al., 2012). Environmental factors, including soil water evaporation (McInerney et al., 2011) and leaf transpiration (Feakins and Sessions, 2010; Sachse et al.,

2010; Kahmen et al., 2013b), result in evaporative deuterium (D)-enrichment of plant waters. Meanwhile, lipid synthesis is associated with a large negative biosynthetic δD fractionation (εbio), the magnitude of which can vary widely (~65‰) among plant species and growth forms (Sessions et al., 1999; Kahmen et al., 2013a). In addition, the seasonal timing and duration of leaf wax synthesis can have significant effects on εapp when plant source water changes through the growing season. Recent evidence indicates that n- alkane production and δD values are averaged across the growing season for temperate deciduous forests

(Sachse et al., 2006; Sachse et al., 2009; Newberry et al., 2015), but biased toward early spring leaf emergence for field grown barley (Sachse et al., 2010) and higher elevation riparian trees (Tipple et al.,

2013). While progress has been made in our understanding of the various controls on δDwax, a large degree of uncertainty around εapp remains. Therefore, we measured the δD of plant waters and resulting leaf waxes from trees in a temperate forest at Brown’s Lake Bog (BLB), Ohio, USA over an entire growing season, capturing snapshots of εapp at weekly to monthly intervals. This is the first study to assess seasonal production of both n-alkanes and n-alkanoic acids even though both compounds are commonly used as δDp proxies.

Chapter 3 expands from individual plants to an entire catchment and the resulting δDwax signals in sediments (Fig. 1). Because εapp is subject to a range of environmental and biological controls, εapp values

2 can range by up to 110‰ among plant species within a single site (Hou et al., 2007; Feakins and

Sessions, 2010; Eley et al., 2014; Cooper et al., 2015). This complicates selection of accurate εapp values for interpretation of sedimentary leaf waxes, which accumulate from a mixture of plant sources at a single site. In addition, transport of leaf waxes from plant source to sediment sink is poorly understood and may bias the composition of sedimentary wax records (Pancost and Boot, 2004; Sachse et al., 2012;

Diefendorf and Freimuth, 2017; Nelson et al., 2018). The goal of this project is to understand how εapp signals at the plant level are integrated in sediments and whether sediments are biased toward particular plant sources. This study used a dual-isotope approach to fingerprint the molecular and isotopic (δD and

δ13C) composition of n-alkanes and n-alkanoic acids in major sources (plants and particulate waxes) and associated sediments within a single catchment at BLB, which is the same site used for Chapter 2.

Chapter 4 extends one step further, from a single lake catchment to a regional survey of sedimentary plant wax (Fig 1). While much recent progress has been made in understanding controls on εapp in modern plants, the processes of catchment-wide mixing, integration and transport that determine εapp values and variability in sediments remain poorly understood (Sachse et al., 2012). Lacustrine records in particular are vulnerable to a high degree of uncertainty because terrestrial plant-derived waxes may be expressed non-uniformly due to sourcing from vegetation assemblages that are unique to each lake catchment. For example, δDwax values in the surface sediments of 20 to 24 lakes within a single region can differ by as much as 44‰ and 60‰ as a result of differences in catchment vegetation or local aridity in the Arctic and

Central America, respectively (Douglas et al., 2012; Daniels et al., 2017). Considerable variability in surface sediments presents a challenge for estimating εapp values to apply to plant waxes preserved deeper in lake sediments, limiting our ability to reconstruct past δDp with a high degree of confidence. This study seeks to determine the variability in lake sediment δDwax that occurs in the absence of major differences in

δDp. The Adirondack Mountains, NY, USA are uniquely suited to this approach due to the abundance of undeveloped lakes with diverse depositional characteristics and vegetation structure. We surveyed δDwax in 12 Adirondack lakes that receive the similar δDp but vary markedly in shoreline and catchment

3 vegetation composition, lake to watershed area ratio, fluvial complexity and bulk organic and inorganic sediment characteristics.

Together, this series of calibration studies on δDwax systematics in modern plants and recent sediments advances our understanding of this proxy system and its utility as a tracer for water isotopes in the geologic past.

4 REFERENCES

Cooper, R.J., Krueger, T., Hiscock, K.M. and Rawlins, B.G. (2015) High-temporal resolution fluvial sediment source fingerprinting with uncertainty: a Bayesian approach. Earth Surface Processes and Landforms 40, 78-92. Daniels, W.C., Russell, J.M., Giblin, A.E., Welker, J.M., Klein, E.S. and Huang, Y. (2017) Hydrogen isotope fractionation in leaf waxes in the Alaskan Arctic tundra. Geochim. Cosmochim. Acta 213, 216-236. Diefendorf, A.F. and Freimuth, E.J. (2017) Extracting the most from terrestrial plant-derived n-alkyl lipids and their carbon isotopes from the sedimentary record: A review. Org. Geochem. 103, 1- 21. Douglas, P.M.J., Pagani, M., Brenner, M., Hodell, D.A. and Curtis, J.H. (2012) Aridity and vegetation composition are important determinants of leaf-wax δD values in southeastern Mexico and Central America. Geochim. Cosmochim. Acta 97, 24-45. Eley, Y., Dawson, L., Black, S., Andrews, J. and Pedentchouk, N. (2014) Understanding 2H/1H systematics of leaf wax n-alkanes in coastal plants at Stiffkey saltmarsh, Norfolk, UK. Geochim. Cosmochim. Acta 128, 13-28. Feakins, S.J. and Sessions, A.L. (2010) Controls on the D/H ratios of plant leaf waxes from an arid ecosystem. Geochim. Cosmochim. Acta 74, 2128-2141. Hou, J.Z., D'Andrea, W.J., MacDonald, D. and Huang, Y.S. (2007) Hydrogen isotopic variability in leaf waxes among terrestrial and aquatic plants around Blood Pond, Massachusetts (USA). Org. Geochem. 38, 977-984. Kahmen, A., Hoffmann, B., Schefuß, E., Arndt, S.K., Cernusak, L.A., West, J.B. and Sachse, D. (2013a) Leaf water deuterium enrichment shapes leaf wax n-alkane δD values of angiosperm plants II: Observational evidence and global implications. Geochim. Cosmochim. Acta 111, 50-63. Kahmen, A., Schefuß, E. and Sachse, D. (2013b) Leaf water deuterium enrichment shapes leaf wax n- alkane δD values of angiosperm plants I: Experimental evidence and mechanistic insights. Geochim. Cosmochim. Acta 111, 39-49. McInerney, F.A., Helliker, B.R. and Freeman, K.H. (2011) Hydrogen isotope ratios of leaf wax n-alkanes in grasses are insensitive to transpiration. Geochim. Cosmochim. Acta 75, 541-554. Milly, P.C., Dunne, K.A. and Vecchia, A.V. (2005) Global pattern of trends in streamflow and water availability in a changing climate. Nature 438, 347-350. Milly, P.C.D., Wetherald, R.T., Dunne, K. and Delworth, T.L. (2002) Increasing risk of great floods in a changing climate. Nature 415, 514. Nelson, D.B., Ladd, S.N., Schubert, C.J. and Kahmen, A. (2018) Rapid atmospheric transport and large- scale deposition of recently synthesized plant waxes. Geochim. Cosmochim. Acta 222, 599-617. Newberry, S.L., Kahmen, A., Dennis, P. and Grant, A. (2015) n-Alkane biosynthetic hydrogen isotope fractionation is not constant throughout the growing season in the riparian tree Salix viminalis. Geochim. Cosmochim. Acta 165, 75-85. Palmer, T. and Räisänen, J. (2002) Quantifying the risk of extreme seasonal precipitation events in a changing climate. Nature 415, 512. Pancost, R.D. and Boot, C.S. (2004) The palaeoclimatic utility of terrestrial biomarkers in marine sediments. Marine Chemistry 92, 239-261. Sachse, D., Radke, J. and Gleixner, G. (2006) δD values of individual n-alkanes from terrestrial plants along a climatic gradient – Implications for the sedimentary biomarker record. Org. Geochem. 37, 469-483. Sachse, D., Kahmen, A. and Gleixner, G. (2009) Significant seasonal variation in the hydrogen isotopic composition of leaf-wax lipids for two deciduous tree ecosystems (Fagus sylvativa and Acer pseudoplatanus). Org. Geochem. 40, 732-742.

5 Sachse, D., Gleixner, G., Wilkes, H. and Kahmen, A. (2010) Leaf wax n-alkane δD values of field-grown barley reflect leaf water δD values at the time of leaf formation. Geochim. Cosmochim. Acta 74, 6741-6750. Sachse, D., Billault, I., Bowen, G.J., Chikaraishi, Y., Dawson, T.E., Feakins, S.J., Freeman, K.H., Magill, C.R., McInerney, F.A., van der Meer, M.T.J., Polissar, P., Robins, R.J., Sachs, J.P., Schmidt, H.- L., Sessions, A.L., White, J.W.C., West, J.B. and Kahmen, A. (2012) Molecular paleohydrology: Interpreting the hydrogen-isotopic composition of lipid biomarkers from photosynthesizing organisms. Annu. Rev. Earth Planet. Sci. 40, 221-249. Sessions, A.L., Burgoyne, T.W., Schimmelmann, A. and Hayes, J.M. (1999) Fractionation of hydrogen isotopes in lipid biosynthesis. Org. Geochem. 30, 1193-1200. Tipple, B.J., Berke, M.A., Doman, C.E., Khachaturyan, S. and Ehleringer, J.R. (2013) Leaf-wax n- alkanes record the plant–water environment at leaf flush. Proceedings of the National Academy of Sciences 110, 2659-2664. Trenberth, K.E. (1999) Conceptual framework for changes of extremes of the hydrological cycle with climate change, Weather and Climate Extremes. Springer, pp. 327-339.

6

Figure 1. Evolution of δDwax from plant source water to leaf waxes in living biomass and in sedimentary records, as well as signal stability over geologic timescales in the geo-archive. The δD values of plant source waters (blue) and both modern (green) and sedimentary (brown) leaf waxes are modified by factors indicated with gray arrows. Level of understanding is based on knowledge for n- alkanes. Chapter 2 focuses on the drivers of the δDwax signal in modern plants; Chapter 3 and 4 address poorly-understood processes that determine the molecular and isotopic composition of plant waxes in lake sediments. T = temperature; rH = relative humidity; NADPH = Nicotinamide adenine dinucleotide phosphate (reduced; additional source of hydrogen for wax synthesis). Figure modified from Sachse et al. (2012).

7 Chapter 2

Seasonality of n-alkyl lipid production and δD composition in a temperate forest and implications for molecular paleohydrology1

ABSTRACT

The hydrogen isotopic composition of leaf waxes (δDwax) primarily reflects that of plant source water.

Therefore, sedimentary δDwax records are increasingly used to reconstruct the δD of past precipitation

(δDp) and to investigate paleohydrologic changes. Such reconstructions rely on estimates of apparent fractionation (εapp) between δDp and the resulting δDwax. However, εapp values are modified by numerous environmental and biological factors during leaf wax production. As a result, εapp can vary widely among plant species and growth forms. This complicates estimation of accurate εapp values and presents a central challenge to leaf wax paleohydrology. During the 2014 growing season, we examined εapp in the five deciduous angiosperm tree species (Prunus serotina, Acer saccharinum, Quercus rubra, Quercus alba, and Ulmus americana) that dominate the temperate forest at Brown’s Lake Bog, Ohio, USA. We sampled individuals of each species at weekly to monthly intervals from March to October and report δD values of n-C29 alkanes (δDn-C29 alkane) and n-C28 alkanoic acids (δDn-C28 acid), as well as xylem (δDxw) and leaf water

(δDlw). n-Alkane synthesis was most intense 2-3 weeks after leaf emergence and ceased thereafter, whereas n-alkanoic acid synthesis continued throughout the entire growing season. During bud swell and leaf emergence, δDlw was a primary control on δDn-C29 alkane and δDn-C28 acid values, which stabilized once leaves became fully expanded. Metabolic shifts between young and mature leaves may be an important

1 Freimuth, E.J., Diefendorf, A.F. and Lowell, T.V., 2017. Hydrogen isotopes of n-alkanes and n-alkanoic acids as tracers of precipitation in a temperate forest and implications for paleorecords. Geochimica et Cosmochimica Acta, 206, pp.166-183.

8 secondary driver of δDwax changes during leaf development. In mature autumn leaves of all species, the mean εapp for n-C29 alkane (-107‰) was offset by approximately -19‰ from the mean εapp for n-C28 alkanoic acid (-88‰). These results indicate that in temperate settings n-alkanes and n-alkanoic acids from deciduous trees are distinct with respect to their abundance, timing of synthesis, and εapp values.

9 1. Introduction

The hydrogen isotope composition (δD) of precipitation is governed by isotope effects during hydrologic processes (Craig, 1961). Plants synthesize leaf waxes using hydrogen derived from plant source waters

(including soil, stream and lake water), which are fed by precipitation. Therefore, the hydrogen isotopic composition of leaf waxes (δDwax) primarily reflects that of local precipitation (δDp) both in modern plants (Sachse et al., 2006; Smith and Freeman, 2006; Tipple and Pagani, 2013; Feakins et al., 2016) and in sediments (Sachse et al., 2004; Hou et al., 2008; Polissar and Freeman, 2010; Sachse et al., 2012).

Thus, δDwax from lacustrine and marine sediments has been used to investigate paleohydrology throughout the Cenozoic (Pagani et al., 2006; Tierney et al., 2008; Rach et al., 2014). The net offset between sedimentary δDwax (measured) and past δDp (calculated) is known as apparent fractionation (εapp).

Estimates of εapp are therefore critical for accurate δDp reconstruction. However, εapp remains the largest source of uncertainty in leaf wax paleohydrology (Sachse et al., 2012; Polissar and D’Andrea, 2014).

εapp values are influenced by a suite of environmental and biological factors. For example, deuterium (D)- enrichment of the biosynthetic water pool from which lipids are synthesized can be driven by soil evaporation (McInerney et al., 2011) and leaf transpiration (Feakins and Sessions, 2010; Sachse et al.,

2010; Kahmen et al., 2013b). In addition, leaf wax biosynthesis results in considerable D-depletion of the wax product relative to the biosynthetic water pool from which it is synthesized (Sachse et al., 2012). This

δD offset is known as biosynthetic fractionation (εbio). Some evidence suggests that εbio is constant within a species (Sessions et al., 1999; Tipple et al., 2015) while other studies have found that εbio may vary throughout the growing season (Newberry et al., 2015; Sachse et al., 2015). Among species, εbio can vary by up to 65‰ (Kahmen et al., 2013b). This may contribute to considerable interspecies εapp differences of up to 65‰ among species in a temperate deciduous forest (Hou et al., 2007), 99‰ in an arid shrub ecosystem (Feakins and Sessions, 2010) and 109‰ in a temperate saltmarsh (Eley et al., 2014). Wide

10 variability in εapp among plants presents a challenge to the accuracy of εapp estimates applied to sedimentary δDwax.

The seasonal timing and duration of leaf wax synthesis is also important for interpretation of δDwax values. If leaf waxes are produced during discrete time intervals, δDwax records in sediments may be biased toward those specific portions of the growing season. Recent evidence indicates that n-alkane production and δD values are an average of the growing season for temperate deciduous forests (Sachse et al., 2006; Sachse et al., 2009; Newberry et al., 2015), but biased toward early spring leaf emergence for field grown barley (Sachse et al., 2010) and for higher elevation riparian trees (Tipple et al., 2013).

Further, while both n-alkanes and n-alkanoic acids are commonly used for leaf wax paleohydrology, relatively few calibration studies of δDwax in modern vegetation have included n-alkanoic acids

(Chikaraishi and Naraoka, 2007; Hou et al., 2007; Gao et al., 2014; Feakins et al., 2016). Therefore, the implications of using either compound for δDp reconstructions remain largely unknown. Several surveys of vegetation and sediments suggest that leaf wax inputs from trees, especially angiosperm trees, dominate the sedimentary record over other growth forms (Sachse et al., 2006; Seki et al., 2010; Sachse et al., 2012; Tipple and Pagani, 2013; Schwab et al., 2015). Therefore, this study is focused on deciduous angiosperm trees as the primary source of leaf wax to sediments. We investigated differences in the timing and amount of n-alkane and n-alkanoic acid production and seasonal changes in leaf wax and source water δD composition among five deciduous angiosperm tree species in a temperate forest. Our central objective was to better understand the controls on the δDwax signal produced in species within the tree growth form and thereby improve the accuracy of εapp estimates from temperate forests.

2. Materials and methods

11 2.1. Site description and environmental waters

All sample collection took place at Brown’s Lake Bog State Nature Preserve (BLB; 40.6818 ºN, 82.0645

ºW; 291 masl; 0.4 km2) in northern Ohio, USA. The biome type for BLB is a temperate deciduous broadleaf forest (Kaplan et al., 2002), and the five tree species that grow on the preserve (with mean diameter at breast height (DBH) of sampled individuals) are Prunus serotina (42 cm), Acer saccharinum

(82 cm), Quercus rubra (144 cm), Quercus alba (99 cm), and Ulmus americana (34 cm). The site has a mean annual temperature of 10.1 ºC and mean annual precipitation of 998 mm (PRISM). During the study period (March to October 2014) the mean temperature, relative humidity and total precipitation recorded at an Ohio Agricultural Research and Development Center (OARDC) weather station 15 km from BLB were 16.9 ± 6.4 ºC, 78 ± 9%, and 583 mm, respectively. Additionally, a weather station (Davis

Instruments, Hayward, CA, USA) was installed to log meteorological data on-site. Precipitation was sampled during each site visit from collection bottles fitted with a funnel and a layer of mineral oil to minimize evaporation. Surface waters including ephemeral pools in the forested lowlands as well as bog and lake water were also sampled during each site visit for isotope analysis.

2.2. Plant sampling

Plant material was collected over 12 visits to BLB at 1-5 week intervals from March to October 2014, with higher frequency during spring leaf emergence and expansion. Samples are referenced with the day of year (DOY) of collection (corresponding dates in Table 1 and 2). When sampling took place over two consecutive days (e.g., DOY 152-153), we reference all data using the first day of collection (e.g., DOY

152). Each round of sampling included the same 2-3 individuals from each of the 5 species (11 individuals total). We analyzed a subset of one particular individual from each species (n = 5) over the entire time series, in addition to all individuals (n = 11) sampled on DOY 110, 128, 152, 235, and 274.

12 Except where otherwise indicated, the data presented refer to the subset of 5 individuals analyzed over the entire time series.

Sampling occurred continuously over two consecutive days from approximately 9:00 am to dusk. Due to the number of trees and time-intensive sample collection from the uppermost canopy (>15 m), repeated sampling through the diel cycle to capture δDlw variability was not feasible (see Section 3.4). During each sampling round, one slender branch was removed from the uppermost canopy of each individual using an arborist’s slingshot. We collected sun-exposed canopy material in order to minimize potential canopy effects in leaf water or lipid isotope composition (Graham et al., 2014). Collection of plant material for molecular and isotope analysis followed the methods of Feakins and Sessions (2010) and Kahmen et al.,

(2013a). Between 5-12 randomly selected leaves were stripped from the collected branch and merged into a single bulk sample per individual. Leaves (with mid-vein and petiole removed) and the branch (<1.0 cm diameter, with outer bark removed) from which leaves were stripped were immediately sealed in separate pre-ashed glass exetainer vials with airtight septa screwcaps. All samples were kept in a cooler in the field and frozen (-5 ºC) in the lab until water extraction. From leaf emergence until DOY 206, an additional set of leaf samples was collected to determine leaf dry mass per area (LMA) following Cornelissen et al.

(2003). Within 48 hours of collection, 10 intact leaves from each individual were scanned to determine leaf area using ImageJ software (http://imagej.nih.gov/ij). The same leaves were then freeze-dried and the dry mass and area of each leaf were pooled to determine the mean LMA for each individual and sampling date.

2.3. Leaf and xylem water extraction and hydrogen isotope analysis

Leaf and xylem water were extracted using cryogenic vacuum distillation following the methods of West et al. (2006). Exetainer vials containing frozen leaves and stems were evacuated to a pressure <8 Pa (<60 mTorr), isolated from the vacuum pump, and heated to 100 ºC. Water vapor was collected in borosilicate

13 test tubes immersed in liquid nitrogen for a minimum of 60 minutes, with mean extraction times of 73 minutes for xylem and 63 minutes for leaves. To verify extraction completion, samples were weighed following cryogenic vacuum distillation and then again after freeze drying. Based on the mass difference, the recovery of plant water was >99%. Collected water was thawed and pipetted into 2 ml crimp-top vials and refrigerated at 4 ºC until analysis.

Analysis of leaf and xylem water δD was made by headspace equilibration using 200 µl of water transferred to exetainer vials with a Pt catalyst added. Samples were purged using 2% H2 in He for 10 minutes at 120 ml/min and equilibrated at 25 ºC for 1 hour. The isotopic composition of equilibrated headspace gas was analyzed on a Thermo Delta V Advantage isotope ratio mass spectrometer (IRMS) with a Thermo Gasbench II connected via a Conflo IV interface. Data were normalized to the V-

SMOW/SLAP scale using three in-house reference standards. Precision and accuracy based on an independent standard were 1.85‰ (1σ, n = 43) and 0.39‰ (n = 43), respectively.

2.4. Lipid extraction

After extraction of bud/leaf waters, dry samples were ground to a powder. Separate aliquots of the ground bud or leaf material were used for lipid extraction and bulk δ13C analysis (see Section 2.7). For lipid extraction, ~200 mg of powdered leaves were extracted by sonicating twice with 20 ml of DCM/MeOH

(2:1, v/v), centrifuging and pipetting the lipid extract into a separate vial after each round of sonication.

The total lipid extract was dried under a gentle stream of nitrogen and base saponified to cleave fatty esters with 3 ml of 0.5 N KOH in MeOH/H2O (3:1, v/v) for 2h at 75 ºC. Once cool, 2.5 ml of NaCl in water (5%, w/w) was added and acidified with 6 N HCl. The solution was extracted with hexanes/DCM

(4:1, v/v), neutralized with NaHCO3/H2O (5%, w/w), and water was removed through addition of

Na2SO4.

14 Compound classes were separated using 0.5 g of aminopropyl-bonded silica gel in 6 ml solid phase extraction columns. Hydrocarbons were eluted with 4 ml of hexanes, ketones were eluted with 8 ml of hexanes/DCM (6:1, v/v), alcohols were eluted with 8 ml of DCM/acetone (9:1, v/v), and acids were eluted last with 8 ml of DCM/85% formic acid (49:1, v/v). The acid fraction was evaporated under a gentle stream of nitrogen and methylated by adding ~1.5 ml of 95:5 MeOH/12 N HCl (v/v) of known δD composition and heating at 70 ºC for 12-18 h. HPLC grade water was added and fatty acid methyl esters

(FAMEs) were extracted with hexanes and eluted through Na2SO4 to remove water.

2.5. Lipid identification and quantification

n-Alkanes and FAMEs were identified by GC-MS using an Agilent 7890A GC and Agilent 5975C quadrupole mass selective detector system and quantified using a flame ionization detector (FID).

Compounds were separated on a fused silica column (Agilent J&W DB-5ms) and the oven ramped from an initial temperature of 60 ºC (held 1 min) to 320 ºC (held 15 min) at 6 ºC/min. Compounds were identified using authentic standards, fragmentation patterns and retention times. All samples were diluted in hexanes spiked with the internal standard 1,1’-binaphthyl. Compound peak areas were normalized to those of 1,1’-binaphthyl and converted to concentration using response curves for an in-house mix of n- alkanes and FAMEs at concentrations ranging from 0.5 to 100 µg/ml. Quantified concentrations were normalized to the mass of dry bud or leaf material extracted. Leaf wax concentration is reported as µg wax/g dry bud or dry leaf. Average chain length (ACL) is the weighted average concentration of all the long-chain waxes, defined as

( )( ACL = & (Eq. 1) $%& (+$ )( where a-b is the range of chain lengths and Ci is the concentration of each wax compound with i carbon atoms. ACL25-35 is used to indicate ACL values for n-C25 to n-C35 alkanes. ACL24-30 is used to indicate

ACL values for n-C24 to n-C30 alkanoic acids.

15

2.6. Lipid hydrogen isotope analysis

The δD composition of n-alkanes and n-alkanoic acids was determined using GC-isotope ratio mass spectrometry (GC-IRMS). A Thermo Trace GC Ultra was coupled to an IRMS via GC Isolink with pyrolysis reactor at 1420 ºC and Conflo IV interface. The GC oven program ramped from 80 ºC (held 2 min) to 320 ºC (held 15 min) at a rate of 8 ºC/min. For FAME analysis, samples and standards were run with the backflush valve open to exclude high-abundance compounds eluting before n-C22 alkanoic acid.

+ -1 The H3 factor was tested daily and averaged 3.8 ppm mV during the period of analysis. Standards of known δD composition (Mix A5, F8-3; A. Schimmelmann, Indiana University) were run every 6-8 samples. Data were normalized to the V-SMOW/SLAP scale by calibrating the H2 reference gas hydrogen isotope composition using Mix A5 and F8-3 standard analyses (see Polissar and D’Andrea,

2014). The long-term standard deviation of the isotope standards was 2.3‰ (1σ, n = 80). The analytical uncertainty (1-standard error of the mean, SEM) is reported along with the δDn-C29 alkane and δDn-C28 acid of each sample in Table 2. The mean analytical uncertainty was 4.2‰ (1s SEM, n = 83) for δDn-C29 alkane samples and 3.7‰ (1s SEM, n = 64) for δDn-C28 acid samples. The δD value of hydrogen added to n- alkanoic acids during derivatization was determined by mass balance of phthalic acid of known δD composition (A. Schimmelmann, Indiana University) and derivatized methyl phthalate. Hydrogen isotopic fractionations are reported as enrichment factors (in units of per mil, ‰), using the following equation:

(./012) -$%& = − 1 (Eq. 2) (./412) where a is the product and b is the substrate.

2.7. Bulk carbon isotope analysis

16 A separate aliquot of the powdered leaves was used to determine the δ13C of bulk organic carbon. Bulk

δ13C analysis was performed via continuous flow (He; 120 ml/min) on a Costech elemental analyzer (EA) interfaced with an IRMS via a Conflo IV. δ13C values were corrected for sample size dependency and normalized to the VPDB scale using a two-point calibration with IAEA calibrated (NBS-19, L-SVEC) in- house standards (-38.26‰ and -11.35‰) following Coplen et al. (2006). Error was determined by analyzing additional independent standards with a precision of 0.02‰ (1σ, n = 28) and accuracy of

−0.01‰ (n = 28). All statistical analyses were performed using JMP Pro 12.0 (SAS, Cary, NC, USA).

3. Results

3.1. Hydrogen isotope composition of environmental waters at Brown’s Lake Bog

Hydrogen isotope data for meteoric and environmental waters at BLB are presented in Figure 1 and Table

3. The δD composition of precipitation (δDp) sampled weekly to monthly at BLB from March to October

2014 ranged from -51.5‰ (DOY 89) to -9.4‰ (DOY 152) and had a mean of -31.9 ± 13.7‰ (1σ; n = 16;

Fig. 1). The mean modeled δDp from the Online Isotopes in Precipitation Calculator (OIPC; Bowen and

Revenaugh, 2003; Bowen, 2015) is -42.6‰ for the study period months (March to October) and -53‰ for the entire year. Similarly, precipitation collected monthly from 1966-1971 at a Global Network of

Isotopes in Precipitation (GNIP) station in Coshocton, OH (~50 km southwest of BLB) had a mean δDp of -37.2‰ for the study period months and -48.9‰ annually. For calculations including the δD composition of mean annual precipitation we use the OIPC value (δDMAP = -53‰) because it includes the contribution of winter precipitation, which we did not sample at BLB, and because modeled δDMAP values are widely used in leaf wax paleohydrology. Mean δD values for bog water and ephemeral surface waters at BLB over the study period were -43.3‰, and -47.1‰, respectively (Fig. 1).

17 3.2. Leaf mass per area during leaf development

Leaf area and LMA values were used to normalize wax concentration to leaf area, track the stages of leaf development, and estimate leaf photosynthetic capacity per unit area (Ellsworth and Reich, 1993;

Niinemets and Tenhunen, 1997; Niinemets, 1999). All LMA data are reported in Table 2. The timing of leaf emergence at BLB was staggered among species by up to 18 days, which is typical for temperate sites (Maycock, 1961; Lechowicz, 1984; Wesołowski and Rowiński, 2006). LMA values in all species decreased as leaves emerged and expanded, reflecting a greater rate of increase in leaf area than leaf dry mass. Minimum LMA values, indicating full leaf expansion, were observed on DOY 139 for early-leafing species (P. serotina, A. saccharinum, and U. americana) and DOY 152 for late-leafing species (Q. rubra and Q. alba). The duration of leaf expansion (from leaf emergence to minimum LMA) was approximately

11 days for U. americana, 22-24 days for A. saccharinum, Q. rubra and Q. alba, and 29 days for P. serotina. LMA minima were followed in all species by rapid increases in LMA, reflecting mass accumulation on leaves of constant area. Mean LMA values in mature leaves were lower for U. americana (48.1 g/m2) than for all other species, which ranged from 64.5 to 69.5 g/m2. These are similar to reported values for other deciduous angiosperms (e.g., Ellsworth and Reich, 1993; Niinemets and

Tenhunen, 1997; Poorter et al., 2009).

3.3.1. Leaf n-alkane abundance and chain length distribution

Adopting terminology from Piasentier et al. (2000), we define three phenologic stages of leaf wax development (indicated on Fig. 2):

(i) late bud, from first collection of buds until leaf emergence;

(ii) young leaf, from leaf emergence until full leaf expansion (as indicated by LMA); and

(iii) mature leaf, from full expansion until final sample collection.

In addition, we refer to new buds produced during the summer of 2014 as summer buds.

18

n-Alkane concentration and ACL data are reported in Table 2 and Fig. 2a. Total n-alkane concentrations

(odd chain lengths from n-C25 to n-C35) varied widely among species and phenologic stages. Total, n-C29 and n-C31 alkane concentrations generally decreased during the bud to leaf transition before increasing at the onset of leaf maturation (DOY 152) and either stabilizing or decreasing thereafter. During the late bud stage, mean total n-alkane concentrations ranged among species from 35 and 57 µg/g dry bud (Q. rubra and Q. alba, respectively) to 1053 and 1822 µg/g dry bud (P. serotina and A. saccharinum, respectively).

U. americana buds sampled on DOY 110 had a total n-alkane concentration of 420 µg/g dry bud.

Combined, n-C29 and n-C31 alkane comprised between 42% (A. saccharinum) and 90% (P. serotina and

U. americana) of total n-alkane concentrations in late buds (Fig. 2a).

During the young leaf stage, patterns of total, n-C29 and n-C31 alkane concentration changes were unique in each species. With the exception of the Quercus species, n-alkane concentrations decreased rapidly at leaf emergence (Fig. 2a). The lowest leaf n-alkane concentrations were observed during the young leaf stage and were followed by increases at the onset of the mature leaf stage (DOY 152-274) in all species except U. americana (Fig. 2a). The largest increases in total n-alkane concentration from leaf emergence to full expansion (DOY 152) were observed in Q. rubra and P. serotina (15- and 7-fold increases, respectively). Following full leaf expansion, n-alkane concentrations stabilized in A. saccharinum and gradually decreased by 26-63% through the summer and fall (DOY 152-274) in all other species. Mean total n-alkane concentrations in mature leaves ranged among species by an order of magnitude, from 132

µg/g in U. americana, 166 µg/g in Q. alba, and 578 µg/g in A. saccharinum, to 1968 µg/g in Q. rubra and

2586 µg/g in P. serotina. These are similar to previously reported n-alkane concentrations in leaves of the same genera and species (Diefendorf et al., 2011). Combined, n-C29 and n-C31 alkane comprised between

72-98% of total n-alkane concentrations in mature leaves of each species except Q. alba, in which they comprised 40%.

19 In all species, ACL25-35 values decreased during bud swell and increased during leaf expansion (Fig. 2a).

ACL25-35 values were stable in mature leaves (1σ < 0.3 in all species), consistent with a prior study that observed no significant change in ACL from summer to fall (Bush and McInerney, 2013). Mean ACL25-35 values for mature leaves (DOY 152-274) were 27.5 for Q. alba, 29.1 for U. americana, 29.2 for A. saccharinum, 29.7 for Q. rubra, and 31.0 for P. serotina.

3.3.2. Leaf n-alkanoic acid abundance and chain length distribution

n-Alkanoic acid concentration and ACL data are reported in Table 2 and Fig. 2b. Total n-alkanoic acid concentrations (even chain lengths from n-C24 to n-C30) were relatively high in late buds, decreased during leaf emergence, and increased throughout the mature leaf stage (Fig. 2b). Total n-alkanoic acid concentrations in the late buds of each species ranged from 114-279 µg/g dry bud. All species had sharp decreases in ACL24-30 during leaf emergence and expansion, reflecting complete turnover of n-C30 alkanoic acid in all species and n-C28 alkanoic acid in most species (Fig. 2b). Total, n-C28 and n-C30 alkanoic acid concentrations increased throughout the mature leaf stage and reached their highest respective concentrations in each species at the end of the growing season (DOY 274). Mean total n- alkanoic acid concentrations in mature leaves ranged among species from 141 µg/g dry leaf in P. serotina to 243 µg/g dry leaf in Q. rubra. Combined, n-C28 and n-C30 alkanoic acid comprised between 41-46% of total n-alkanoic acid concentrations in mature leaves of each species except A. saccharinum in which they comprised 63%. Minimum ACL24-30 values were observed in all species during the young leaf stage due to turnover of longer chain-lengths. Thereafter, ACL24-30 increased due to higher n-C28 and n-C30 alkanoic acid production, and stabilized between 25.9-27.4 in the mature leaves of all species and individuals.

There were several key differences in the abundance and distribution of n-alkanes and n-alkanoic acids.

First, total n-alkanoic acid concentrations and interspecies variability in mature leaves (187 ± 123 µg/g dry leaf) were both lower than observed for n-alkanes (1077 ± 1020 µg/g dry leaf; Fig. 2). Second, while

20 n-alkane concentrations in mature leaves of all species were dominated by n-C29, and n-C31 (mean ACL25-

35 = 29.3 ± 0.2), there was a less pronounced preference for any particular n-alkanoic acid chain length

(mean ACL24-30 = 26.7 ± 0.6; Fig. 2). Third, loss of longer chain lengths during leaf expansion was common to both compound classes, but complete turnover during this stage was unique to n-alkanoic acids (see Section 4.2.1). Lastly, while n-alkane concentrations were either stable or decreased through the mature leaf stage, n-alkanoic acid concentrations increased in all species over the same interval.

3.4. Plant source waters: xylem and leaf water δD

Plant water δD data are reported in Table 2 and Figure 3. Mean δDxw for all sampled species and individuals was -50.0 ± 10‰ (1σ; n = 70) over the study period and -51.8 ± 6‰ (1σ, n = 35) during the mature leaf stage. This is similar to the OIPC modeled δDMAP (-53‰), and significantly D-depleted relative to modeled δDp (-42.6 ± 10.6‰, 1σ, t-test, p = 0.048) and measured δDp (-31.9 ± 13.7‰, 1σ, t- test, p < 0.0001) for the same period (March-October). Mean δDxw values during each stage of leaf development were not significantly different for P. serotina, A. saccharinum, or U. americana (one-way

ANOVA, p = 0.2, 0.4, 0.8, respectively) but were for Q. alba and Q. rubra (one-way ANOVA, p =

0.0004, 0.001, respectively). Xylem water was significantly more D-enriched prior to leaf maturity in Q. alba and Q. rubra (-35‰ and -42‰, respectively) than during the mature leaf stage (-51‰ and -55‰, respectively).

Mean δDlw (including both buds and leaves) for all species and individuals over the study period was -

21.5 ± 21.7‰ (1σ, n = 77), which was significantly more positive and more variable than δDxw (-50.0 ±

10‰ (1σ, n = 70); t-test, p < 0.0001). The fractionation between δDlw and δDxw (εlw-xw) was significantly higher during the young leaf stage (55.1‰) than either the late bud (20.8‰) or mature leaf (28.2‰) stages (t-test, p < 0.0001), indicating D-enrichment of leaf water during the transition from bud to leaf, and D-depletion as leaves reached full expansion and maturity (Fig. 3). The mean δDlw values for the late

21 bud, young leaf and mature leaf stages (-32.4‰, 1.5‰, and -24.2‰, respectively) were significantly different (one-way ANOVA, p < 0.0001). The magnitude of leaf water D-enrichment during bud swelling and leaf emergence was 57‰ in A. saccharinum, 63‰ in P. serotina and Q. rubra, and 73‰ in Q. alba

(Fig. 3). Sampling only permitted one bud water sample for U. americana.

Single point measurements of δDlw are problematic due to observed diurnal δDlw variability (Cernusak et al., 2002; Eley et al., 2014; Cernusak et al., 2015). Therefore, we applied an expanded Craig-Gordon model described by Roden et al. (2000) to estimate the diurnal range (9:00-18:00 hrs) of δDlw using the following inputs: measured δDxw for each individual and sampling date, measured atmospheric water vapor δD sampled on DOY 116, 128, 138, 152, 205, and 275 (reported in Table 3) using the cryogenic sampling apparatus described in Helliker et al. (2002); atmospheric pressure from a weather station installed at BLB; hourly relative humidity and temperature data from the OARDC weather station; and the mean stomatal conductance of 69 mmol m-2 s-1 measured for broadleaf angiosperm trees Quercus prinus and Acer rubrum at a temperate site in Pennsylvania in August (Song et al., 2013). Modeled δDlw values were typically lowest in the morning and increased to peak values between 14:00-18:00 hrs. The diurnal range in modeled δDlw varied from 2-22‰ for each individual and date. By comparison, measured

δDlw ranged by 10-30‰ among individuals on a given date.

3.5.1. n-Alkane hydrogen isotope composition

Hydrogen isotope analysis was carried out on all n-alkane chain lengths of adequate concentration. The following discussion focuses on the δD composition of n-C29 alkane, which is frequently used in leaf wax paleohydrology and was generally the most abundant chain length in trees sampled at BLB. The δD composition of additional chain lengths is reported in Table 2. The δD values of major long-chain n- alkane homologs (n-C27, n-C29, n-C31) of all sampled individuals and species over the entire growing

22 2 2 season were strongly correlated (n-C27 and n-C29, R = 0.79, p < 0.0001; n-C29 and n-C31, R = 0.87, p <

0.0001).

In all species, δDn-C29 alkane fluctuated between DOY 89-152 and generally stabilized thereafter. The δD composition of n-C29 alkane was most negative in late buds and most positive either at or immediately following leaf emergence (Fig. 2a). Total D-enrichment during the transition from late bud to young leaf was 66‰ for P. serotina, 56‰ for Q. alba, 55‰ for Q. rubra, 46‰ for A. saccharinum, and 44‰ for U. americana, comparable to prior observations of woody broadleaf angiosperms (Tipple et al., 2013; Oakes and Hren, 2016). The timing of n-alkane D-enrichment corresponded with the greatest bud and leaf water

D-enrichment (Fig. 3). As leaves reached full expansion on DOY 152, δDn-C29 alkane values of all species became more negative, dropping by ~24-29‰ relative to peak young leaf values. After DOY 152, δDn-C29 alkane values were stable in P. serotina, A. saccharinum and Q. rubra and decreased by 13‰ and 18‰ in

U. americana and Q. alba, respectively, before stabilizing in late summer (Fig. 2a). As mature leaves approached senescence interspecies δDn-C29 alkane variability decreased. The mean δDn-C29 alkane for all species and individuals was -154.4 ± 5.4‰ (1σ, n = 11) on DOY 235 and -156.7 ± 3.8‰ (1σ, n = 9) on

DOY 274.

3.5.2. Hydrogen isotope fractionation in n-alkanes

All hydrogen isotope fractionation data are included in Table 2 and summarized by DOY and phenologic stage in Table 1. The net δD fractionation between the biosynthetic water pool and resulting leaf wax

(εbio) is not directly measurable. However, the biosynthetic water pool is fed in large part by leaf water.

Therefore, prior studies have approximated εbio values using δDwax and measured or modeled δDlw

(Kahmen et al., 2013a; Kahmen et al., 2013b; Newberry et al., 2015; Sachse et al., 2015). We approximate εbio for n-C29 alkane (εn-C29/lw) based on measured δDn-C29 alkane and measured δDlw. Because

εbio is a measure of δD fractionation during wax synthesis, we restrict εbio estimates to portions of the

23 growing season when n-alkanes are being produced (approximately from leaf emergence to full expansion). Mean εn-C29/lw estimates for this period were -147‰ for Q. rubra and P. serotina (± 11‰ and

14‰, 1σ, respectively), -138 ± 14‰ (1σ) for A. saccharinum, -135 ± 11‰ (1σ) for Q. alba, and -129 ±

10‰ (1σ) for U. americana.

Values of apparent fractionation based on δDn-C29 alkane and either δDxw (εn-C29/xw) or the δD of modeled mean annual precipitation (εn-C29/MAP) were in close agreement throughout the growing season (Table 1).

Mean εn-C29/xw values were significantly different during the late bud, young leaf and mature leaf stages

(one-way ANOVA, p < 0.0001), and interspecies variability decreased through the late summer and autumn. Mean εn-C29/MAP values for all species and individuals prior to senescence were similar on DOY

235 (-107.1 ± 6‰, 1σ, n = 11) and 274 (-109.6 ± 4‰, 1σ, n = 9). For comparison, the mean εn-C29/MAP for all previously published dicot trees sampled in temperate sites from August to November (when we expect δDwax values are stable and n-alkane synthesis has ceased), is -121.2 ± 14.9‰ (1σ, n = 67) (Sachse et al., 2006; Hou et al., 2007; Sachse et al., 2009; Tipple and Pagani, 2013).

3.6.1. n-Alkanoic acid hydrogen isotope composition

Hydrogen isotope analysis was carried out on all long-chain n-alkanoic acid homologs (n-C26, n-C28, n-

C30) where concentration permitted. Our discussion is focused on n-C28 alkanoic acid and δD values for additional n-alkanoic acid chain lengths are reported in Table 2. The δD values of long-chain n-alkanoic acid homologs (n-C26, n-C28, n-C30) of all sampled individuals and species over the entire growing season

2 2 were positively correlated (n-C26 and n-C28, R = 0.73, p < 0.0001; n-C30 and n-C28, R = 0.88, p <

0.0001).

The magnitude of D-enrichment of δDn-C28 acid during the transition from late buds to young leaves was

70‰ for A. saccharinum, 67‰ for Q. alba and 57‰ for U. americana. This is 11-24‰ greater than

24 observed for δDn-C29 alkane in the same individuals. Due to complete turnover of n-C28 alkanoic acid during the young leaf stage, δDn-C28 acid data were sparse for the young leaves of P. serotina and Q. rubra, in which the D-enrichment was 18‰ and 19‰, respectively (48‰ and 36‰ less than observed for δDn-C29 alkane in the same individuals). δDn-C28 acid values fluctuated prior to full leaf expansion and stabilized thereafter in all species except U. americana, for which δDn-C28 acid increased by 12‰ throughout the mature leaf stage (Fig. 2b). Overall, δDn-C28 acid values in mature leaves (-138.0 ± 7.5‰, 1σ, n = 34) were significantly more positive than in late buds (-182 ± 9‰, 1σ, n = 15; t-test, p < 0.0001). Prior to leaf shed

(DOY 274) mean δDn-C28 acid among all species and individuals was -136.8 ± 9.1‰ (1σ, n = 9).

3.6.2. Hydrogen isotope fractionation in n-alkanoic acids

We approximate εbio for n-C28 alkanoic acid (εn-C28/lw) based on measured δDn-C28 acid and δDlw. The majority of n-C28 alkanoic acid synthesis took place throughout the mature leaf stage following full leaf expansion. Mean εn-C28/lw estimates during the mature leaf stage were -119‰ for Q. rubra and P. serotina

(±14‰ and 15‰, 1σ, respectively), -114‰ for A. saccharinum and Q. alba (± 16‰ and 13‰, 1,

respectively), and -106 ± 23‰ (1σ) for U. americana. εn-C28/lw estimates were 21-28‰ more positive than

εn-C29/lw in the same species (Table 1).

Values of apparent fractionation based on δDn-C28 acid and either δDxw (εn-C28/xw) or modeled δDMAP (εn-

C28/MAP) were in close agreement throughout the growing season (Table 1). Mean εn-C28/xw values during the late bud stage (-139.8± 12.2‰, 1σ, n = 8) were significantly more negative than the mature leaf stage

(-90.4 ± 10.5‰, 1σ, n = 33; t-test, p < 0.0001). Toward the end of the growing season, mean εn-C28/MAP was constant from DOY 235 to 274 (-88‰) although the interspecies variability (9.6‰, 1σ) was more than twice that of εn-C29/MAP (4.1‰, 1σ). Mean εn-C28/MAP on DOY 274 was 21‰ more positive than mean

εn-C29/MAP.

25

3.7. Bulk foliar δ13C values

Bulk foliar δ13C values generally decreased from young leaves to mature leaves (Fig. 4), as observed in prior studies (Cernusak et al., 2009; McKown et al., 2013; Newberry et al., 2015). Mean bulk δ13C was -

27.1‰ for late buds, -26.6‰ for young leaves, and -28.1‰ for mature leaves. Bulk δ13C increased during the transition from late buds to young leaves, and became increasingly negative as leaves matured through the summer and autumn. On DOY 274 the bulk δ13C of summer buds was identical to that of mature leaves from the same individual (Fig. 4). The total growing season range in bulk δ13C was lowest in P. serotina (2.6‰) and highest in Q. alba (4.1‰).

4. Discussion

4.1.1. Timing of n-alkane synthesis

Leaf wax concentration reflects net wax production and removal through mechanical stresses including ablation via wind and rain (Shepherd and Griffiths, 2006). At times when rates of removal exceed production, leaf wax concentration will decrease despite continued synthesis. Therefore, we use the additional indicators of ACL (indicating production or loss of different chain lengths) and δDwax

(reflecting wax synthesis using H with different δD values) to constrain the timing of de novo leaf wax synthesis.

In all tree species at BLB, n-alkane production occurred at variable rates from the first collection of late buds through the onset of leaf maturity (DOY 152), and ceased thereafter, as indicated by stable or decreasing concentrations through the summer. This seasonal pattern is common in deciduous angiosperm

26 trees and shrubs (Jetter and Schaeffer, 2001; Sachse et al., 2009; Tipple et al., 2013; Newberry et al.,

2015; Oakes and Hren, 2016). By contrast, n-alkane production may continue through the summer and autumn in mature leaves of evergreen gymnosperms and angiosperms (Sachse et al., 2015; Oakes and

Hren, 2016). At BLB, total n-alkane concentrations generally increased in swelling buds (by 20-600%) while δDn-C29 alkane values increased (Fig. 2a). Late bud δDn-C29 alkane increases began 18-39 days prior to leaf emergence, indicating de novo n-alkane synthesis (using an increasingly D-enriched H source) as buds swell. Simultaneous decreases in ACL25-35 indicate that synthesis in buds is dominated by n-C25 through n-C29 alkane, while longer chain lengths are lost.

Leaf area expansion (decreasing LMA) during the young leaf stage coincided with decreasing n-alkane concentrations in most species (Fig. 2a). Leaf wax concentrations are typically determined based on the mass of dry leaf extracted. However, because waxes accumulate in layers over leaf surfaces, either embedded within the cuticle or as an epicuticular film (Eglinton and Hamilton, 1967; Jetter et al., 2006), dispersal of a fixed amount of wax over an increasing leaf surface could result in an apparent drop in n- alkane concentrations during leaf expansion. Converting from mass-based to leaf area-based n-alkane concentration using LMA (e.g., Wright et al., 2004) revealed a significant positive correlation between

2 2 the two measures of n-alkane abundance (µg n-C29/cm and µg n-C29/g dry leaf extracted, R = 0.90, p <

0.0001). Therefore, rapidly decreasing n-alkane concentrations in the young leaf stage following leaf emergence (e.g., P. serotina, A. saccharinum, Q. alba) may be due to dilution of pre-existing bud n- alkanes over an increasing leaf area. Baker and Hunt (1981) suggested that decreasing wax concentrations in Prunus following bud break reflect rapid leaf expansion. Despite apparent loss over this period, bud n- alkanes can be retained and contribute to the wax composition of the forming leaf. However, in certain species (e.g., Q. rubra, P. serotina) this contribution is likely swamped by relatively large amounts of n- alkanes produced de novo during leaf development.

27 We mark the beginning of de novo n-C29 alkane accumulation in leaves at the date of minimum concentrations (determined by leaf mass and area), between 7-24 days after leaf emergence (indicated with dotted vertical lines in Fig. 2, 3, 4). Rapid increases in n-alkane concentrations (total and n-C29) at the onset of the mature leaf stage (Fig. 2a) suggest an increased rate of n-alkane production as leaves reach full expansion. Similarly, Hauke and Schreiber (1998) found that major n-alkane production did not take place until 23-28 days after leaf emergence in Hedera helix (ivy) leaves.

Changes in n-alkane production and composition over the first nine weeks of the BLB growing season may be linked to shifting plant metabolism during the early stages of leaf development. Young leaves are heterotrophic and depend on stored carbohydrate reserves for synthesis of new compounds. By contrast, mature leaves are autotrophic and generally rely on photosynthesis (Turgeon, 1989; Cernusak et al., 2009;

Pantin et al., 2011). In dicotyledonous plants, the metabolic transition from heterotroph to autotroph takes place when the leaf is 30-60% fully expanded (Turgeon, 1989). As the lamina expands, rates of photosynthesis increase while rates of respiration decrease (Collier and Thibodeau, 1995; Pantin et al.,

2011). For all tree species at BLB, maximum leaf area was observed 24-35 days after leaf emergence.

Therefore, we estimate that the metabolic transition occurs ~12-17 days after leaf emergence (equivalent to 30-60% of full expansion). This timing corresponds both with peak δDn-C29 alkane values and minimum n-

C29 alkane concentrations (Fig. 2a, dotted vertical lines). Therefore, we suggest that leaves are net heterotrophs prior to the date of minimum n-C29 alkane concentration, and have developed sufficient photosynthetic capacity to operate as net autotrophs after this point. This is supported by bulk leaf δ13C data (Fig. 4), with young leaves significantly 13C-enriched relative to mature leaves (-26.6‰ and -28.1‰, respectively; t-test, p < 0.0001). Prior studies have established that heterotrophic developing leaves are

13C-enriched relative to autotrophic mature leaves (Leavitt and Long, 1985; Damesin et al., 1998;

Damesin and Lelarge, 2003; Cernusak et al., 2009). All tree species at BLB had peak δ13C values just prior to full leaf expansion (minimum LMA) and decreasing δ13C values thereafter (Fig. 4), indicating a metabolic shift just prior to leaf maturity. Phenologic regulation of leaf wax synthesis based on shifts

28 from heterotrophic to autotrophic growth appears to be a critical control on the timing of leaf wax synthesis as well as the partitioning of hydrogen sources (see Section 4.1.2.).

Results from BLB suggest that n-alkane production ceases in mature leaves, indicated by relatively constant ACL and δDn-C29 alkane and stable or decreasing n-alkane concentrations during the mature leaf stage. Total n-alkane concentrations in autumn were 26-63% lower than at full leaf expansion (for all species except A. saccharinum), perhaps due to ablation from the leaf surface (Jetter and Schaeffer,

2001). If the observed loss of n-alkanes from mature summer leaves is due to ablation, this could contribute to a regionally integrated n-alkane component in atmospheric dust or aerosols.

Summer buds sampled on DOY 274 and during the following winter from A. saccharinum and Q. rubra had total n-alkane concentrations, ACL and δDn-C29 alkane values that were distinct from those in co- occurring mature leaves and instead were similar to those observed in the late buds of the same individuals (Fig. 2a). This indicates that major production of late bud n-alkanes takes place during summer bud set and that these compounds are retained over the winter, establishing set points for leaf wax abundance and δD values in the late buds of the following growing season, as suggested previously

(Tipple et al., 2013).

4.1.2. Controls on δD of n-C29 alkane

One objective of this study was to determine when δDn-C29 alkane values of leaves are established and the factors controlling δDn-C29 alkane as leaves develop. At BLB, δDn-C29 alkane values were lowest in late buds (-

175 ± 14‰, 1σ, n = 22), highest in young leaves (-140 ± 11‰, 1σ, n = 18), and became relatively D- depleted in mature leaves (-153 ± 7‰, 1σ, n = 37). The possible drivers of distinct δDn-C29 alkane values among the phases of leaf development are changes in source water δD, changes in leaf water δD, and

29 changes in leaf metabolism. Below we demonstrate that data from BLB are consistent with the latter two factors.

Vascular plants take up precipitation-fed soil water through roots and into xylem, during which there is no isotopic fractionation (Dawson and Ehleringer, 1993; Dawson et al., 2002). Therefore, δDxw values directly reflect plant source water δD. At BLB mean δDxw across the stages of leaf development was invariable for all species except Q. alba and Q. rubra, which had 10-17‰ higher δDxw during leaf emergence and expansion than during the mature leaf stage. A similar pattern has been previously observed in both arid and humid coastal sites in the western US (Phillips and Ehleringer, 1995; Tipple et al., 2013; Sachse et al., 2015). Higher δDxw values during leaf formation may be due to retention of D- enriched winter sap (Phillips and Ehleringer, 1995; Tipple et al., 2013), and may be specific to late- leafing species (e.g., Q. rubra and Q. alba at BLB). Staggered bud break among species is controlled by the diameter and density of xylem vessels in earlywood (Lechowicz, 1984). Late-leafing species undergo xylem cavitation in winter and regeneration during leaf expansion (Turgeon, 1989). This process may cause evaporative D-enrichment of xylem water in late-leafing species and thus may contribute to D- enrichment of leaf wax prior to leaf maturity. While elevated δDxw was observed in late-leafing Q. alba and Q. rubra, the magnitude (10-17‰) was insufficient to account for δDn-C29 alkane increases of 44-66‰ over the same interval. Based on this evidence, source water δD may have a small influence on δDn-C29 alkane values in late-leafing species, but has no direct influence on other species, consistent with prior observations (Tipple et al., 2013; Newberry et al., 2015; Oakes and Hren, 2016).

The magnitude of D-enrichment in δDn-C29 alkane during the transition from late bud to young leaf (44-66‰ among species) was consistent with that observed in bud and leaf water over the same period (Fig. 3;

57‰ in A. saccharinum, 63‰ in Q. rubra, 73‰ in Q. alba, and 63‰ in P. serotina; D-enrichment was not observed in U. americana). We suspect this rapid increase in δDlw may reflect intensified transpirational D-enrichment in swelling buds and expanding leaves prior to development of the cuticle

30 that limits water loss (Riederer and Schreiber, 2001). Leaf transpiration rates typically increase during leaf emergence and peak as leaves reach 25-100% of the final leaf area (Constable and Rawson, 1980;

Hauke and Schreiber, 1998; Pantin et al., 2012). Indeed, δDlw values at BLB peaked prior to full leaf expansion (DOY 152) (Fig. 3, dotted lines). As leaves reached full expansion and n-alkane production intensified, we observed simultaneous D-depletion in δDn-C29 alkane and δDlw of 19-27‰ and 23-36‰, respectively, perhaps reflecting sufficient cuticular development to prevent further transpirational leaf water D-enrichment (Fig. 3). For the entire growing season, all species except U. americana had a

2 significant positive correlation between δDlw and δDn-C29 alkane (Fig. 3; P. serotina R = 0.53, p = 0.02; A. saccharinum R2 = 0.52, p = 0.005; Q. rubra R2 = 0.70, p = 0.0006; Q. alba R2 = 0.71, p = 0.002). When regressions are limited to the main period of de novo n-alkane synthesis (DOY 82-152), R2 values increase slightly, ranging from 0.56 to 0.78. This suggests that δDlw is the primary control on δDn-C29 alkane, as previously found for dicotyledonous angiosperms in controlled growth experiments and in natural settings (Kahmen et al., 2013b; Tipple et al., 2015).

A secondary driver of δDn-C29 alkane variability may be metabolic shifts as leaves develop, as suggested previously (Schmidt et al., 2003; Sessions, 2006; Newberry et al., 2015; Sachse et al., 2015).

Heterotrophic plant tissues have more positive δDwax values than autotrophic tissues because the former source hydrogen from relatively D-enriched NADPH derived from stored carbohydrates rather than from cellular water as during photosynthesis (Sessions, 2006; Gamarra and Kahmen, 2015). At BLB, δDn-C29 alkane values increased throughout the late bud stage, signaling that de novo n-alkane production prior to the metabolic transition may rely in part on D-enriched NADPH derived from stored reserves. The timing

13 of peak δDn-C29 alkane values coincides with the estimated metabolic transition based on peak bulk leaf δ C values (Fig. 4). Subsequently, δDn-C29 alkane values decreased by 19-27‰, perhaps in part due to the transition toward net autotrophy in mature leaves and reliance on relatively D-depleted NADPH derived from cellular water. Therefore, while δDlw controls up to 56-78% of changes in δDn-C29 alkane during the

31 period of de novo synthesis, the transition from heterotrophic to autotrophic growth may explain the remaining variability.

4.2.1. Timing of n-alkanoic acid synthesis

Full loss of long-chain waxes during the young leaf stage was unique to the n-alkanoic acids and serves as direct evidence that de novo synthesis of these chain lengths is initiated during leaf expansion. It is possible that complete turnover of n-C28 and/or n-C30 alkanoic acids reflects conversion of n-alkanoic acid precursors to n-alkanes in developing leaves (Chikaraishi and Naraoka, 2007). Increasing concentrations through the summer and autumn indicate that n-C28 alkanoic acid is synthesized continuously throughout the growing season, unlike n-C29 alkane. Therefore, continuous n-alkanoic acid synthesis may result in a seasonally-integrated record, while discrete timing of synthesis may bias n-alkane records toward the early growing season.

n-Alkane concentrations decreased through the summer and autumn (Fig 2a). If loss of n-alkanes were driven by ablation of the epicuticular film from pervasive environmental conditions such as wind or rain, we would also expect loss of n-alkanoic acids, which we did not observe. Therefore, either n-alkanoic acids are more resistant to ablation than n-alkanes, perhaps due to their different morphology or position within the cuticle (Kunst et al., 2008), or decreasing n-alkane concentrations reflect conversion in the cuticle to other wax constituents (Jetter and Schaeffer, 2001) rather than physical removal. Regardless of the specific mechanisms, this suggests that n-alkanes and n-alkanoic acids are required at different stages of the growing season and in different amounts.

4.2.2. Controls on δD of n-C28 alkanoic acid

32 For both n-alkanes and n-alkanoic acids, mean δDwax values were more positive in mature leaves than in late buds (Fig. 2). However, we note that given the same xylem and leaf water and comparable late bud

δDwax values, the net D-enrichment of n-C28 alkanoic acids from buds to mature leaves (44‰) was approximately twice that of n-C29 alkanes (22‰). We found no relationship between δDxw and δDn-C28 acid

2 (R = 0.00, p = 0.77), indicating that changes in source water were not influencing δDn-C28 acid. Based on correlations, leaf water δD may be a strong driver of n-C28 alkanoic acid δD in certain species (P. serotina, R2 = 0.78, p = 0.02; Q. alba, R2 = 0.74, p = 0.01; Q. rubra, R2 = 0.66, p = 0.01) but not in others

(A. saccharinum and U. americana, no correlation).

In mature leaves, we found a consistent ~-19‰ offset in the δD values of n-C29 and n-C27 alkanes relative to n-C30 and n-C28 alkanoic acid, respectively (Fig. 5). Prior calibration studies of mature leaves in temperate biomes found similar εalkane/acid offsets of -25‰ in Japan and Thailand (Chikaraishi and

Naraoka, 2007) and -30‰ in Massachusetts, USA (Hou et al., 2007). This offset may reflect different D- discrimination during elongation and decarboxylation of the common precursor n-alkyl acyl-ACP to produce n-alkanoic acids and n-alkanes, respectively (Chikaraishi and Naraoka, 2007). Importantly, the systematic negative εalkane/acid offset was only observed in mature leaves at BLB. Prior to full leaf expansion we observed positive εalkane/acid offsets in all species during the late bud and young leaf stages

(Fig. 5). We speculate that the sign change of εalkane/acid during the metabolic shift may reflect a change in

εbio from developing leaves that rely on carbohydrate metabolism to mature leaves that rely on current photosynthate. While observations from BLB strengthen the evidence for a systematic εalkane/acid offset in temperate settings with relatively low species diversity and a single leaf flush, evidence from locations with exceptionally high species diversity including a tropical forest (Feakins et al., 2016) and botanical garden (Gao et al., 2014) found large interspecies variability in the sign and values of εalkane/acid. This suggests that fractionation between compound classes may depend on species diversity and associated heterogeneity in leaf lifespan and the timing of leaf (and leaf wax) development.

33 4.3. Implications for interpreting δDwax from sediments

Sediments integrate leaf waxes at the regional scale via particulate waxes in aerosols and at the ecosystem scale via direct leaf fall and erosion of soils (Diefendorf and Freimuth, 2017). Here we use observations from BLB to assess the composition and variability of the δDwax signal integrated via direct leaf fall from the tree growth form only. Assuming direct transfer of wax from tree leaf to sediment is a simplification.

However, this allows us to estimate the minimum sensitivity of the plant level δDwax signal exclusive of the additional noise likely incorporated at the sediment level by the comprehensive set of wax sources and transport processes.

4.3.1. Interspecies δDwax variability within a growth form

When multiple plant growth forms are present, trees may be a dominant source of long-chain leaf waxes to sediments (Sachse et al., 2012). At BLB, trees have greater spatial coverage and leaf biomass production than other growth forms including grasses, herbs and shrubs. Therefore, we use all tree species to approximate the ecosystem-wide δDwax signal at BLB but acknowledge major differences among species and growth forms with respect to leaf wax ACL and abundance (Diefendorf et al., 2011) and εapp

(Sachse et al., 2012) that likely influence wax signals integrated across diverse ecosystems. Relatively low δDwax variability among species just prior to leaf shed (DOY 274) for δDn-C29 alkane (-157 ± 3‰, 1σ) and δDn-C28 acid (-137 ± 7‰, 1σ) suggests that the community-integrated δDwax signal will not be weighted toward those species with the highest wax concentration, leaf biomass production or abundance within the forest. This may not be the case for sites with greater interspecies δDwax variability. A comparison with

δDn-C29 alkane reported from other deciduous angiosperm tree species sampled at temperate sites from

August through November indicates that the interspecies δDn-C29 alkane variability at BLB from August-

October (1σ = 5‰, n = 5 species) is less than half that observed at other sites (Fig. 6). For instance, variability for 9 tree species at Blood Pond, Massachusetts was 11‰ (Hou et al., 2007). When including

34 all growth forms at the same site, interspecies variability rose to 65‰. Even within a species, variability can be relatively high, for example 15‰ among 3 biological replicates of the woody shrub Corylus americana sampled in October in Connecticut (Oakes and Hren, 2016). Biological δDwax variability at a single site and even within a single growth form (trees) potentially limits the sensitivity of ecosystem- integrated δDwax records to changes in δDp that are greater in magnitude than interspecies variability (i.e.,

5‰ at BLB, 11‰ at Blood Pond, Massachusetts). However, we note that the full range of δDwax variability among growth forms and species within a site may increase these thresholds considerably. This complicates selection of accurate εapp values for interpretation of sedimentary δDwax and demonstrates the need to understand how biological noise is integrated at the sediment level.

4.3.2. n-Alkanes and n-alkanoic acids record different source water information

We observed several key differences among compound classes that may support the use of either n- alkanes or n-alkanoic acids for paleohydrology applications. First, we found that n-alkane concentrations in mature leaves were 1.5-20 times higher than n-alkanoic acid concentrations in the same species, except for U. americana where n-alkane concentrations were ~20% lower than n-alkanoic acids. Therefore, we anticipate that n-alkanes will be several times more abundant in the ecosystem-integrated wax than n- alkanoic acids. Based on higher interspecies concentration differences for n-alkanes (1σ = 1020 µg/g dry leaf) than for n-alkanoic acids (1σ = 123 µg/g dry leaf), there is higher potential for weighting of the ecosystem-integrated signal toward species with highest n-alkane concentrations.

Second, εn-C29/xw and εn-C28/xw values stabilized early in the mature leaf stage (by DOY 177 and 152, respectively), after which n-C29 alkane synthesis ceased and n-C28 synthesis continued (Fig. 2, Table 1).

Therefore, at BLB and other sites where source water δD is stable through the growing season (i.e., δDxw

= δDMAP; Fig. 1), εapp is likely set for both compound classes at the onset of leaf maturity, regardless of wax production or loss in mature leaves. However, for conditions in which δDxw changes in response to

35 short term δDp variations, continued synthesis of n-alkanoic acids through the growing season could potentially bias δDn-C28 acid records toward summer and autumn source water δD. These conditions could include sites with abundant and reliable summer rainfall (Ehleringer and Dawson, 1992), plants with shallow rooting depths or with DBH < ~12 cm (Dawson and Ehleringer, 1991; Phillips and Ehleringer,

1995), coastal plants influenced by changing source water salinity, trees with limited access to groundwater (White et al., 1985), riparian plants using ephemeral stream water, or annuals and herbaceous plants more likely to use summer precipitation than woody plants (Ehleringer et al., 1991).

Third, we observed that toward the end of the mature leaf stage (DOY 235-274) interspecies δDn-C28 acid variability was ~2-3 times that of δDn-C29 alkane (Table 1). Higher δD variability in n-alkanoic acids relative to n-alkanes has been observed previously (Chikaraishi and Naraoka, 2007; Hou et al., 2007; Feakins et al., 2016) and suggests greater sensitivity of sedimentary δDn-C29 alkane records than δDn-C28 acid records.

Finally, we found that εapp values in mature leaves were consistently ~19‰ more positive for n-alkanoic acids than for n-alkanes (Fig. 5), similar to prior observations in temperate settings (Chikaraishi and

Naraoka, 2007; Hou et al., 2007). The majority of calibrations of εapp in modern vegetation have been developed for n-alkanes and therefore may not be appropriate to apply to reconstructions of past δDp using n-alkanoic acids in sediments. A more comprehensive calibration dataset may be needed for robust estimates of εapp values specific to n-alkanoic acids.

4.3.3. Estimating εapp values and uncertainty

εapp remains the largest source of uncertainty in leaf wax paleohydrology (Sachse et al., 2012; Polissar and

D’Andrea, 2014). Even within the same growth form and biome there can be considerable site-to-site variation in εapp (Fig. 6). Thus, applying εapp values determined at other sites to interpret sedimentary

δDwax may introduce additional uncertainty into δDp reconstructions. Further, site-specific εapp calibrations based on modern vegetation may have limited accuracy for δDwax interpretation from older sediments,

36 especially during past ecological transitions or periods when ancient plant communities were significantly different from the modern.

Hydrogen isotope data from BLB provide an opportunity to quantify the propagated analytical uncertainty in leaf wax and source water δD values and estimate the total uncertainty in resulting εapp at the growth for (tree) level. We estimate that the error propagated uncertainty in our measured εapp values is 6.7‰

(1σ) for εn-C29/MAP and 6.9‰ (1σ) for εn-C28/MAP. This was calculated using a Monte Carlo simulation with

10,000 iterations (e.g., Anderson, 1976) of the weighted mean 1σ for the measured δDwax of all species and individuals during the mature leaf stage, as well as the 95% confidence interval (4‰) reported for the modeled δDMAP value at BLB (OIPC, Bowen, 2015; Bowen and Revenaugh, 2003). The same simulation using the weighted mean 1σ of measured δDxw as the source water input estimates a slightly higher uncertainty for εn-C29/xw (8.7‰) and εn-C28/xw (8.8‰). Therefore, the total uncertainty in εapp estimates based on all species within a single growth form (trees) at BLB is approximately 9‰. While this is considerably lower than other conservative estimates of εapp uncertainty (~25‰; Polissar and D’Andrea, 2014), we note that εapp uncertainty would likely increase with inclusion of multiple growth forms at the sediment level.

Therefore we consider 9‰ to be a minimum expected uncertainty in ecosystem-integrated εapp values at

BLB, which may be similar for other temperate sites. Additional factors including the transport and integration of leaf waxes in sediments and ecological changes through time likely further affect the uncertainty of εapp estimates applied to sedimentary records.

5. Conclusions

We provide a record of seasonal changes in the δD composition of n-alkanes and n-alkanoic acids and plant source waters for five tree species growing in a temperate deciduous forest. This dataset allows us to approximate the δDwax signal integrated across the tree growth form, which may have a strong influence

37 on sedimentary δDwax records. Results indicate that long chain n-alkanes and n-alkanoic acids produced in the same plant differ in key ways including concentration, timing of synthesis and δD composition.

Specifically, in four out of five tree species, n-alkane concentrations in mature leaves were 2 to 29 times higher than n-alkanoic acid concentrations in the same individuals. Further, while the majority of de novo n-C29 alkane synthesis occurs during a discrete interval as leaves reach full expansion (24-35 days after

13 leaf emergence), n-C28 alkanoic acids are produced throughout the growing season. Bulk leaf δ C suggests that phenologic development controls the timing and amount of n-alkanes produced. During leaf emergence and expansion, exposure of leaf tissues with undeveloped cuticle appears to drive a strong transpirational D-enrichment of leaf water which, in turn, acts as a primary control on δDwax values as leaves and the cuticle reach maturity. Despite differences in the timing of n-alkane and n-alkanoic acid synthesis, δDwax values for both compound classes are primarily determined during the period of leaf emergence and expansion. This suggests that the δDwax signal from trees in temperate settings may be biased toward the local timing of spring leaf emergence. A secondary control on δDwax may be a shift in plant metabolism and the δD composition of H used for wax synthesis in developing and mature leaves.

Prior to autumn leaf fall, interspecies δDn-C29 alkane variability (1σ = 4‰) was lower than for δDn-C28 acid (1 σ

= 9‰), suggesting that n-alkanes may be more sensitive tracers of changes in source water δD. During the mature leaf stage we observed a consistent ~-19‰ offset of εn-C29/xw values relative to εn-C28/xw, underscoring that the two compound classes fractionate shared source water to different extents and should be interpreted from sedimentary records accordingly. Overall, these data suggest that the distinct abundance, timing of synthesis and εapp values of n-alkanes and n-alkanoic acids merit consideration when applying either as paleohydrology proxies.

38 Acknowledgments

We thank Michael Hren, two anonymous reviewers and Associate Editor Alex Sessions for their helpful suggestions to improve this manuscript. Thanks go to Dr. Gregory Wiles, Nicholas Wiesenberg and

Douglas Sberna for sample collection assistance and to The Nature Conservancy for site access. This research was supported by the National Science Foundation (EAR-1229114 to AFD). Acknowledgment is made to the Donors of the American Chemical Society Petroleum Research Fund for partial support of this research (PRF #51787-DNI2 to AFD). Fieldwork was supported by a grant from the American

Philosophical Society to EJF. This research was also supported by awards from the Geological Society of

America and Sigma-Xi to EJF.

39 REFERENCES

Baker, E. and Hunt, G.M. (1981) Developmental changes in leaf epicuticular waxes in relation to foliar penetration. New Phytol. 88, 731-747. Bush, R.T. and McInerney, F.A. (2013) Leaf wax n-alkane distributions in and across modern plants: Implications for paleoecology and chemotaxonomy. Geochim. Cosmochim. Acta 117, 161-179. Cernusak, L., Pate, J.S. and Farquhar, G. (2002) Diurnal variation in the stable isotope composition of water and dry matter in fruiting Lupinus angustifolius under field conditions. Plant, Cell Environ. 25, 893-907. Cernusak, L.A., Tcherkez, G., Keitel, C., Cornwell, W.K., Santiago, L.S., Knohl, A., Barbour, M.M., Williams, D.G., Reich, P.B., Ellsworth, D.S., Dawson, T.E., Griffiths, H.G., Farquhar, G.D. and Wright, I.J. (2009) Why are non-photosynthetic tissues generally 13C enriched compared with leaves in C3 plants? Review and synthesis of current hypotheses. Funct. Plant Biol. 36, 199-213. Cernusak, L.A., Barbour, M.M., Arndt, S.K., Cheesman, A.W., English, N.B., Feild, T.S., Helliker, B.R., HollowayPhillips, M.M., Holtum, J.A. and Kahmen, A. (2015) Stable isotopes in leaf water of terrestrial plants. Plant, Cell Environ. Chikaraishi, Y. and Naraoka, H. (2007) δ13C and δD relationships among three n-alkyl compound classes (n-alkanoic acid, n-alkane and n-alkanol) of terrestrial higher plants. Org. Geochem. 38, 198-215. Collier, D.E. and Thibodeau, B.A. (1995) Changes in respiration and chemical content during autumnal senescence of Populus tremuloides and Quercus rubra leaves. Tree Phys 15, 759-764. Constable, G. and Rawson, H. (1980) Effect of leaf position, expansion and age on photosynthesis, transpiration and water use efficiency of cotton. Func. Plant Bio. 7, 89-100. Coplen, T.B., Brand, W.A., Gehre, M., Gröning, M., Meijer, H.A.J., Toman, B. and Verkouteren, R.M. (2006) New guidelines for δ13C measurements. Analytical Chemistry 78, 2439-2441. Cornelissen, J.H.C., Lavorel, S., Garnier, E., #237, az, S., Buchmann, N., Gurvich, D.E., Reich, P.B., Steege, H.t., Morgan, H.D., Heijden, M.G.A.v.d., Pausas, J.G. and Poorter, H. (2003) A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Aust. J. Bot. 51, 335-380. Craig, H. (1961) Isotopic Variations in Meteoric Waters. Science 133, 1702-1703. Damesin, C., Rambal, S. and Joffre, R. (1998) Seasonal and annual changes in leaf 13C in two co occurring Mediterranean oaks: relations to leaf growth and drought progression. Func. Ecol. 12, 778-785. Damesin, C. and Lelarge, C. (2003) Carbon isotope composition of currentyear shoots from Fagus sylvatica in relation to growth, respiration and use of reserves. Plant, Cell & Environment 26, 207-219. Dawson, T.E. and Ehleringer, J.R. (1991) Streamside trees that do not use stream water. Nature 350, 335- 337. Dawson, T.E. and Ehleringer, J.R. (1993) Isotopic enrichment of water in the “woody” tissues of plants: implications for plant water source, water uptake, and other studies which use the stable isotopic composition of cellulose. Geochim. Cosmochim. Acta 57, 3487-3492. Dawson, T.E., Mambelli, S., Plamboeck, A.G., Templer, P.H. and Tu, K.P. (2002) Stable isotopes in plant ecology. Annu. Rev. Ecol. Syst. 33, 507-559. Diefendorf, A.F., Freeman, K.H., Wing, S.L. and Graham, H.V. (2011) Production of n-alkyl lipids in living plants and implications for the geologic past. Geochim. Cosmochim. Acta 75, 7472-7485. Diefendorf, A.F. and Freimuth, E.J. (2017) Extracting the most from terrestrial plant-derived n-alkyl lipids and their carbon isotopes from the sedimentary record: A review. Org. Geochem. 103, 1- 21. Eglinton, G. and Hamilton, R.J. (1967) Leaf epicuticular waxes. Science 156, 1322-1335.

40 Ehleringer, J. and Dawson, T. (1992) Water uptake by plants: perspectives from stable isotope composition. Plant, Cell Environ. 15, 1073-1082. Ehleringer, J.R., Phillips, S.L., Schuster, W.S. and Sandquist, D.R. (1991) Differential utilization of summer rains by desert plants. Oecologia 88, 430-434. Eley, Y., Dawson, L., Black, S., Andrews, J. and Pedentchouk, N. (2014) Understanding 2H/1H systematics of leaf wax n-alkanes in coastal plants at Stiffkey saltmarsh, Norfolk, UK. Geochim. Cosmochim. Acta 128, 13-28. Ellsworth, D. and Reich, P. (1993) Canopy structure and vertical patterns of photosynthesis and related leaf traits in a deciduous forest. Oecologia 96, 169-178. Feakins, S.J. and Sessions, A.L. (2010) Controls on the D/H ratios of plant leaf waxes from an arid ecosystem. Geochim. Cosmochim. Acta 74, 2128-2141. Feakins, S.J., Bentley, L.P., Salinas, N., Shenkin, A., Blonder, B., Goldsmith, G.R., Ponton, C., Arvin, L.J., Wu, M.S. and Peters, T. (2016) Plant leaf wax biomarkers capture gradients in hydrogen isotopes of precipitation from the Andes and Amazon. Geochim. Cosmochim. Acta 182, 155-172. Gamarra, B. and Kahmen, A. (2015) Concentrations and δ2H values of cuticular n-alkanes vary significantly among plant organs, species and habitats in grasses from an alpine and a temperate European grassland. Oecologia 178, 981-998. Gao, L., Edwards, E.J., Zeng, Y. and Huang, Y. (2014) Major Evolutionary Trends in Hydrogen Isotope Fractionation of Vascular Plant Leaf Waxes. PLoS ONE 9, e112610. Graham, H.V., Patzkowsky, M.E., Wing, S.L., Parker, G.G., Fogel, M.L. and Freeman, K.H. (2014) Isotopic characteristics of canopies in simulated leaf assemblages. Geochim. Cosmochim. Acta 144, 82-95. Hauke, V. and Schreiber, L. (1998) Ontogenetic and seasonal development of wax composition and cuticular transpiration of ivy (Hedera helix L.) sun and shade leaves. Planta 207, 67-75. Helliker, B.R., Roden, J.S., Cook, C. and Ehleringer, J.R. (2002) A rapid and precise method for sampling and determining the oxygen isotope ratio of atmospheric water vapor. Rapid Communications in Mass Spectrometry 16, 929-932. Hou, J., D'Andrea, W.J. and Huang, Y. (2008) Can sedimentary leaf waxes record D/H ratios of continental precipitation? Field, model, and experimental assessments. Geochim. Cosmochim. Acta 72, 3503-3517. Hou, J.Z., D'Andrea, W.J., MacDonald, D. and Huang, Y.S. (2007) Hydrogen isotopic variability in leaf waxes among terrestrial and aquatic plants around Blood Pond, Massachusetts (USA). Org. Geochem. 38, 977-984. Jetter, R. and Schaeffer, S. (2001) Chemical composition of the Prunus laurocerasus leaf surface. Dynamic changes of the epicuticular wax film during leaf development. Plant Physiol. 126, 1725- 1737. Jetter, R., Kunst, L. and Samuels, A.L. (2006) Composition of plant cuticular waxes, in: Rieder, M., Muller, C. (Eds.), Biology of the plant cuticle. Blackwell Publishing, Oxford, pp. 145-181. Kahmen, A., Hoffmann, B., Schefuß, E., Arndt, S.K., Cernusak, L.A., West, J.B. and Sachse, D. (2013a) Leaf water deuterium enrichment shapes leaf wax n-alkane δD values of angiosperm plants II: Observational evidence and global implications. Geochim. Cosmochim. Acta 111, 50-63. Kahmen, A., Schefuß, E. and Sachse, D. (2013b) Leaf water deuterium enrichment shapes leaf wax n- alkane δD values of angiosperm plants I: Experimental evidence and mechanistic insights. Geochim. Cosmochim. Acta 111, 39-49. Kaplan, J.O., Prentice, I.C. and Buchmann, N. (2002) The stable carbon isotope composition of the terrestrial biosphere: Modeling at scales from the leaf to the globe. Global Biogeochemical Cycles 16, 1060. Kunst, L., Jetter, R. and Samuels, A.L. (2008) Biosynthesis and transport of plant cuticular waxes. Annual Plant Reviews 23, 182-215. Leavitt, S.W. and Long, A. (1985) Stable-carbon isotopic composition of maple sap and foliage. Plant Physiol. 78, 427-429.

41 Lechowicz, M.J. (1984) Why do temperate deciduous trees leaf out at different times? Adaptation and ecology of forest communities. The American Naturalist 124, 821-842. Maycock, P.F. (1961) Botanical studies on Mont St. Hilaire, Rouville County, Quebec. Canadian Journal of Botany 39, 1293-1325. McInerney, F.A., Helliker, B.R. and Freeman, K.H. (2011) Hydrogen isotope ratios of leaf wax n-alkanes in grasses are insensitive to transpiration. Geochim. Cosmochim. Acta 75, 541-554. McKown, A.D., Guy, R.D., Azam, M.S., Drewes, E.C. and Quamme, L.K. (2013) Seasonality and phenology alter functional leaf traits. Oecologia 172, 653-665. Newberry, S.L., Kahmen, A., Dennis, P. and Grant, A. (2015) n-Alkane biosynthetic hydrogen isotope fractionation is not constant throughout the growing season in the riparian tree Salix viminalis. Geochim. Cosmochim. Acta 165, 75-85. Niinemets, Ü. and Tenhunen, J. (1997) A model separating leaf structural and physiological effects on carbon gain along light gradients for the shadetolerant species Acer saccharum. Plant, Cell Environ. 20, 845-866. Niinemets, Ü. (1999) Research review. Components of leaf dry mass per area–thickness and density–alter leaf photosynthetic capacity in reverse directions in woody plants. New Phytol. 144, 35-47. Oakes, A.M. and Hren, M.T. (2016) Temporal variations in the δD of leaf n-alkanes from four riparian plant species. Org. Geochem. 97, 122-130. Pagani, M., Pedentchouk, N., Huber, M., Sluijs, A., Schouten, S., Brinkhuis, H., Damsté, J.S.S., Dickens, G.R., Backman, J. and Clemens, S. (2006) Arctic hydrology during global warming at the Palaeocene/Eocene thermal maximum. Nature 442, 671-675. Pantin, F., Simonneau, T., Rolland, G., Dauzat, M. and Muller, B. (2011) Control of leaf expansion: a developmental switch from metabolics to hydraulics. Plant Physiol. 156, 803-815. Pantin, F., Simonneau, T. and Muller, B. (2012) Coming of leaf age: control of growth by hydraulics and metabolics during leaf ontogeny. New Phytol. 196, 349-366. Phillips, S.L. and Ehleringer, J.R. (1995) Limited uptake of summer precipitation by bigtooth maple (Acer grandidentatum Nutt) and Gambel's oak (Quereus gambelii Nutt). Trees 9, 214-219. Piasentier, E., Bovolenta, S. and Malossini, F. (2000) The n-alkane concentrations in buds and leaves of browsed broadleaf trees. The Journal of Agricultural Science 135, 311-320. Polissar, P.J. and Freeman, K.H. (2010) Effects of aridity and vegetation on plant-wax δD in modern lake sediments. Geochim. Cosmochim. Acta 74, 5785-5797. Polissar, P.J. and D’Andrea, W.J. (2014) Uncertainty in paleohydrologic reconstructions from molecular δD values. Geochim. Cosmochim. Acta 129, 146-156. Poorter, H., Niinemets, U., Poorter, L., Wright, I.J. and Villar, R. (2009) Causes and consequences of variation in leaf mass per area (LMA): a meta-analysis. New Phytol. 182, 565-588. Rach, O., Brauer, A., Wilkes, H. and Sachse, D. (2014) Delayed hydrological response to Greenland cooling at the onset of the Younger Dryas in western Europe. Nature Geoscience 7, 109-112. Riederer, M. and Schreiber, L. (2001) Protecting against water loss: analysis of the barrier properties of plant cuticles. J. Exp. Bot. 52, 2023-2032. Roden, J.S., Lin, G. and Ehleringer, J.R. (2000) A mechanistic model for interpretation of hydrogen and oxygen isotope ratios in tree-ring cellulose. Geochim. Cosmochim. Acta 64, 21-35. Sachse, D., Radke, J. and Gleixner, G. (2004) Hydrogen isotope ratios of recent lacustrine sedimentary n- alkanes record modern climate variability. Geochim. Cosmochim. Acta 68, 4877-4889. Sachse, D., Radke, J. and Gleixner, G. (2006) δD values of individual n-alkanes from terrestrial plants along a climatic gradient – Implications for the sedimentary biomarker record. Org. Geochem. 37, 469-483. Sachse, D., Kahmen, A. and Gleixner, G. (2009) Significant seasonal variation in the hydrogen isotopic composition of leaf-wax lipids for two deciduous tree ecosystems (Fagus sylvativa and Acer pseudoplatanus). Org. Geochem. 40, 732-742.

42 Sachse, D., Gleixner, G., Wilkes, H. and Kahmen, A. (2010) Leaf wax n-alkane δD values of field-grown barley reflect leaf water δD values at the time of leaf formation. Geochim. Cosmochim. Acta 74, 6741-6750. Sachse, D., Billault, I., Bowen, G.J., Chikaraishi, Y., Dawson, T.E., Feakins, S.J., Freeman, K.H., Magill, C.R., McInerney, F.A., van der Meer, M.T.J., Polissar, P., Robins, R.J., Sachs, J.P., Schmidt, H.- L., Sessions, A.L., White, J.W.C., West, J.B. and Kahmen, A. (2012) Molecular paleohydrology: Interpreting the hydrogen-isotopic composition of lipid biomarkers from photosynthesizing organisms. Annual Review of Earth and Planetary Sciences 40, 221-249. Sachse, D., Dawson, T.E. and Kahmen, A. (2015) Seasonal variation of leaf wax n-alkane production and δ2H values from the evergreen oak tree, Quercus agrifolia. Isotopes Environ. Health Stud. 51, 124-142. Schmidt, H.-L., Werner, R.A. and Eisenreich, W. (2003) Systematics of 2H patterns in natural compounds and its importance for the elucidation of biosynthetic pathways. Phytochem. Rev. 2, 61-85. Schwab, V.F., Garcin, Y., Sachse, D., Todou, G., Séné, O., Onana, J.-M., Achoundong, G. and Gleixner, 13 G. (2015) Effect of aridity on δ C and δD values of C3 plant- and C4 graminoid-derived leaf wax lipids from soils along an environmental gradient in Cameroon (Western Central Africa). Org. Geochem. 78, 99-109. Seki, O., Nakatsuka, T., Shibata, H. and Kawamura, K. (2010) A compound-specific n-alkane δ13C and δD approach for assessing source and delivery processes of terrestrial organic matter within a forested watershed in northern Japan. Geochim. Cosmochim. Acta 74, 599-613. Sessions, A.L., Burgoyne, T.W., Schimmelmann, A. and Hayes, J.M. (1999) Fractionation of hydrogen isotopes in lipid biosynthesis. Org. Geochem. 30, 1193-1200. Sessions, A.L. (2006) Seasonal changes in D/H fractionation accompanying lipid biosynthesis in Spartina alterniflora. Geochim. Cosmochim. Acta 70, 2153-2162. Shepherd, T. and Griffiths, D.W. (2006) The effects of stress on plant cuticular waxes. New Phytol. 171, 469-499. Smith, F.A. and Freeman, K.H. (2006) Influence of physiology and climate on δD of leaf wax n-alkanes from C3 and C4 grasses. Geochim. Cosmochim. Acta 70, 1172-1187. Song, X., Barbour, M.M., Farquhar, G.D., Vann, D.R. and Helliker, B.R. (2013) Transpiration rate relates to withinand acrossspecies variations in effective path length in a leaf water model of oxygen isotope enrichment. Plant, Cell Environ. 36, 1338-1351. Tierney, J.E., Russell, J.M., Huang, Y., Damsté, J.S.S., Hopmans, E.C. and Cohen, A.S. (2008) Northern hemisphere controls on tropical southeast African climate during the past 60,000 years. Science 322, 252-255. Tipple, B.J., Berke, M.A., Doman, C.E., Khachaturyan, S. and Ehleringer, J.R. (2013) Leaf-wax n- alkanes record the plant–water environment at leaf flush. Proceedings of the National Academy of Sciences 110, 2659-2664. Tipple, B.J. and Pagani, M. (2013) Environmental control on eastern broadleaf forest species’ leaf wax distributions and D/H ratios. Geochim. Cosmochim. Acta 111, 64-77. Tipple, B.J., Berke, M.A., Hambach, B., Roden, J.S. and Ehleringer, J.R. (2015) Predicting leaf wax n- alkane 2H/1H ratios: controlled water source and humidity experiments with hydroponically grown trees confirm predictions of Craig–Gordon model. Plant, Cell Environ. 38, 1035-1047. Turgeon, R. (1989) The sink-source transition in leaves. Annual review of plant biology 40, 119-138. Wesołowski, T. and Rowiński, P. (2006) Timing of bud burst and tree-leaf development in a multispecies temperate forest. For Ecol Manag 237, 387-393. West, A.G., Patrickson, S.J. and Ehleringer, J.R. (2006) Water extraction times for plant and soil materials used in stable isotope analysis. Rapid Communications in Mass Spectrometry 20, 1317- 1321. White, J.W., Cook, E.R. and Lawrence, J.R. (1985) The DH ratios of sap in trees: Implications for water sources and tree ring DH ratios. Geochim. Cosmochim. Acta 49, 237-246.

43 Wright, I.J., Reich, P.B., Westoby, M., Ackerly, D.D., Baruch, Z., Bongers, F., Cavender-Bares, J., Chapin, T., Cornelissen, J.H.C., Diemer, M., Flexas, J., Garnier, E., Groom, P.K., Gulias, J., Hikosaka, K., Lamont, B.B., Lee, T., Lee, W., Lusk, C., Midgley, J.J., Navas, M.-L., Niinemets, U., Oleksyn, J., Osada, N., Poorter, H., Poot, P., Prior, L., Pyankov, V.I., Roumet, C., Thomas, S.C., Tjoelker, M.G., Veneklaas, E.J. and Villar, R. (2004) The worldwide leaf economics spectrum. Nature 428, 821-827.

44

Figure 1. Seasonal changes in environmental waters measured at BLB (closed symbols) throughout the study period (March to October). Open symbols show monthly δDp modeled for BLB (OIPC) and measured at the GNIP station in Coshocton, OH from 1966-71. Modeled δDMAP for BLB is represented with horizontal black line. Phenologic stages prior to and during leaf maturity for trees at BLB are indicated along DOY axis.

45

46 Figure 2. Seasonal changes in lipid concentration (shaded layers), ACL (thin black line) and δD values

(thick black lines and points) for n-alkanes (A) and n-alkanoic acids (B) from a single individual from each species. Black squares show δDwax values from biological replicates. The stages of leaf development are initialed LB (late bud), YL (young leaf, shaded), and ML (mature leaf). Vertical lines mark the DOY of minimum n-C29 alkane concentration (based on leaf mass and area, solid lines), minimum LMA (long dashes), or both (short dashes). Open symbols indicate data for buds; closed symbols indicate leaves. Summer buds are indicated with open symbols on DOY 274. Right panels labeled “SB” for A. saccharinum and Q. rubra represent lipid concentration (stacked bars), ACL

(triangles) and δD data (open circles) from buds collected the following winter. Note different scales for leaf wax concentration and ACL. Analytical uncertainty for δDwax measurements is 4‰.

47

Figure 3. Seasonal changes in the δD composition

of xylem water (gray squares), leaf water (black

squares), n-alkanes (blue circles) and n-alkanoic

acids (red circles) in buds (open symbols) and

leaves (closed symbols). Correlations of δDlw with

n-alkane δD and n-alkanoic acid δD are given in

blue and red text, respectively. The young leaf

stage is indicated with gray shading. Summer buds

of A. saccharinum and Q. rubra are indicated with

open symbols on DOY 274. Data for a single

individual from each species.

48

Figure 4. Seasonal changes in δDn-C29 alkane (blue

circles) and bulk foliar δ13C (black diamonds)

observed in buds (open symbols) and leaves

(closed symbols) of one individual from each

species. Summer buds sampled the following

winter from A. saccharinum and Q. rubra are in

panels labeled SB. The young leaf stage is

indicated with gray shading.

49

Figure 5. Fractionation between corresponding n-alkanes and n-alkanoic acids, εn-C27 alkane/n-C28 acid

(squares) and εn-C29 alkane/n-C30 acid (circles). Open symbols represent late buds (blue) and summer buds

(red). Closed symbols represent young leaves (orange) and mature leaves (green). For illustrative purposes, smoothing splines (lambda = 0.05) mark general seasonal trends for εn-C27 alkane/n-C28 acid (dotted line) and εn-C29 alkane/n-C30 acid (solid line). Dotted vertical lines represent the earliest and latest dates for minimum LMA across all species. Right panel labeled SB includes data from A. saccharinum and Q. rubra buds collected the following winter.

50

Figure 6. Comparison of εn-C29/MAP observed in this study with comparable published data from deciduous angiosperm trees and woody shrubs sampled in temperate biomes during the mature leaf stage, August to November. Panels include data from single locations (closed symbols); and from transect studies (open symbols). Data are numbered by study as follows: 1) Sachse et al., 2009 (Acer pseudoplatanus only); 2) Hou et al., 2007; 3) Oakes and Hren, 2016; 4) this study; 5) Sachse et al.,

2006; 6) Tipple and Pagani, 2013. The OIPC modeled δDMAP for the sites(s) of sample collection are listed above each boxplot. Mean εn-C29/MAP values ± 1σ, and the number of individuals and species (in parentheses) appear below each boxplot.

51

Table 1. Mean fractionation of all individuals analyzed for n-C29 alkane and n-C28 alkanoic acid by DOY (top panel, summer buds Table 1. excluded)Mean fractionation and phonologic of all individualsstage (bottom analyzed panel). for n-C29 alkane and n-C28 alkanoic acid by DOY (top, summer buds excluded) and phenologic stage (bottom).

a b c a b c εn-C29/xw (‰) εn-C29/MAP (‰) εn-C29/lw (‰) εn-C28/xw (‰) εn-C28/MAP (‰) εn-C28/lw (‰) Date DOY Mean (1σ) n Mean (1σ) n Mean (1σ) n Mean (1σ) n Mean (1σ) n Mean (1σ) n 3/23/14 82 -- -142.7 (6) 4 -136.5 (4) 2 -- -142.9 (8) 3 -134.3 (4) 2 3/30/14 89 -- -141.1 (6) 4 -137.3 (8) 4 -- -138.8 (7) 4 -135.0 (8) 4 4/5/14 95 -128.7 (7) 4 -128.2 (3) 4 -133.7 (10) 4 -127.7 (2) 2 -131.3 (15) 2 -143.9 (17) 2 4/20/14 110 -131.7 (19) 9 -119.3 (16) 9 -162.5 (10) 9 -135.5 (27) 6 -123.6 (25) 6 -167.6 (20) 6 4/27/14 117 -102.1 (25) 5 -96.9 (16) 5 -129.7 (16) 5 -112.5 (32) 2 -120.5 (14) 2 -154.5 (9) 2 5/8/14 128 -94.5 (16) 10 -95.6 (10) 9 -145.7 (16) 10 -85.3 (18) 4 -79.2 (17) 4 -133.9 (18) 5 5/19/14 139 -100.3 (10) 4 -95.7 (12) 4 -134.4 (13) 4 -86.3 (20) 3 -83.5 (16) 3 -119.5 (22) 3 6/1/14 152 -104.0 (10) 10 -101.3 (9) 11 -140.6 (10) 11 -96.6 (9) 9 -93.0 (7) 10 -131.7 (7) 10 6/26/14 177 -108.4 (8) 4 -104.8 (8) 4 -125.6 (10)d 4 -92.1 (4) 3 -89.4 (1) 3 -109.2 (8) 3 7/25/14 206 -108.2 (5) 2 -109.6 (10) 2 -145.6 (--) d 1 -92.4 (5) 2 -93.8 (1) 2 -123.8 (--) 1 8/23/14 235 -107.9 (7) 11 -107.1 (6) 11 -125.2 (10) d 11 -87.8 (11) 10 -87.2 (9) 10 -105.8 (12) 10 10/1/14 274 -107.2 (9) 8 -109.6 (4) 9 -126.0 (17) d 9 -87.9 (12) 8 -88.4 (10) 9 -105.3 (17) 9 Late bud -132.3 (12) 14 -129.2 (15) 22 -145.9 (18) 20 -139.8 (12) 8 -136.0 (10) 15 -151.9 (21) 14 Young leaf -94.7 (14) 18 -94.5 (10) 17 -141.4 (16) 18 -86.1 (15) 9 -84.0 (17) 9 -131.2 (18) 10 Mature leaf -106.7 (8) 35 -105.9 (8) 37 -130.7 (14) d 36 -90.4 (11) 33 -89.8 (8) 34 -114.4 (16) 33 Summer bud -135.6 (11) 4 -144.1 (4) 4 -141.1 (19) 4 -133.7 (12) 4 -142.3 (5) 4 -139.3 (16) 4

a b c d εapp based on measured δDxw; εapp based on modeled δDMAP; εbio estimated using measured δDlw; εbio estimates for the period when n-alkane synthesis ceased (DOY 177-274) are shown for completeness but note εbio values are only relevant for periods when wax synthesis is taking place.

52 Table 2. Leaf wax abundance, δD and εapp values (‰ VSMOW) for every individual and sampling date. Material is listed by growth stage (LB,

Late Bud; YL, Young Leaf; ML, Mature Leaf; SB, Summer Bud) and sample type (B, Bud; L, Leaf). Also included are bulk leaf δ13C, leaf mass

per area (LMA), leaf wax abundance by leaf area (µg/cm2) and plant water δD for xylem (δDxw) and leaves (δDlw).

ACL n-acid (µg/g) n-alkane (µg/g) n-acid δD n-alkane δD εalkane-acid εwax/lw εwax/xw εwax/MAP conc. (µg/cm2) Sampling Bulk Date n-C24-30 n-C25-35 n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- leaf/bud LMA Individual Site Species DOY (M/D/YR) Stage Type acid alkane C26 C28 C30 C25 C27 n-C29 n-C31 C33 C35 C26 C28 C30 C25 C27 C29 C31 C33 C29/30 C27/28 n-C29 n-C28 n-C29 n-C28 n-C29 n-C28 δ13C (g/m2) n-C29 n-C28 δDlw δDxw ε(lw/xw) 1BC 1 P. serotina 82 3/23/14 LB B 27.2 30.2 6 4 6 0 11 233 278 25 0 ------196 -196 -179 ------151 ------1SM 1 A. saccharinum 82 3/23/14 LB B 27.7 30.5 59 140 117 130 222 320 471 611 59 -172 -182 -185 -159 -170 -184 -191 -188 1.4 13.7 -139 -137 - - -138 -136 -29.2 - - - -51.7 - - 2RO 2 Q. rubra 82 3/23/14 LB B 25.7 29.6 22 17 31 0 0 17 7 0 0 -185 -187 -197 - -167 -189 -186 - 9.7 23.6 -134 -132 - - -143 -141 -29.2 - - - -63.4 - - 2WO 2 Q. alba 82 3/23/14 LB B 27.1 29.2 6 7 12 0 4 16 7 0 0 - -197 -177 - - -184 -178 - -9.0 ------139 -152 -28.2 ------1BC 1 P. serotina 89 3/30/14 LB B 26.5 30.2 50 27 36 15 27 483 622 90 0 -162 -175 -186 - - -180 -188 - 7.0 - -135 -130 - - -134 -129 -27.4 - - - -51.8 - - 1SM 1 A. saccharinum 89 3/30/14 LB B 27.6 30.4 63 139 117 141 206 317 396 548 56 -175 -186 -188 -156 -170 -186 -191 -189 1.4 19.1 -143 -142 - - -141 -140 -27.8 - - - -50.9 - - 2RO 2 Q. rubra 89 3/30/14 LB B 25.8 32.0 22 16 35 0 4 18 8 0 26 -188 -190 -201 -176 - -193 -192 - 9.7 - -144 -141 - - -148 -144 -26.9 - - - -57.0 - - 2WO 2 Q. alba 89 3/30/14 LB B 27.0 28.5 38 46 79 0 5 16 0 0 0 -180 -187 -195 -187 -190 -187 -182 - 9.7 -4.1 -127 -127 - - -141 -141 -26.0 - - - -68.9 - - 1BC 1 P. serotina 95 4/5/14 LB B 26.7 30.3 87 46 58 0 32 543 690 107 0 - -168 -176 - - -172 -182 - 4.8 - -137 -132 -131 -126 -126 -121 -25.8 - - - -41.1 -47.5 6.6 1SM 1 A. saccharinum 95 4/5/14 LB B 27.1 30.0 32 57 40 212 260 297 357 492 55 -179 -187 -197 -150 -160 -178 -180 -184 24.8 33.7 -146 -156 -119 -129 -131 -142 -29.5 - - - -37.0 -66.3 31.4 2WO 2 Q. alba 95 4/6/14 LB B 26.5 29.0 17 19 24 0 0 14 0 0 0 ------171 - - - - -121 - -131 - -125 - -27.4 - - - -56.6 -46.5 -10.6 2RO 2 Q. rubra 95 4/6/14 LB B 24.0 29.6 0 0 0 0 0 16 7 0 0 ------176 - - - - -131 - -134 - -130 - -25.6 - - - -52.8 -49.2 -3.9

1BC 1 P. serotina 110 4/20/14 YL L 24.7 30.1 16 0 0 22 14 166 201 54 0 ------140 -143 - - - -150 - -101 - -92 - -26.1 - - - 11.5 -44.0 58.1 1SM 1 A. saccharinum 110 4/20/14 LB B 26.7 29.6 36 54 40 251 438 431 504 515 0 -169 -190 -189 -153 -163 -171 -163 -171 22.0 33.6 -162 -181 -133 -152 -125 -145 -27.8 - - - -10.7 -44.8 35.7 2RO 2 Q. rubra 110 4/20/14 LB B 26.0 28.5 21 19 27 0 7 21 0 0 0 - -167 - - - -172 - - - - -157 -152 -144 -139 -125 -121 -26.3 - - - -17.3 -32.8 16.0 2SM 2 A. saccharinum 110 4/20/14 LB B 26.5 29.4 25 32 27 271 470 211 487 388 26 -137 -172 -185 -150 -155 -169 -142 -151 18.7 20.8 -173 -175 -135 -138 -123 -126 -25.6 - - - 3.7 -39.7 45.2 2WO 2 Q. alba 110 4/20/14 LB B 26.8 27.6 17 20 27 12 10 26 0 0 0 - - -181 -138 -153 -170 - - 13.5 - -154 - -141 - -124 - -25.9 - - - -19.4 -33.5 14.6 3AE 3 U. americana 110 4/21/14 LB B 25.8 28.9 25 13 25 7 33 350 30 0 0 -178 -175 -178 - - -169 - - 10.1 - -163 -168 -130 -136 -123 -128 -25.7 - - - -7.7 -44.8 38.8 3BC 3 P. serotina 110 4/21/14 YL L 25.5 29.6 30 14 16 17 7 73 73 11 0 - -126 -156 -123 - -146 -148 - 12.9 - -155 -136 -105 -85 -98 -78 -26.3 - - - 11.6 -45.6 59.9 3RO 3 Q. rubra 110 4/21/14 LB B 25.9 28.6 15 14 31 6 8 37 9 0 0 - - - -179 -167 -191 -182 - - - -182 - -160 - -146 - -27.9 - - - -11.1 -37.7 27.6 3WO 3 Q. alba 110 4/21/14 LB B 27.1 28.1 54 61 132 13 15 31 10 0 0 - -189 -200 -146 -152 -164 -177 - 45.0 45.6 -166 -191 -137 -163 -117 -144 -27.0 - - - 2.6 -31.4 35.1 2RO 2 Q. rubra 117 4/27/14 LB B 25.8 28.4 24 20 24 0 15 30 0 0 0 - - -164 - -122 -140 - - 28.9 - -119 - -110 - -92 - -26.3 - - - -23.0 -33.7 11.0 2WO 2 Q. alba 117 4/27/14 LB B 25.6 26.7 18 24 0 77 44 54 0 0 0 - - -154 -128 -139 -144 - - 11.7 - -130 - -119 - -96 - -25.8 - - - -16.5 -28.4 12.2 3RO 3 Q. rubra 117 4/27/14 LB B 25.7 28.5 18 15 26 7 11 39 10 0 0 -164 -176 -184 -162 -144 -170 -167 - 16.8 38.9 -155 -161 -129 -135 -124 -130 -27.0 - - - -18.2 -47.6 30.9 1BC 1 P. serotina 117 4/28/14 YL L 24.0 30.1 0 0 0 0 0 18 21 0 0 -117 - - - - -130 -129 - - - -115 - -83 - -81 - -27.7 43.7 0.08 - -17.1 -51.5 36.2 1SM 1 A. saccharinum 117 4/28/14 YL L 24.6 30.4 13 15 0 77 119 110 513 263 0 -106 -158 -170 -120 -150 -140 -130 -128 35.7 8.9 -130 -148 -71 -90 -92 -111 -27.3 95.4 1.05 - -11.6 -74.8 68.3 3BC 3 P. serotina 117 4/27/14 - L ------50.2 - - - - - 2SM 2 A. saccharinum 117 4/27/14 - L ------94.3 - - - - - 2RO 2 Q. rubra 128 5/8/14 YL L 24.4 28.5 17 0 0 8 35 67 21 0 0 - - - -116 -129 -138 -132 - - - -138 - -99 - -90 - -26.2 - - - 0.1 -43.5 45.5 2SM 2 A. saccharinum 128 5/8/14 YL L 24.7 31.1 13 5 0 13 10 16 182 84 0 ------112 -117 ------25.0 59.3 0.10 0.03 12.6 -48.3 64.0 2WO 2 Q. alba 128 5/8/14 YL L 25.7 27.8 16 33 0 71 49 114 31 0 0 - -120 - -132 -139 -133 - - - -21.9 -137 -124 -102 -88 -84 -70 -27.4 94.8 1.08 0.31 4.8 -34.5 40.7 3AE 3 U. americana 128 5/8/14 YL L 25.6 28.1 16 7 12 38 67 173 16 0 0 - -117 - -103 -108 -126 -134 - - 9.5 -118 -109 -84 -75 -77 -67 -27.3 58.5 1.01 0.04 -8.8 -45.2 38.1 3BC 3 P. serotina 128 5/8/14 YL L 25.8 29.7 9 5 5 4 0 19 26 0 0 - - - -114 - -152 -141 -135 - - -158 - -90 - -105 - -26.4 54.4 0.11 0.03 7.0 -68.0 80.5 3RO 3 Q. rubra 128 5/8/14 YL L 25.7 28.4 9 8 11 11 34 71 16 0 0 - -152 -169 -144 -131 -153 -155 - 19.2 24.0 -157 -155 -111 -109 -106 -104 -26.9 76.6 0.55 0.06 4.2 -47.9 54.7 3WO 3 Q. alba 128 5/8/14 YL L 24.6 27.6 4 0 0 176 173 292 54 0 0 - - - -139 -141 -143 -141 - - - -151 - -110 - -95 - -24.7 - - - 10.1 -36.5 48.4 1AE 1 U. americana 128 5/9/14 YL L 26.2 28.9 10 8 6 0 20 74 13 0 0 - -124 - - -91 -112 -105 - - 37.9 -125 -137 -56 -69 - -75 -29.1 38.3 0.28 0.03 15.0 -59.3 79.0 1BC 1 P. serotina 128 5/9/14 YL L 25.9 30.7 39 21 21 4 4 82 188 48 0 ------149 -154 -130 - - -157 - -97 - -102 - -27.6 40.0 0.33 0.09 9.4 -58.3 71.9 1SM 1 A. saccharinum 128 5/9/14 YL L 25.8 30.3 27 23 22 43 70 79 271 138 0 - -140 - -122 -140 -147 -138 -137 - - -151 - -101 - -99 - -26.0 62.0 0.49 0.14 5.6 -50.5 59.1 2BC 2 P. serotina 128 5/9/14 YL L 25.6 30.7 17 7 6 0 2 25 53 12 0 ------150 -144 -135 - - -164 - -94 - -103 - -27.3 43.0 0.11 0.03 16.3 -62.2 83.7 3AE 3 U. americana 138 5/18/14 - L ------44.9 - - - - - 3BC 3 P. serotina 138 5/18/14 - L ------59.8 - - - - - 3WO 3 Q. alba 138 5/18/14 - L ------72.3 - - - - - 1BC 1 P. serotina 139 5/19/14 YL L 26.0 30.9 50 28 29 8 12 190 481 177 0 - -149 -148 - - -158 -162 -153 -10.9 - -147 -139 -108 -100 -110 -102 -28.2 39.6 0.75 0.11 -12.4 -55.1 45.3 1SM 1 A. saccharinum 139 5/19/14 YL L 25.6 30.2 33 49 0 16 25 77 144 49 0 -116 -120 - - - -142 -144 -139 - - -119 -96 -87 -63 -94 -71 -28.0 51.5 0.40 0.25 -26.6 -60.9 36.6 2RO 2 Q. rubra 139 5/19/14 YL L 24.9 28.5 19 17 0 6 38 76 15 0 0 - - - - -126 -144 - - - - -144 - -106 - -96 - -26.1 53.8 0.41 0.09 0.0 -42.8 44.7 2WO 2 Q. alba 139 5/19/14 YL L 25.9 28.8 19 27 39 10 13 51 24 0 0 - -127 - - - -131 - - - - -128 -124 -101 -96 -82 -78 -24.2 55.6 0.28 0.15 -3.6 -33.9 31.3 1AE 1 U. americana 139 5/19/14 - L ------32.8 - - - - - 2BC 2 P. serotina 139 5/19/14 - L ------50.6 - - - - -

53 Table 2. (continued)

ACL n-acid (µg/g) n-alkane (µg/g) n-acid δD n-alkane δD εalkane-acid εwax/lw εwax/xw εwax/MAP conc. (µg/cm2) Sampling Bulk Date n-C24-30 n-C25-35 n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- leaf/bud LMA Individual Site Species DOY (M/D/YR) Stage Type acid alkane C26 C28 C30 C25 C27 n-C29 n-C31 C33 C35 C26 C28 C30 C25 C27 C29 C31 C33 C29/30 C27/28 n-C29 n-C28 n-C29 n-C28 n-C29 n-C28 δ13C (g/m2) n-C29 n-C28 δDlw δDxw ε(lw/xw) 2SM 2 A. saccharinum 139 5/19/14 - L ------56.0 - - - - - 2BC 2 P. serotina 152 6/1/14 ML L 24.9 30.9 14 0 0 6 15 572 1489 545 0 ------144 -145 -138 - - -144 - -94 - -96 - -27.5 62.5 3.57 - 0.6 -55.1 58.9 2RO 2 Q. rubra 152 6/1/14 ML L 25.4 29.6 18 20 0 51 181 1106 905 58 0 -144 -148 -148 - - -166 -166 - -21.5 - -161 -143 -119 -100 -120 -100 -27.2 52.6 5.82 0.11 -6.5 -53.5 49.6 3AE 3 U. americana 152 6/1/14 ML L 25.3 29.3 23 16 0 7 13 113 59 0 0 -152 -138 - - - -140 -127 - - - -136 -134 -99 -97 -92 -90 -26.4 55.1 0.62 0.09 -4.3 -45.9 43.6 3BC 3 P. serotina 152 6/1/14 ML L 25.7 30.7 21 17 0 7 19 448 953 266 0 -148 -126 -124 - - -140 -142 -134 -17.3 - -136 -122 -87 -73 -91 -77 -27.4 65.2 2.92 0.11 -4.4 -57.2 56.0 3RO 3 Q. rubra 152 6/1/14 ML L 26.5 29.7 23 27 23 20 142 989 749 55 0 -140 -147 -140 - - -154 -153 - -16.3 - -137 -130 -104 -96 -107 -100 -27.6 56.6 5.59 0.15 -20.1 -56.3 38.4 3WO 3 Q. alba 152 6/1/14 ML L 25.7 27.0 25 22 0 43 34 26 7 0 0 -164 -145 -145 -139 -149 -147 - - -2.3 -4.0 -139 -138 -102 -101 -99 -98 -26.9 57.2 0.15 0.13 -9.0 -50.0 43.1 1AE 1 U. americana 152 6/2/14 ML L 25.7 29.2 26 24 0 0 20 266 62 0 0 -129 -138 -114 - - -144 - - -33.6 - -139 -132 -111 -104 -96 -89 -29.7 41.1 1.09 0.10 -5.9 -37.0 32.3 1BC 1 P. serotina 152 6/2/14 ML L 26.9 31.0 20 20 23 8 21 515 1359 546 0 - -142 -132 - - -153 -158 -151 -24.5 - -145 -134 -110 -99 -106 -94 -27.5 52.0 2.68 0.10 -9.3 -47.8 40.5 1SM 1 A. saccharinum 152 6/2/14 ML L 26.3 29.3 21 38 0 16 77 206 142 21 0 -129 -141 -139 - -146 -154 -149 - -17.4 -5.4 -137 -124 -115 -102 -107 -93 -28.1 65.5 1.35 0.25 -19.3 -43.8 25.6 2SM 2 A. saccharinum 152 6/2/14 ML L 27.3 29.7 19 39 24 10 29 77 95 17 0 -135 -146 -146 - - -158 -152 - -13.8 - -150 -138 -85 -72 -111 -98 -28.2 62.9 0.48 0.25 -9.2 -80.2 77.2 2WO 2 Q. alba 152 6/2/14 ML L 26.5 27.2 27 28 21 83 71 66 23 0 0 -138 -137 - -136 -139 -138 - - - -2.3 -122 -120 -98 -97 -90 -89 -26.8 49.5 0.33 0.14 -19.2 -44.6 26.6 1BC 1 P. serotina 177 6/26/14 ML L 26.8 31.0 29 23 29 6 20 610 1624 615 0 ------152 -158 -151 - - -129 - -111 - -105 - -28.4 69.4 4.23 0.16 -26.2 -46.0 20.8 1SM 1 A. saccharinum 177 6/26/14 ML L 27.1 29.1 27 52 30 19 113 313 156 18 0 -131 -138 -133 - -142 -159 -150 - -30.1 -4.4 -122 -101 -110 -88 -112 -90 -28.5 82.4 2.58 0.42 -41.6 -55.0 14.2 2RO 2 Q. rubra 177 6/26/14 ML L 26.4 29.8 42 49 38 42 173 1067 1022 74 0 -137 -136 -146 - -149 -156 -153 - -12.1 -14.4 -136 -116 -116 -95 -109 -88 -26.8 97.1 10.36 0.48 -22.9 -45.6 23.8 2WO 2 Q. alba 177 6/26/14 ML L 27.0 27.7 45 53 29 33 54 42 22 0 0 -137 -138 -136 -137 -144 -142 -140 - -6.9 -6.6 -114 -110 -97 -93 -94 -90 -27.0 87.0 0.36 0.46 -31.0 -49.7 19.6 1AE 1 U. americana 177 6/26/14 - L ------56.7 - - - - - 2SM 2 A. saccharinum 177 6/26/14 - L ------78.5 - - - - - 3AE 3 U. americana 178 6/27/14 - L ------54.6 - - - - - 3BC 3 P. serotina 178 6/27/14 - L ------80.3 - - - - - 3RO 3 Q. rubra 178 6/27/14 - L ------101.3 - - - - - 3WO 3 Q. alba 178 6/27/14 - L ------113.5 - - - - - 1BC 1 P. serotina 205 7/24/14 ML L 26.9 31.0 36 29 53 8 21 782 2159 774 0 - -141 -140 - - -150 -155 -146 -11.5 - ##### ##### -105 -96 -102 -93 -27.0 84.7 6.62 0.25 - -50.0 1SM 1 A. saccharinum 205 7/25/14 ML L 26.9 29.3 41 74 40 13 96 265 177 18 0 -138 -142 -136 - -158 -164 -158 - -31.9 -17.8 -146 -124 -111 -89 -117 -94 -29.7 67.3 1.78 0.50 -21.1 -58.9 40.2 3AE 3 U. americana 205 7/24/14 - L ------78.5 - - - - - 3BC 3 P. serotina 205 7/24/14 - L ------91.5 - - - - - 3RO 3 Q. rubra 205 7/24/14 - L ------99.8 - - - - - 3WO 3 Q. alba 205 7/24/14 - L ------105.1 - - - - - 1AE 1 U. americana 205 7/25/14 - L ------56.1 - - - - - 2SM 2 A. saccharinum 205 7/25/14 - L ------70.4 - - - - - 2BC 2 P. serotina 205 7/25/14 - L ------82.9 - - - - - 2RO 2 Q. rubra 205 7/25/14 - L ------90.7 - - - - - 1BC 1 P. serotina 235 8/24/14 ML L 26.6 31.1 30 22 33 2 6 337 1192 503 0 - -136 - - - -157 -158 -152 - - -138 -116 -119 -96 -110 -87 -27.5 - - - -22.4 -43.5 22.0 1SM 1 A. saccharinum 235 8/24/14 ML L 27.0 29.0 36 70 36 23 139 330 154 17 0 -129 -128 -131 - -151 -159 -153 - -32.2 -26.8 -135 -103 -106 -73 -112 -79 -28.4 - - - -27.4 -59.0 33.6 2BC 2 P. serotina 235 8/23/14 ML L 26.9 31.2 25 20 33 0 0 269 943 399 0 -144 - -136 - - -154 -149 -141 -21.0 - -124 - -109 - -107 - -27.9 - - - -34.1 -50.2 17.0 2RO 2 Q. rubra 235 8/23/14 ML L 26.6 29.9 61 82 57 27 147 732 875 80 0 -147 -150 -152 - - -160 -160 - -9.5 - -136 -126 -116 -105 -113 -103 -28.6 - - - -27.9 -50.5 23.8 2SM 2 A. saccharinum 235 8/23/14 ML L 27.2 29.1 34 72 48 18 112 239 130 19 0 -134 -141 -143 - -149 -163 -152 - -23.3 -9.6 -136 -112 -119 -95 -117 -93 -27.9 - - - -31.9 -50.8 19.8 2WO 2 Q. alba 235 8/23/14 ML L 27.0 27.6 42 44 25 34 65 42 19 0 0 - -132 - -148 -152 -156 -152 - - -23.6 -129 -105 -108 -82 -109 -83 -27.0 - - - -30.5 -54.0 24.8 3AE 3 U. americana 235 8/23/14 ML L 26.6 29.0 55 35 58 0 0 95 0 0 0 -130 -127 -131 - - -153 -143 - -25.1 - -115 -88 -104 -76 -105 -78 -27.5 - - - -42.2 -54.7 13.3 3RO 3 Q. rubra 235 8/23/14 ML L 27.0 29.0 40 59 55 0 0 549 - - 0 -139 -140 -149 - - -150 -144 - -1.3 - -113 -102 -97 -87 -102 -92 -27.9 - - - -42.0 -58.3 17.3 1AE 1 U. americana 235 8/24/14 ML L 26.9 29.0 44 33 58 0 0 98 0 0 0 -131 -126 -134 - - -152 - - -21.0 - -122 -94 -105 -77 -105 -77 -30.4 - - - -35.0 -52.7 18.7 3BC 3 P. serotina 235 8/24/14 ML L 27.0 30.9 38 27 57 0 0 354 928 278 0 - -132 - - - -145 -149 -139 - - -109 -95 -102 -88 -97 -83 -29.1 - - - -40.8 -47.8 7.4 3WO 3 Q. alba 235 8/24/14 ML L 26.8 27.9 77 62 32 57 114 100 48 0 0 -141 -145 -143 -143 -151 -148 -142 - -6.4 -6.9 -121 -117 -102 -98 -101 -97 -28.4 - - - -31.5 -51.5 21.1 1SM 1 A. saccharinum 274 10/1/14 SB B 27.2 31.2 67 120 57 18 78 136 235 355 42 -182 -189 -188 - -194 -195 -191 -185 -9.3 -6.3 -154 -147 -149 -142 -150 -143 -28.6 - - - -49.1 -54.4 5.6 2RO 2 Q. rubra 274 10/1/14 SB B 25.6 29.3 30 20 40 1 3 22 11 0 0 -186 -191 -205 - - -186 -182 - 23.3 - -138 -143 -138 -143 -141 -146 -26.7 - - - -55.5 -55.9 0.4 3AE 3 U. americana 274 10/1/14 ML L 26.8 29.2 57 34 69 0 6 88 16 0 0 -129 -126 -130 - - -153 -137 - -26.2 - -107 -79 -94 -65 -106 -77 -28.1 - - - -51.6 -64.8 14.1 3RO 3 Q. rubra 274 10/1/14 ML L 27.0 30.0 40 63 56 19 161 1030 1256 123 0 -138 -143 -147 - - -155 -152 - -8.8 - -118 -105 -98 -85 -107 -95 -29.0 - - - -41.8 -63.0 22.6 3WO 3 Q. alba 274 10/1/14 ML L 27.3 27.6 26 29 15 52 75 65 24 0 0 -148 -149 -146 -157 -155 -153 -160 - -8.1 -6.8 -114 -110 -100 -96 -105 -101 -28.9 - - - -43.6 -58.3 15.6 1AE 1 U. americana 274 10/2/14 ML L 27.5 29.1 72 55 143 0 7 117 16 0 0 - -121 - - - -154 - - - - -140 -106 - - -107 -72 -30.4 - - - -17.1 - - 1BC 1 P. serotina 274 10/2/14 ML L 26.7 31.0 41 36 50 0 0 348 1081 392 0 -157 -141 -141 - - -156 -158 -154 -17.4 - -144 -129 -119 -103 -109 -93 -28.1 - - - -13.5 -42.1 29.8 1SM 1 A. saccharinum 274 10/2/14 ML L 27.1 28.9 88 187 95 23 135 277 127 16 0 -138 -136 -135 - -154 -163 -154 - -32.2 -20.9 -152 -125 -115 -86 -116 -88 -28.8 - - - -13.1 -54.4 43.7 2RO 2 Q. rubra 274 10/2/14 ML L 26.4 29.9 75 90 65 15 85 557 622 52 0 -140 -135 -143 - -146 -158 -158 - -18.2 -12.8 -136 -112 -108 -83 -111 -86 -27.1 - - - -25.7 -55.9 32.0 2SM 2 A. saccharinum 274 10/1/14 ML L 27.8 28.9 93 293 189 23 137 255 109 21 0 -143 -146 -148 - -163 -163 -150 - -17.7 -19.8 -106 -88 -116 -98 -116 -98 -29.1 - - - -63.8 -53.0 -11.4 2WO 2 Q. alba 274 10/2/14 ML L 26.9 27.5 66 79 42 25 42 30 12 0 0 -136 -134 -136 -147 -155 -156 -147 - -23.6 -23.3 -117 -95 -108 -85 -109 -86 -27.9 - - - -43.9 -54.4 11.1 1SM 1 A. saccharinum 390 1/25/15 SB B 27.5 30.4 63 126 90 81 247 408 373 546 61 -179 -190 -191 - -183 -190 -194 -189 2.0 9.4 -116 -117 -131 -132 -144 -145 -28.4 - - - -83.6 -67.4 -17.4 2RO 2 Q. rubra 390 1/25/15 SB B 25.7 29.1 26 19 36 1 4 22 8 0 0 -180 -181 -192 - -171 -187 -182 - 6.0 12.6 -157 -151 -124 -118 -141 -135 -27.3 - - - -35.8 -71.6 38.6

54 Table 3. Environmental water δD values collected at BLB in 2014.

Date Sample type DOY (M/D/YR) δD water Accumulated Precipitation 89 3/30/14 -51.5 Accumulated Precipitation 95 4/5/14 -48.6 Accumulated Precipitation 95 4/5/14 -46.9 Accumulated Precipitation 118 4/28/14 -16.4 Accumulated Precipitation 118 4/28/14 -18.6 Accumulated Precipitation 128 5/8/14 -36.0 Accumulated Precipitation 138 5/18/14 -35.3 Accumulated Precipitation 138 5/18/14 -20.0 Accumulated Precipitation 152 6/1/14 -12.6 Accumulated Precipitation 152 6/1/14 -9.4 Accumulated Precipitation 177 6/26/14 -26.2 Accumulated Precipitation 177 6/26/14 -29.1 Accumulated Precipitation 235 8/23/14 -48.5 Accumulated Precipitation 235 8/23/14 -42.4 Accumulated Precipitation 275 10/2/14 -34.5 Accumulated Precipitation 275 10/2/14 -34.1 Atm Water Vapor 117 4/27/14 -92.0 Atm Water Vapor 128 5/8/14 -63.8 Atm Water Vapor 138 5/18/14 -113.5 Atm Water Vapor 152 6/1/14 -86.2 Atm Water Vapor 205 7/24/14 -137.8 Atm Water Vapor 275 10/2/14 -89.8 Bog Water 82 3/23/14 -72.5 Bog Water 89 3/30/14 -62.7 Bog Water 96 4/6/14 -60.5 Bog Water 110 4/20/14 -47.4 Bog Water 118 4/28/14 -51.9 Bog Water 128 5/8/14 -45.8 Bog Water 139 5/19/14 -25.8 Bog Water 153 6/2/14 -28.2 Bog Water 177 6/26/14 -30.1 Bog Water 206 7/25/14 -25.8 Bog Water 236 8/24/14 -35.8 Bog Water 275 10/2/14 -32.5 Ephemeral Surface Water 82 3/23/14 -58.0 Ephemeral Surface Water 89 3/30/14 -68.2 Ephemeral Surface Water 95 4/5/14 -62.5 Ephemeral Surface Water 110 4/20/14 -54.8 Ephemeral Surface Water 118 4/28/14 -55.7 Ephemeral Surface Water 128 5/8/14 -46.3 Ephemeral Surface Water 129 5/9/14 -44.0 Ephemeral Surface Water 139 5/19/14 -40.6 Ephemeral Surface Water 139 5/19/14 -50.5 Ephemeral Surface Water 139 5/19/14 -49.5 Ephemeral Surface Water 177 6/26/14 -29.3 Ephemeral Surface Water 177 6/26/14 -39.2 Ephemeral Surface Water 178 6/27/14 -31.5 Ephemeral Surface Water 235 8/23/14 -37.9 Ephemeral Surface Water 236 8/24/14 -39.2 Lake Water 89 3/30/14 -44.5 Lake Water 95 4/5/14 -59.9 Lake Water 129 5/9/14 -42.5 Lake Water 129 5/9/14 -39.2 Rain Event 118 4/28/14 -39.6 Rain Event 129 5/9/14 7.7 Snow Event 89 3/30/14 -160.0

55 Chapter 3

Sedimentary plant waxes in a temperate bog are biased toward woody vegetation2

ABSTRACT

Leaf waxes preserved in sediments have emerged as an important tool for investigating past hydrologic change. Such reconstructions require estimates of the δD offset between precipitation and leaf wax, known as apparent fractionation (εapp). However, εapp values are controlled by numerous factors and remain a large source of uncertainty in leaf wax paleohydrology. Prior studies have observed a wide range of εapp values in modern plants, both among growth forms and species. This complicates selection of accurate εapp values for sedimentary leaf waxes, which accumulate from a mixture of plant sources at a single site. Further, the transport of leaf waxes from source (plants) to sink (sediments) is poorly understood and may further influence sedimentary wax records.

This project addresses these uncertainties by comparing the abundance, molecular distribution and isotopic composition (δD and δ13C) of n-alkanes and n-alkanoic acids in bog sediments with all major plant species growing in the catchment of Brown’s Lake Bog (BLB), Ohio, USA. There are two distinct plant assemblages at BLB, including a forest dominated by trees and a bog shoreline dominated by shrubs, woody groundcover, herbs and grasses. A large range of δDwax values among individual plants

(77‰ for n-C29 alkane and 84‰ for n-C28 alkanoic acid) was driven by differences in biosynthetic δD fractionation as well as source water differences between forest and shoreline plants. This range in δDwax was much smaller in the upper 40 cm of bog sediments (9‰ and 11‰, respectively), suggesting that

2 Freimuth, E.J., Diefendorf, A.F., Lowell, T.V., Wiles, G. Sedimentary plant waxes in a temperate bog are biased toward woody vegetation. Organic Geochemistry (in prep.)

56 sediments are either biased toward specific plants, or that there is consistent pre- or post-deposition signal averaging. Tree species had n-alkane concentrations between 10 and 300 times higher than shoreline species, while n-alkanoic acid concentrations were generally lower and comparable among all species and growth forms. Comparing the δD and δ13C fingerprint of all plant sources as well as rain-scavenged aerosol waxes suggests that sediments are biased toward trees and other woody species (including shrubs and woody groundcover) growing around the basin. This suggests that sedimentary plant wax records can be strongly influenced by catchment vegetation. We observed a ~30‰ offset in εapp values between n- alkanes (-133‰) and n-alkanoic acids (-103‰), both at the plant-level and in sediments. These insights into sediment bias toward woody vegetation identify a critical area for further investigation and provide a framework for more accurate interpretations of lake sediment leaf waxes.

57 1. Introduction

Plant waxes are preserved on geologic timescales in diverse sedimentary archives including paleosols, rocks and lacustrine and marine sediments. These organic compounds have become valuable proxies for past environmental conditions, owing to their widespread spatial coverage and excellent preservation potential. Plant waxes are resistant to alteration below ~150 °C (Yang and Huang, 2003; Sessions et al.,

2004; Schimmelmann et al., 2006; Diefendorf et al., 2015). Consequently, they can retain their original molecular and isotopic composition from the time of synthesis more or less in perpetuity. This study focuses on n-alkanes and n-alkanoic acids, two classes of n-alkyl lipids that contribute to the diverse mixture of compounds that comprise plant epicuticular waxes (Eglinton and Hamilton, 1967). While n- alkanes are more widely studied in modern systems, both n-alkanes and n-alkanoic acids are commonly used in studies of past environments.

Different plant groups produce waxes with characteristic molecular distributions that can be used as indicators of plant growth environments (e.g., terrestrial or aquatic) or growth forms (e.g., tree, shrub, grass, moss) that were dominant in the past (Eglinton and Hamilton, 1967; Ficken et al., 2000;

Rommerskirchen et al., 2006; Aichner et al., 2010; Bush and McInerney, 2013; Freeman and Pancost,

2014). Additional paleoenvironmental information is contained in the stable carbon (δ13C) and hydrogen

(δD) isotopic composition of plant waxes. Plants using the C3 photosynthetic pathway generally have smaller carbon isotope fractionation during wax biosynthesis than C4 plants (Collister et al., 1994;

Chikaraishi et al., 2004; Diefendorf and Freimuth, 2017). While there can be wide variability in lipid carbon isotope fractionation within and among species growing at the same location (e.g., Eley et al.,

2016) or in the same biome (e.g., Diefendorf et al., 2010; Wu et al., 2017), the broad isotopic differences between photosynthetic pathways can be exploited to reconstruct the relative abundance of C3 and C4 vegetation in the past (Collins et al., 2013; Garcin et al., 2014; Tierney et al., 2017).

58 The hydrogen isotopic composition of precipitation (δDp) exerts a primary control on that of plant waxes

(δDwax) both in modern plants (Sachse et al., 2006; Smith and Freeman, 2006; Feakins and Sessions,

2010; Tipple and Pagani, 2013) and in lake sediments (Sachse et al., 2004; Hou et al., 2008; Polissar and

Freeman, 2010). Thus, δDwax in lacustrine (e.g., Tierney et al., 2008; Feakins et al., 2014; Rach et al.,

2014) and marine (e.g., Pagani et al., 2006; Collins et al., 2013) sediments is commonly used to study past continental hydroclimate. Example applications include constraining the onset and duration of past drying events (e.g., Rach et al., 2014) and reconstructing changes in moisture balance (e.g., Jacob et al., 2007) or the strength of monsoon systems (e.g., Bird et al., 2014). However, quantitative reconstruction of δDp using sedimentary δDwax is hindered by lingering uncertainties regarding: 1) biological drivers of plant wax hydrogen isotope fractionation (εapp) in plants, and 2) the mechanisms and associated biases of leaf wax transport from plant source to sediment sink. Together, these processes contribute a high degree of uncertainty to the application of sedimentary plant waxes, and therefore are the focus of this study.

1.1. Drivers of εapp variability in living biomass

A suite of secondary climatic and plant physiological factors modifies the primary relationship between

δDp and δDwax. These secondary factors include soil evaporation (McInerney et al., 2011) and leaf transpiration (Feakins and Sessions, 2010; Sachse et al., 2010; Kahmen et al., 2013a; Kahmen et al.,

2013b), which can drive greater evaporative D-enrichment of plant source water in more arid settings. In addition, differences in plant physiology, leaf architecture and rooting depth can lead to differences in the extent of evaporative D-enrichment of stem and leaf water among plant groups (Smith and Freeman,

2006; Feakins and Sessions, 2010; Kahmen et al., 2013a; Kahmen et al., 2013b). In addition, all plants undergo a large (>100‰) negative biosynthetic fractionation (εbio) during lipid biosynthesis from intracellular water (Sachse et al., 2006; Sachse et al., 2012). The cumulative result of these secondary factors is the net apparent fractionation (εapp) between δDp and δDwax. Ultimately, estimates of εapp are critical for plant wax paleohydrology because εapp links measured δDwax to calculated δDp. However,

59 because εapp depends on the numerous contributing factors discussed above, it is subject to large uncertainties that hinder quantitative reconstructions of δDp. Differences in εapp among species growing at a single site can be large (between 70‰ and 110‰; Hou et al., 2007b; Feakins and Sessions, 2010; Eley et al., 2014; Cooper et al., 2015). This complicates selection of accurate εapp values for sedimentary leaf waxes, which accumulate from a mixture of plant sources at a single site.

1.2. Leaf wax transfer from plant to sediment

The transport of leaf waxes from plant source to sediment sink is poorly understood and may bias the composition of sedimentary wax records (Pancost and Boot, 2004; Sachse et al., 2012; Nelson et al.,

2018), with implications for their accurate interpretation. There are three main mechanisms that deliver plant waxes to sediments: direct leaf fall, dry and wet deposition of particulate waxes, and erosion of soil- derived waxes (Diefendorf and Freimuth, 2017, and references therein). Each of these transport mechanisms operates on different temporal and spatial scales, and may therefore introduce bias toward different plant sources. For these reasons, it is critical to constrain the influence that plant wax sourcing and transport exert on the composition of plant wax records archived in sediments.

Plant wax transport is difficult to measure directly. Previous approaches have included the use of sediment traps (Daniels et al., 2017), particulate organic matter filtering (Giri et al., 2015) and high- volume air filters (Simoneit et al., 1977; Nelson et al., 2017; Nelson et al., 2018) or rain collectors

(Meyers and Kites, 1982). This study uses a dual-isotope approach to fingerprint the molecular and isotopic (δD and δ13C) composition of n-alkanes and n-alkanoic acids in major sources (plants and particulate waxes) and associated sediments within a single catchment at Brown’s Lake Bog (BLB), Ohio,

USA. While prior studies have used a similar dual-isotope approach to fingerprint the sources of lake sediment n-alkanes in riverine systems (Seki et al., 2010; Cooper et al., 2015), this study is designed to assess only direct leaf fall and aerosol plant wax transport mechanisms because it uses a hydrologically closed basin with no surface water inputs or outlets. We consider BLB a representative site for

60 undisturbed temperate deciduous forests on formerly-glaciated terrain (Lutz et al., 2007; Glover et al.,

2011). The site includes a bog with a well-developed shoreline vegetation zone within a larger forest (Fig.

1), providing an opportunity to address the following questions: (1) How and why do leaf wax abundance and isotopic composition vary among species and growth forms at a single site?; (2) Are sedimentary plant lipids biased toward specific sources?; (3) Do the molecular and isotopic compositions of two lipid classes (n-alkanes and n-alkanoic acids) differ at the plant and sediment level? This approach extends our understanding of δDwax development beyond the level of individual plants to examine the catchment-wide integration of plant waxes in sediments. Characterizing these integration mechanisms and associated sediment bias is a critical step in reducing the uncertainties in precipitation δD reconstructions based on sedimentary plant waxes.

2. Materials and methods

2.1. Site description and sediment collection

Brown’s Lake Bog (BLB; 40.6818 °N, 82.0645 °W; 291 m above sea level) is an ombrotrophic peatland

(Sanger and Crowl, 1979) consisting of a floating Sphagnum moss mat supporting a shrub-dominated buffer of shoreline vegetation (4,329 m2) between the open water of the bog (288 m2) and the surrounding lowland fen forest, which grades into relatively well-drained oak-dominated kames (Fig. 1). The bog is hydrologically closed and is recharged by precipitation and groundwater. Bog water was collected approximately monthly during the 2014 and 2016 growing season using 30 ml Nalgene bottles.

Precipitation was collected approximately monthly during the 2014 and 2015 growing season using a 500 ml Nalgene bottle fitted with a funnel. A layer of mineral oil was used to prevent evaporation of collected precipitation. Approximately monthly, an aliquot of the collected precipitation was transferred to an airtight exetainer vial. Water samples were stored at 4 °C until isotope analysis.

61 The upper 1 m of sediment, including the sediment-water interface, was collected from the depositional center (2.5 m water depth) of BLB on January 25, 2015 using a Bolivian piston corer from an ice platform. The core was extruded in the field, transferred to Whirl-Pak bags at 1.0 cm increments, and kept frozen until lipid analysis. An additional 2 m Bolivian piston core was collected on January 30, 2016 and was used to constrain the age of the uppermost sediments using 210Pb gamma counting on 22 samples from the upper 163 cm at the St. Croix Watershed Research Station, MN, USA (Fig. 2). The same core was also used to determine magnetic susceptibility using a Barington MS2 meter in units of 10-7 m3/kg, as well as percent total organic matter and total carbonate based on the mass loss after baking dried (60 °C overnight) sediment at 550 °C (4 h) and 1000 °C (2 h), respectively (Heiri et al., 2001).

2.2. Plant sampling for lipid and stem water analysis

Leaf samples were collected from 20 species at BLB on September 20, 2015, including five deciduous angiosperm tree species (Prunus serotina, Acer saccharinum, Quercus rubra, Quercus alba, and Ulmus americana) that were sampled to track the evolution of source water and leaf wax δD over the 2014 growing season (Freimuth et al., 2017). All trees were sampled in the forested area of BLB; the remaining

15 species were sampled in the bog shoreline area (Fig. 1), and included woody shrubs (n = 2 species;

Toxicodendron vernix, Vaccinium corymbosum), woody groundcover (n = 2 species; Salix petiolaris,

Vaccinium macrocarpon), herbaceous plants (n = 7 species, including two ferns), C3 graminoids (n = 3 species), and one moss (Sphagnum sp.). In addition, leaves from one C4 graminoid (Zea mays, corn) were sampled from a farm bordering the BLB nature preserve to the west (Fig. 1). All leaves were collected in paper bags, frozen within 12 hours of collection and freeze dried prior to lipid analysis (see Section 2.5).

An additional round of plant sampling was carried out on September 22, 2016 to assess variations in plant source water δD. The same individual trees from 2015 were resampled using an arborist’s sling shot to

62 remove one branch (<1.0 cm) from the uppermost canopy of each tree. The outer bark was immediately removed and each xylem sample was transferred to a pre-ashed glass exetainer vial, sealed with an airtight septa screwcap and kept in a dry cooler in the field. Xylem samples from woody bog plants were prepared the same way. Herbaceous bog plants were also stored in airtight exetainer vials but the entire green stems were collected. All xylem samples were kept frozen (-20 ºC) until water extraction.

2.3. Monitoring wet deposition of particulate n-alkanes

Rain water was collected at BLB from July to December 2016 in order to quantify the flux of n-alkanes from aerosols via precipitation scavenging (e.g., Meyers and Kites, 1982; Gagosian and Peltzer, 1986).

Two identical collectors were placed 5 m apart in a forest clearing with no overhead foliage within 200 m of the bog shoreline (Fig. 1). Each collector consisted of a solvent-rinsed 12-liter glass carboy stabilized and shaded within a bucket buried approximately 30 cm into the ground. A 25.4 cm-diameter funnel with internal and external mesh debris guards was fitted into the opening of each carboy to direct rainwater into the collector. The collectors were decanted monthly into solvent-rinsed glass bottles. Collected water was then filtered through ashed and pre-weighed glass fiber filters in the lab to collect particles > 0.7 µm

(Whatman GE, Little Chalfont, UK). Filters from each collector and sampling month were wrapped separately in combusted foil and frozen at -20 ºC until freeze drying prior to lipid extraction (see Section

2.5). After the final rainwater collection, each collector was solvent-rinsed with dichloromethane (DCM) and Methanol (MeOH) to collect any residual lipids adsorbed to the carboy. These two samples were treated identically to the total lipid extract (TLE) from the filtered rainwater (see Section 2.5).

2.4. Xylem water extraction and δD analysis

Xylem water was extracted using cryogenic vacuum distillation following the methods of West et al.

(2006). Exetainer vials containing frozen leaves and stems were evacuated to a pressure <8 Pa (<60

63 mTorr), isolated from the vacuum pump, and heated to 100 ºC. Water vapor was collected in borosilicate test tubes immersed in liquid nitrogen for a minimum of 60 minutes. To verify extraction completion, samples were weighed following cryogenic vacuum distillation and then again after freeze drying. Based on the mass difference, the recovery of plant water was >95%. Collected water was thawed and pipetted into 2 ml crimp-top vials and refrigerated at 4 ºC until analysis.

Analysis of xylem water δD was made by headspace equilibration using 200 µl of water transferred to exetainer vials with a Pt catalyst added. Samples were purged using 2% H2 in He for 10 minutes at 120 ml/min and equilibrated at 25 ºC for 1 hour. The isotopic composition of equilibrated headspace gas was analyzed on a Thermo Delta V Advantage isotope ratio mass spectrometer (IRMS) with a Thermo

Gasbench II connected via a Conflo IV interface. Data were normalized to the V-SMOW/SLAP scale using three in-house reference standards. Precision and accuracy based on independent standards were

3.4‰ (1σ, n = 13) and -0.48‰ (n = 5), respectively.

2.5. Lipid extraction, quantification and stable isotope analysis

Dry sediment and leaf samples were ground to a powder. Separate aliquots of the dry material were used for lipid extraction and bulk δ13C analysis (see Section 2.6). Dried sediments and filtered rainwater samples (see Section 2.3) were solvent-extracted using an accelerated solvent extractor (ASE; Dionex

350) with 9:1 (v/v) DCM/MeOH with three extraction cycles at 100 ºC and 10.3 MPa. For leaf lipid extraction, ~200 mg of powdered leaves were extracted by sonicating twice with 20 ml of DCM/MeOH

(2:1, v/v), centrifuging and pipetting the lipid extract into a separate vial after each round of sonication.

For all samples, the TLE was dried under a gentle stream of nitrogen and base saponified to cleave fatty esters with 3 ml of 0.5 N KOH in MeOH/H2O (3:1, v/v) for 2 h at 75 ºC. Once cool, 2.5 ml of NaCl in water (5%, w/w) was added and acidified with 6 N HCl. The solution was extracted with hexanes/DCM

64 (4:1, v/v), neutralized with NaHCO3/H2O (5%, w/w), and water was removed through addition of

Na2SO4. Neutral and acid fractions were separated using DCM/IPA (2:1, v/v) and 4% formic acid in diethyl ether, respectively, over aminopropyl-bonded silica gel. The neutral fraction was then separated into aliphatic and polar fractions using alumina gel column chromatography. The aliphatic fraction was separated over 5% silver silica gel with hexanes to collect saturated compounds. The acid fraction of sediment and leaf samples was evaporated under a gentle stream of nitrogen and methylated by adding

~1.5 ml of 95:5 MeOH/12 N HCl (v/v) of known δD composition and heating at 70 ºC for 12-18 h. HPLC grade water was added and fatty acid methyl esters (FAMEs) were extracted with hexanes and eluted through Na2SO4 to remove water before separation over 5% deactivated silica gel with DCM to collect

FAMEs.

n-Alkanes and FAMEs were identified by GC-MS using an Agilent 7890A GC and Agilent 5975C quadrupole mass selective detector system and quantified using a flame ionization detector (FID).

Compounds were separated on a fused silica column (Agilent J&W DB-5ms) and the oven ramped from an initial temperature of 60 ºC (held 1 min) to 320 ºC (held 15 min) at 6 ºC/min. Compounds were identified using authentic standards, fragmentation patterns and retention times. All samples were diluted in hexanes spiked with the internal standard 1,1’-binaphthyl at known concentration. Compound peak areas were normalized to those of 1,1’-binaphthyl and converted to concentration using response curves for an in-house mix of n-alkanes and FAMEs at a range of concentrations. Quantified concentrations were normalized to the dry mass of material extracted and are reported as µg wax/g dry leaf or µg wax/g dry sediment in Table 1. FAME concentrations were converted to equivalent n-alkanoic acid concentrations using respective molar masses. Average chain length (ACL) is the weighted average concentration of all the long-chain waxes, defined as:

( )( ACL = & (Eq. 1) $%& (*$ )(

65 where a-b is the range of chain lengths and Ci is the concentration of each wax compound with i carbon atoms. ACL27-35 is used to indicate ACL values for n-C27 to n-C35 alkanes. ACL26-30 is used to indicate

ACL values for n-C26 to n-C30 alkanoic acids.

The isotopic composition of n-alkanes and n-alkanoic acids was determined using a Thermo Trace GC

Ultra coupled to an IRMS via a Conflo IV interface. The GC was connected to an Isolink pyrolysis reactor at 1420 ºC for δD analysis and an Isolink combustion reactor at 1000 ºC for δ13C analysis. The GC oven program ramped from 80 ºC (held 2 min) to 320 ºC (held 15 min) at a rate of 8 ºC/min. For FAME analysis, samples and standards were run with the backflush valve open to exclude high-abundance

13 compounds eluting before n-C22 alkanoic acid. A standard n-alkane mix of known δD and δ C composition (Mix A5, F8-3; A. Schimmelmann, Indiana University) was run every 6-8 samples and used to normalize the isotopic composition of samples to the Vienna Standard Mean Ocean Water (VSMOW) scale (for δD) and the Vienna Peedee Belemnite (VPDB) scale (for δ13C) following Coplen et al. (2006).

The δD value of hydrogen added to n-alkanoic acids during derivatization was determined by mass balance of phthalic acid of known δD composition (A. Schimmelmann, Indiana University) and derivatized methyl phthalate. Based on repeat injections for a subset of samples, the mean sample

13 + uncertainty was 3.3‰ for δD and 0.22‰ for δ C. The H3 factor was tested daily during δD analysis and

-1 averaged 3.8 ppm mV . Hydrogen isotopic fractionations (εapp, εbio, etc.) are reported as enrichment factors (in units of per mil, ‰), using the following equation:

(./012) -$%& = − 1 (Eq. 2) (./412) where a is the product and b is the substrate.

2.6. Bulk foliar carbon isotope analysis

66 A separate aliquot of the powdered leaves was used to determine the δ13C of bulk organic carbon. Bulk

δ13C analysis was performed via continuous flow (He; 120 ml/min) on a Costech elemental analyzer (EA) interfaced with a Thermo Delta V Advantage isotope ratio mass spectrometer (IRMS) via a Conflo IV.

All δ13C values were corrected for sample size dependency and normalized to the VPDB scale using a two-point calibration with in-house standards which were calibrated with IAEA standards (NBS-19, L-

SVEC) to -38.26‰ and -11.35‰ VPDB following Coplen et al. (2006). Based on additional independent standards, sample precision was 0.03‰ (1s, n = 11) and accuracy was 0.03‰ (1s, n = 11). All data analyses were performed using JMP Pro 12.0 (SAS, Cary, NC, USA) and significance was attributed to alpha levels at or below 0.05.

3. Results

3.1. Environmental waters at BLB

Environmental and plant waters are reported in Fig. 3. Bog water δD values increased steadily from

March (-73‰) to May (-26‰), and remained relatively constant through autumn (May to October mean =

-30 ± 4‰, n = 6). Precipitation collected approximately monthly from March to November 2015 ranged from -79‰ in March to -13‰ in September, with a mean of -44‰. The modeled mean annual precipitation δD value for BLB is -50 ± 2‰ from the Online Isotopes in Precipitation Calculator (OIPC;

Welker, 2000; Bowen and Revenaugh, 2003; IAEA/WMO, 2015; Bowen, 2017). Plants growing in the bog shoreline zone had significantly higher xylem water δD values (-27 ± 8‰, n = 25) than trees occupying the forested zones of the catchment (-44 ± 10‰, n = 5; t-statistic = 0.0007, p ≤ 0.024). There were no significant differences in xylem water δD values among shoreline plant growth forms (t-test, p >

0.26).

67 3.2. Leaf wax molecular and isotopic composition in plants

Leaf wax abundance varied widely among plant growth forms at BLB (Fig 4; Table 1). We focus on n-C29 alkane and n-C28 alkanoic acid as chain-lengths indicative of terrestrial plants and molecules commonly used in paleohydrology; additional chain-lengths are reported in Table 1. Mean n-C29 alkane concentrations were an order of magnitude higher in trees (520 µg/g dry weight) than in the second most waxy growth form (shrubs, 31 µg/g). The lowest concentrations of n-C29 alkane were observed in ferns, mosses and C3 graminoids growing in the bog shoreline (ranging from 1.8 to 8.6 µg/g). Within the tree growth form, there were significant interspecific differences in n-C29 alkane concentration; lowest in Q. alba and U. americana (34 and 64 µg/g), intermediate in A. saccharinum (289 µg/g) and highest in Q. rubra and P. serotina (676 and 731 µg/g). The ACL27-35 in all plants ranged from 28.8 in forbs to 31.4 in

Z. mays (Fig. 5b), with comparable values in trees (29.6) and shrubs (29.9). Particulate waxes had similar

ACL27-35 (29.5). By contrast, the concentration of n-C28 acid, and its variability among growth forms was considerably smaller, from 4.0 µg/g in moss to 33.8 µg/g in trees (Fig. 4a; Table 1). The ACL26-30 was consistently lower for trees (27.5) than for all other growth forms, which ranged from 28.7 in C3 grass to

29.2 in shrubs and woody vines (Fig. 5).

The overall range in δDwax values among individual plants was 77‰ for n-C29 alkane and 84‰ for n-C28 alkanoic acid (Table 3). Isotopic variability (1s) among species was approximately 2-4 times greater for shoreline plants (n-C29 alkane and n-C28 alkanoic acid 1s = 23‰ and 17‰, respectively) than for trees in the forested areas of the catchment (1s = 6‰ and 9‰, respectively). Shrubs and C3 graminoids had the lowest mean n-C29 alkane δD values (both -196‰), although variability among the four shrub species was notably high (27‰). Mean n-C29 alkane δD values were intermediate in forbs (-167 ± 9‰, n = 4) and trees (-163 ± 6‰, n = 5), and highest in the C4 graminoid (Z. mays, -153‰). C3 graminoids had the lowest mean n-C28 alkanoic acid δD values (-182 ± 3‰, n =3). Shrubs and forbs had similar mean n-C28

68 alkanoic acid δD values (-156 ± 18‰, n = 9; and -152 ± 11‰, n = 9); trees were higher (-139 ± 9‰, n =

5); and Z. mays had the highest observed value (-101‰). Prior studies have observed similar distributions, with δD values for long-chain n-alkanes being lowest in C3 graminoids, highest in trees, and intermediate in herbaceous plants (Liu et al., 2006; Hou et al., 2007b; Eley et al., 2014; Cooper et al.,

2015).

Apparent fractionation for n-C29 alkanes relative to mean annual precipitation δD (ε29-MAP) ranged among growth forms, from -151‰ in C3 graminoids and shrubs, -121‰ in forbs, -116‰ in trees, and -105‰ in

Z. mays (Table 3). The corresponding apparent fractionation for n-C28 alkanoic acid relative to mean annual precipitation δD (ε28-MAP) was approximately 15-55‰ smaller than ε29-MAP in each growth form, from -136‰ in C3 graminoids, -107‰ in shrubs, -105‰ in forbs, -91‰ in trees, and -51‰ in Z. mays.

13 The overall range in leaf wax δ C values among individual plants was 21‰ for n-C29 alkane and 16‰

13 for n-C28 alkanoic acid. Interspecies n-C29 alkane δ C variability was higher for shoreline plants (2.4‰) than for trees (1.2‰). Isotopic variability in n-C28 alkanoic acid among species was comparable for

13 shoreline plants and trees (1s = 2.3‰ and 2.2‰, respectively). Mean n-C29 alkane δ C values were lowest in shoreline vegetation, ranging from forbs (-39.3 ± 1.8‰, n = 3) to C3 graminoids (-36.6 ± 0.9‰,

13 n = 2) and shrubs (-35.2 ± 1.9‰, n = 4). Trees had a higher mean n-C29 alkane δ C value (-33.8 ±1.2 ‰, n = 5), and Z. mays was highest (-20.7‰), consistent with prior studies (Collister et al., 1994; Pancost and

Boot, 2004; Chikaraishi and Naraoka, 2007). The distribution of δ13C values by growth form was similar for n-C28 alkanoic acid, with the lowest mean values in forbs (-39.0 ± 1.1, n = 4), shrubs (-36.5 ± 2.4 ‰, n

13 = 3) and C3 graminoids (-35.5 ± 2.1 ‰, n = 3). Trees had a higher mean n-C28 alkanoic acid δ C value (-

34.5 ± 2.2‰, n = 5), and Z. mays was highest (-24.5‰).

3.3. Particulate wax flux and composition

69

The particulate wax samples were only analyzed for n-alkanes and not for n-alkanoic acids. The apolar lipid fractions were dominated by n-C29 and n-C31 alkanes, with smaller amounts of n-C25, n-C27 and n-C33 alkanes (Fig. 4b; Table 2). These molecular distributions are similar to those found previously in aerosols and precipitation and indicate that the long-chain (≥ n-C27) alkanes are sourced primarily from terrestrial plants (Simoneit et al., 1977; Broddin et al., 1980; Meyers and Kites, 1982). The concentration of total n- alkanes (odd-carbon numbers from n-C23 to n-C31) in precipitation at BLB ranged from 0.2 to 1.2 µg/L; n-

C29 alkane ranged from 0.1 to 0.5 µg/L and n-C31 alkane ranged from 0 to 0.3 µg/L. These concentrations are comparable to, or smaller than, those found in rain and snow collected in the Midwestern USA, in which total n-alkane concentrations ranged from 0.4 to 8 µg/L (Meyers and Kites, 1982). Due to low abundance of monthly particulate wax samples, all samples were combined prior to compound-specific

13 isotope analysis. Pooled δ C values for particulate wax n-C29 and n-C31 alkanes were -30.7‰ and -

31.1‰, slightly more 13C-enriched compared to most forest and bog shoreline plants (Fig. 6a). Pooled δD values for particulate wax n-C29 and n-C31 alkanes were -167‰ and -156‰, closely matching the n-C29 alkane values measured in trees and forest litter (Fig. 6a).

3.4. Bog sediment chronology and bulk properties

Bog sediment chronology was established using 210Pb on the core collected in 2016 with an oldest reliable date of 1848 ± 10 CE at 92.5 cm (Fig. 2) and a mean sediment accumulation rate of 0.81 cm yr-1 since

~1900 CE. The sediment was uniformly fine-grained and dark brown over the entire collected interval.

The magnetic susceptibility was < 6 x 10-7 m3/kg for most of the 0-40 cm interval, with the exception of an abrupt increase at 26 cm to 36 x 10-7 m3/kg. The mean %TOM was 66% in the upper 24 cm, 87% from

26 to 32 cm, and 69% from 34 to 80 cm. The mean %TC from 0 to 80 cm was 1%.

70 Based on sediment chronology and bulk characteristics we limited our plant wax sampling strategy

(Section 2.5) to the upper 40 cm of sediment, which corresponds to the years 1998-2015 CE. During this time, the vegetation assemblage and distribution within the BLB catchment would have been comparable to today because the site has remained an undeveloped state nature preserve since 1967 CE.

3.5. Plant wax molecular and isotopic composition in bog sediments

The upper 40 cm of bog sediments contained plant waxes with a mean ACL27-35 of 29.3 and ACL26-30 of

27.4. The most abundant compound was generally n-C29 alkane (17 µg/g dry sediment), followed by n-

C27 and n-C31 alkane (12 and 9 µg/g, respectively; Table 1). There was a strong odd-over-even chain- length preference for n-alkanes (mean CPI = 14.6) indicating that sedimentary waxes are derived from terrestrial plant sources (Freeman and Pancost, 2014). Long-chain n-alkanoic acids were dominated by n-

C26 alkanoic acid (12 µg/g). Average concentrations for n-C28 and n-C30 alkanoic acids were lower (5 and

4 µg/g, respectively; Table 1).

The total range in δDwax values for the upper 40 cm of sediments was 9‰ for n-C29 alkane and 11‰ for n-

C28 alkanoic acid, with mean values of -176 ± 3‰ (n = 11) and -148 ± 5‰ (n = 6), respectively (Table 3).

Corresponding ε29-MAP and ε28-MAP values in sediments were -133 ± 3‰ and -103 ± 5‰ (n = 6). The range

13 in δ Cwax values in sediments was < 1‰, with a mean of -32.9 ± 0.2‰ (n = 11) for n-C29 alkane and -

33.9 ± 0.36‰ (n = 10) for n-C28 alkanoic acid.

4. Discussion

4.1. Drivers of differences in δDwax values among plant species and growth forms

71

One major challenge for interpreting sedimentary δDwax values is the wide range in δDwax values observed for different plant species and growth forms growing at a single site receiving the same precipitation. The range in δD values observed for n-C29 alkane (77‰) and n-C28 alkanoic acid (84‰) among all plants at

BLB is comparable with that observed among diverse growth forms in other temperate ecosystems (75-

98‰ and 100-194‰, respectively, Chikaraishi and Naraoka, 2007; Hou et al., 2007b; Eley et al., 2014;

Cooper et al., 2015). One possible driver of the range in δDwax values observed at BLB is differences in source water used by plants growing in different zones of the catchment. Plants in the bog shoreline zone had xylem water δD values that reflected May to October bog water δD values (between -27‰ and -

30‰). The xylem water of shoreline plants was D-enriched by 17‰ on average compared to trees, in which xylem water δD (-44‰) was closer to mean annual precipitation δD (-50‰). Because xylem water samples were collected in September, it is possible that their δD values are not representative of the source water used by plants during springtime wax synthesis. However, in an earlier study (Freimuth et al., 2017), xylem water δD values in trees at BLB were relatively stable throughout the growing season (-

50 ± 10‰ 1s; n = 70). Therefore, September xylem water δD values reported here for trees should be comparable to source water used by trees during spring lipid synthesis.

There is less available information about the isotopic variability of source water accessed by the shoreline plants. Bog water δD was relatively stable over much of the growing season (Fig. 3). Therefore, if shoreline plants are primarily accessing bog water, their September xylem water δD values may be representative of the growing season. However, we did not explore seasonal changes in lipid abundance or δDwax, and it may be the case that the diverse species present in the shoreline zone synthesize lipids at different (or multiple) points throughout the growing season (Pedentchouk et al., 2008; Sachse et al.,

2009; Newberry et al., 2015). Further, synthesis of n-alkanes may occur earlier in the growing season than n-alkanoic acids (Freimuth et al., 2017), which could further affect source water used to produce the two compounds. Thus, a shoreline plant that produces leaf wax in March or April could be accessing bog

72 water that is D-depleted by ~20‰ relative to the values across most of the growing season (May to

October; Fig. 3). Therefore, interspecies differences in the timing of leaf maturation and lipid synthesis may be a factor contributing to the pronounced variability in δDwax values among shoreline species, though ~20‰ differences in early-spring bog water cannot account for the total range in δDwax values observed for plants.

The isotopic differences between the primary source water for shoreline plants (bog water) and trees

(precipitation-fed soil water) may be amplified by physiological factors particular to the plants growing in different areas within the catchment. For instance, shoreline graminoids, herbs and shrubs generally have shallower rooting depths than the trees, and may therefore access source water that is more evaporatively

D-enriched. Further, many shoreline plants at BLB grow on Sphagnum hummocks that are subaerially exposed and therefore more vulnerable to evaporative D-enrichment than forest soils (Nichols et al.,

2010), which are shielded from evaporation by canopy and groundcover. These factors, along with a greater diversity of growth forms, physiology and leaf morphologies in the shoreline zone, may contribute to the 2-4 times larger variability in n-C29 alkane among shoreline plants than among the tree species

(Levene test, p = 0.04).

Nevertheless, differences in source water δD cannot account for the overall range in δDwax values among

2 plants. Species-level xylem water δD had no significant relationship with n-C29 alkane δD (R = 0.17, p =

2 0.08, n = 19) but could explain approximately one third of the variation in n-C28 alkanoic acid δD (R =

0.36, p = 0.002, n = 24). Therefore, differences in biosynthetic fractionation among species must therefore account for the majority of the observed range in δDwax values among species sharing common growth conditions and similar source water. We did not analyze leaf water δD values, and therefore cannot approximate εbio for each species. However, in a similar study, εwax/leaf water (~εbio) varied by 150‰ among species sampled at the same time at a single temperate site (Eley et al., 2014). While Eley et al. (2014) included saltmarsh plants growing in saline soils, the total range in plant source (xylem) water (38‰) was

73 comparable to that at BLB for species representing multiple growth forms (graminoids, C3 herbaceous plants and shrubs). Therefore, we would expect a comparable range in εbio at BLB. A wide range (>70‰) of biologically-driven δDwax variability among species within a single site can be expected (e.g., Hou et al., 2007; Eley et al., 2014), owing mainly to differences in εbio. While it is important to understand the specific drivers of this variability, the issue most relevant to the application of the δDwax paleohydrology proxy is how this variability is represented in sediments. Identifying which plant sources, if any, are preferentially represented in sediments will provide a means to increase the accuracy and reduce the range of uncertainty in εapp estimations for interpretation of past precipitation δD values based on sedimentary

δDwax values.

4.2. Sediment bias toward n-alkane sources

Comparing the molecular and isotopic composition of plant waxes in vegetation with the associated sediments can help identify bias in what is ultimately preserved in sediments. The range in δDwax values at

BLB was far greater among plant species (77 to 84‰) than in bog surface sediments (6 to 11‰; Fig. 6,

7). Reduced variability in δDwax values at the sediment-level indicates that the integration of plant waxes in sediments involves either consistent mixing of all sources or biased incorporation of lipids from a subset of sources.

Terrestrial plant lipids are delivered to terrestrial sediments through three primary transport mechanisms: direct leaf fall, dry or wet deposition of particulate wax in aerosols or dust, and erosion of soil-derived waxes (Diefendorf and Freimuth, 2017, and references therein). The latter process is likely minimal at

BLB, where there are no surface water inputs to the bog. Additionally, total relief of the site is low (12 m), and most of the topographic relief is focused around several kames in the interior of the forest, which have extensive groundcover that acts to stabilize their soils. Therefore, direct leaf fall and deposition of particulate waxes are likely the main modes of plant lipid transport to sediments at BLB. While we did

74 not directly measure the flux of lipids to sediments, comparing the molecular and isotopic signatures of n- alkanes and n-alkanoic acids in the main sources at BLB (plant leaves and rain-scavenged particulate wax) to those found in recent sediments offers insight into their relative contribution to the sedimentary record.

If sedimentary plant waxes at BLB were simply a mixture derived from leaves of all plant sources, the

13 expected δDwax and δ Cwax in sediments could be approximated using the concentration-weighted mean

13 isotopic composition of all plant species. The concentration-weighted mean n-C29 alkane δD and δ C from plants at BLB (-161‰ and -33.1‰, respectively; Fig. 7) is indistinguishable from sediment δ13C (-

32.9‰; t-test, p = 0.88), but significantly D-enriched relative to sediment δD (-176 ± 2.5‰; t-test, p =

0.05). The calculated weighted mean for δDwax is likely D-enriched compared to sediments because it is heavily weighted toward P. serotina and Q. rubra (-160‰ and -155‰, respectively), which had n-C29 alkane concentrations that were 2 to 50 times higher than all other trees, and 10 to 300 times higher than shoreline species. This is evidence that, even when plants within a catchment differ in n-alkane abundance by orders of magnitude, the signal in sediments does not simply reflect only the waxiest species but is a complex mixture of different sources. For n-C31 alkane, the discrepancy between the concentration-weighted mean δD value for plants (-155‰) and that measured in sediments (-176‰) was even more pronounced, indicating that there must be significant sedimentary contributions from sources that are more D-depleted than trees, but with a similar δ13C value. This rules out significant contributions from the particulate waxes we measured, which were consistently D-enriched and 13C-enriched relative to both trees and sediments.

For the same reasons, significant contributions from Z. mays can be ruled out; its n-C29 alkane values were consistently enriched in both 13C (by > 10‰) and deuterium relative to sediments (Fig. 6). Further, the ACL for Z. mays is distinct from the sediments, indicating little to no influence on sediments based on molecular distribution (Fig. 5). If waxes from Z. mays were transferred to sediments, direct leaf fall is not

75 likely to be a dominant transport mechanism because of its annual lifespan, the intervention of seasonal harvesting, and the forest buffer between the adjacent agricultural fields and the bog (Fig. 1). Instead, if lipids from Z. mays are mobilized they would most likely be delivered to sediments as particulate wax and therefore distinctly 13C-enriched wax values may serve as an additional tracer for the flux of particulate waxes to sediments. It may be the case that any potential input of n-C29 alkane from Z. mays is swamped by trees given the far higher rates of n-C29 alkane production in the latter. As a test for potential

13 contributions from Z. mays we compared the δ C composition of n-C33 alkane in sediments (-33.3 ±

0.5‰, 1s) with that in the rain-scavenged wax (-30‰) and Z. mays (-19.7‰), which had relatively high

13 n-C33 alkane concentrations relative to other plant sources. Both sources were C-enriched by more than

13 2s of the mean δ Cwax in sediments.

Therefore, based on the δD and δ13C composition of multiple long-chain n-alkanes, particulate wax sources (rain-scavenged waxes collected adjacent to the bog and Z. mays likely to be transported as particulate wax) have limited to no influence on sediments at BLB, which instead are more likely to reflect direct leaf fall from surrounding vegetation. A limited contribution of particulate waxes to sediments at BLB is consistent with recent evidence from a temperate forested region in Europe where the dry deposition of particulate waxes was estimated to contribute only ~20% of the total wax flux to sediments (Nelson et al., 2017; Nelson et al., 2018).

Based on the limited contribution of particulate sources, we excluded the particulate wax and Z. mays data when comparing all of the sediment and species-level n-alkane isotope data using the Stable Isotope

Mixing Models in R (simmr) package (Parnell et al., 2010; Parnell et al., 2013). This dual-isotope mixing

13 model uses summary (mean ±1s) δDwax and δ Cwax values for each species and applies a Bayesian framework to estimate the fractional contributions of various sources (i.e., BLB plants) to a mixture (i.e.,

BLB sediments). This model does not take into account plant wax concentrations or molecular

76 distributions among the different plant sources, and therefore offers a simplified analysis of this system.

Nevertheless, we used this model as a tool to explore the BLB data in δD and δ13C space and found that, for n-C29 alkane, the species with the greatest fractional contribution to bog sediments were Q. rubra

(0.59) and V. macrocarpon (0.24), with smaller contributions from P. serotina (0.028) and T. vernix

(0.026). This includes the tree species that produce the highest abundance of long-chain n-alkanes (Q. rubra and P. serotina) and one of the two dominant shrubs in the bog shoreline zone (T. vernix). The woody groundcover species V. macrocarpon had a relatively large modeled contribution, perhaps because it’s extremely D-depleted n-C29 alkane value (-229‰; Table 1, 3) was necessary to bring a mainly tree- derived δDwax signal in line with values in sediments (Fig. 7). This is an interesting result in that, even though the simmr model does not include concentration data, the dominance of the waxiest trees was apparent in the model results on the basis of δD and δ13C values alone. It is also notable in that all of the species with a high modeled contribution to sediments were woody, but include three different growth forms. This suggests the dominance of woody plant sources in close proximity to the basin, regardless of growth form or species-dependent plant wax concentration.

Together, these results indicate that the major sources of n-C29 alkane to BLB sediments are trees and select woody shoreline vegetation (likely via direct leaf fall), with a negligible contribution from particulate waxes. These findings are in agreement with prior studies of sediment apportionment using molecular and isotopic analysis of plant lipids also found that trees were a dominant source of sedimentary lipids (Sachse et al., 2004; Seki et al., 2010; Tipple and Pagani, 2013; Cooper et al., 2015).

This indicates that the n-alkane signal incorporated in sediments in forested closed basin at BLB are not regional in nature (e.g., particulate waxes) but instead are derived primarily from the plants growing in close proximity to the basin.

Weighting the εn-C29/MAP values of each species by the simmr model estimates of the fractional contribution of each source gives an overall εn-C29/MAP of -132‰, which yields a δDMAP of -50‰ for BLB

77 when applied to sedimentary n-C29 alkane δD values. This is identical to both the OIPC-modeled δDMAP

(Welker, 2000; Bowen and Revenaugh, 2003; IAEA/WMO, 2015; Bowen, 2017) and mean xylem water

δD across the 2014 growing season (Freimuth et al., 2017). This confirms that, while the modeled proportional contributions are just one possible scenario based only on δD and δ13C values, the resulting estimates are reasonable and reflect the fidelity of n-alkanes from woody plants as recorders of precipitation. Therefore, at BLB and perhaps similar temperate deciduous forested sites without surface water inputs as a mechanism for soil transport, sedimentary n-C29 alkanes primarily reflect direct leaf fall from trees and select woody shoreline species, with minor contributions from herbs, graminoids and particulate aerosol waxes.

4.3 Sediment bias toward n-alkanoic acid sources

13 In contrast to n-alkanes, the concentration-weighted mean n-C28 alkanoic acid δD and δ C values for all species (-142‰ and -34‰; Fig. 7) are within approximately one standard deviation of the measured mean

δD and δ13C in sediments (-148 ± 5‰ and -33.9 ± 0.4‰, respectively). This likely reflects the more uniform distribution of n-C28 alkanoic acid concentrations across growth forms. As with n-C29 alkane, the simmr model predicts that trees have the greatest contribution to sedimentary n-C28 alkanoic acid. The four species with the greatest fractional contribution to n-C28 alkanoic acid δD values in sediment were all trees: P. serotina (0.42), A. saccharinum (0.1), Q. alba (0.07), and Q. rubra (0.06). The modeled fractional contributions from all remaining species showed little variation, ranging from 0.03 to 0.04. If these modeled contributions are accurate, then sedimentary n-C28 alkanoic acids may be more strongly biased toward trees than n-C29 alkanes, which may be more generally derived from woody trees, shrubs and groundcover. The tree-dominated model results for n-C28 alkanoic acid isotopes are consistent with molecular signatures, which suggest that sediment ACL26-30 reflects that of trees more than any other plant source (Fig. 4, 5).

78 To assess model estimates, we weighted the measured εn-C28/MAP of each plant species by its modeled contribution to sediments, which yields an overall εn-C28/MAP of -58‰ and a reconstructed δDMAP of -95‰, which is widely inaccurate compared to the expected δDMAP of -50‰. This suggests that the model may in fact overestimate the combined n-C28 alkanoic acid contributions from sources other than trees. Using the εn-C28/MAP for trees only (-92‰) yields a δDMAP estimate of -61‰ based on sediment δDwax. When using the εn-C28/xyl value for trees, reconstructed δDMAP (-48‰) is much closer to the expected value.

Together, these result support that sedimentary n-C28 aklanoic acids are weighted toward contributions from trees, but small contributions from other plant sources are possible given the similar n-alkanoic acid concentrations among species and the close agreement between sediment and concentration-weighted plant wax isotope values (Fig. 7).

4.4. Molecular and isotopic differences between n-alkanes and n-alkanoic acids

There were several molecular and isotopic differences between n-alkanes and n-alkanoic acids at BLB that may influence their application as proxies for past precipitation δD. First, leaf wax concentrations among species varied widely (by two orders of magnitude) for n-alkanes but not for n-alkanoic acids (Fig.

4). This is consistent with prior studies that found relatively consistent n-alkanoic acid concentrations among species compared to greater variability in n-alkanes (Diefendorf et al., 2011; Bush and McInerney,

2013; Polissar et al., 2014). The evidence that sedimentary n-alkanoic acids are more strongly biased toward trees than n-alkanes (which may reflect woody plants more generally) suggests that leaf wax abundance among species or compound classes is not necessarily a direct indicator of the dominant sources to sediments. Given the differences in source water for trees (precipitation-fed soil water) and shoreline plants (D-enriched bog water) at BLB, a stronger bias toward trees may make sedimentary n-C28 alkanoic acid a higher-fidelity proxy for precipitation δD. By contrast, if sedimentary n-C29 alkanes are derived from more varied woody plant sources, their δD values may reflect more mixed source waters including the bog itself.

79

Second, we observed a consistent negative offset in n-C29 alkane δD values relative to n-C28 alkanoic acid, both in plants and in sediments. The offset in plants was variable, ranging from -19‰ (C3 graminoids) to -61‰ (woody groundcover); however, mean offsets ranged from -26‰ to -23‰ in forbs, trees and shrubs, closely reflecting that in sediments (-27‰). An offset of comparable magnitude and sign has previously been observed in temperate settings and attributed to fractionation associated with biosynthesis of the two compounds (Chikaraishi and Naraoka, 2007; Hou et al., 2007b; Freimuth et al.,

2017). Persistence of this offset in sediments may warrant application of different εapp values, especially considering that existing calibrations of εapp in plants have mainly been developed for n-alkanes (Sachse et al., 2012). For example, applying εapp values based on calibration data from n-alkanes to sedimentary n- alkanoic acids at BLB would overestimate δDMAP by approximately 30‰. However, we note that while systematic δD offsets between compound classes is a common feature in temperate settings dominated by plants with a single annual leaf flush, this may not be a relevant consideration in settings with a higher diversity of plant species and leaf lifespans (Gao et al., 2014; Feakins et al., 2016).

Lastly, in trees and shrubs we observed a positive linear relationship between the δD and δ13C values of n-

2 2 C29 alkane (R = 0.66, p = 0.0004) and a weaker, negative relationship for n-C28 alkanoic acid (R = 0.32, p = 0.028; Fig. 8). This is consistent with a prior study, which found that long-chain n-alkanoic acids from trees (Hou et al., 2007a) became more 13C-enriched as they became more D-depleted, similar to n- alkanoic acids at BLB. This may be because plants with higher water use efficiency (WUE; reflected in

13 more positive δ Cwax values) undergo less transpirational D-enrichment of intracellular water during the production of leaf waxes, which therefore have lower δDwax values. We observed a stronger relationship of opposite sign for n-C29 alkanes, similar to observations by Bi et al. (2005) for n-alkanes from trees. It remains unclear why the sign of this relationship would change for different compounds or photosynthetic pathways, but we speculate that if there is a mismatch in the seasonal timing of production of n-alkanes during the brief period of leaf expansion and n-alkanoic acids throughout the entire growing season

80 (Tipple et al., 2013; Freimuth et al., 2017), then this might reflect contrasting metabolic and WUE states in young, rapidly expanding leaves (n-alkanes) and mature, opportunistic leaves (n-alkanoic acids).

5. Conclusions

We surveyed leaf waxes (n-alkanes and n-alkanoic acids) in all major plant species growing in the BLB catchment and compared their concentration, molecular distribution and isotopic composition (δD and

δ13C) with that in bog sediments. Sampled species encompassed a range of growth forms occupying the bog shoreline and primarily accessing bog water, as well as trees that dominate forested areas and primarily access precipitation-fed soil water. n-Alkane concentrations were variable among species and were highest in trees, while long-chain n-alkanoic acid concentrations were generally lower and more evenly distributed among growth forms. The total range in sedimentary δDwax (6 to 11‰) was far smaller than in plants at BLB (77 to 84‰), indicating some predictable integration processes in sediments.

Our results suggest that for both n-alkanes and n-alkanoic acids, direct leaf fall from trees and woody shrubs growing near the bog was a more significant wax transport mechanism than the deposition of

13 particulate aerosol waxes. Based on δD and δ C signatures, both n-C29 alkanes and n-C28 alkanoic acids in sediments primarily reflect trees, although n-C29 alkanes may have more mixed woody sources including shoreline shrubs and groundcover. Therefore, n-C28 alkanoic acid may be more reflective of soil water (i.e., precipitation) whereas n-C29 alkane in sediments may reflect more mixed water sources including bog water. We found a ~30‰ offset in εapp values for n-C29 alkane (-133‰) and n-C28 alkanoic acid (-103‰) in sediments, a feature previously observed in temperate settings. These insights into sediment bias toward plant sources and lipid transport processes provide a framework for more accurate interpretations of lake sediment leaf waxes.

81 Acknowledgements

We thank The Nature Conservancy for granting research access to Brown’s Lake Bog State Nature

Preserve. Thanks also go to Nicholas Wiesenberg, Katherine McNulty and Kelly Grogan for field sampling assistance, and again to Kelly Grogan for laboratory assistance. We thank Erin Mortenson and

Daniel Engstrom at the St. Croix Watershed Research Station for sediment 210Pb analysis. This research was supported in part by the US National Science Foundation (EAR-1229114 to AFD) and by the Donors of the American Chemical Society Petroleum Research Fund (PRF #51787-DN12 to AFD). This research was also supported by awards from the Geological Society of America and Sigma Xi to EJF.

82 REFERENCES

Aichner, B., Herzschuh, U., Wilkes, H., Vieth, A. and B√∂hner, J.r. (2010) dD values of n-alkanes in Tibetan lake sediments and aquatic macrophytes - A surface sediment study and application to a 16 ka record from Lake Koucha. Org. Geochem. 41, 779-790. Bi, X., Sheng, G., Liu, X., Li, C. and Fu, J. (2005) Molecular and carbon and hydrogen isotopic composition of n-alkanes in plant leaf waxes. Org. Geochem. 36, 1405-1417. Bird, B.W., Polisar, P.J., Lei, Y., Thompson, L.G., Yao, T., Finney, B.P., Bain, D.J., Pompeani, D.P. and Steinman, B.A. (2014) A Tibetan lake sediment record of Holocene Indian summer monsoon variability. Earth Planet. Sci. Lett. 399, 92-102. Bowen, G. J. (2017) The Online Isotopes in Precipitation Calculator, version 3.1. http://www.waterisotopes.org. Bowen, G.J. and Revenaugh, J. (2003) Interpolating the isotopic composition of modern meteoric precipitation. Water Resources Research 39. Broddin, G., Cautreels, W. and Van Cauwenberghe, K. (1980) On the aliphatic and polyaromatic hydrocarbon levels in urban and background aerosols from Belgium and the Netherlands. Atmospheric Environment (1967) 14, 895-910. Bush, R.T. and McInerney, F.A. (2013) Leaf wax n-alkane distributions in and across modern plants: Implications for paleoecology and chemotaxonomy. Geochim. Cosmochim. Acta 117, 161-179. Chikaraishi, Y., Naraoka, H. and Poulson, S.R. (2004) Hydrogen and carbon isotopic fractionations of lipid biosynthesis among terrestrial (C3, C4 and CAM) and aquatic plants. Phytochemistry 65, 1369-1381. Chikaraishi, Y. and Naraoka, H. (2007) δ13C and δD relationships among three n-alkyl compound classes (n-alkanoic acid, n-alkane and n-alkanol) of terrestrial higher plants. Org. Geochem. 38, 198-215. Collins, J.A., Schefuß, E., Mulitza, S., Prange, M., Werner, M., Tharammal, T., Paul, A. and Wefer, G. (2013) Estimating the hydrogen isotopic composition of past precipitation using leaf-waxes from western Africa. Quat. Sci. Rev. 65, 88-101. Collister, J.W., Rieley, G., Stern, B., Eglinton, G. and Fry, B. (1994) Compound-specific δ13C analyses of leaf lipids from plants with differing carbon dioxide metabolisms. Org. Geochem. 21, 619-627. Cooper, R.J., Krueger, T., Hiscock, K.M. and Rawlins, B.G. (2015) High-temporal resolution fluvial sediment source fingerprinting with uncertainty: a Bayesian approach. Earth Surface Processes and Landforms 40, 78-92. Coplen, T.B., Brand, W.A., Gehre, M., Gröning, M., Meijer, H.A.J., Toman, B. and Verkouteren, R.M. (2006) New guidelines for δ13C measurements. Anal. Chem. 78, 2439-2441. Daniels, W.C., Russell, J.M., Giblin, A.E., Welker, J.M., Klein, E.S. and Huang, Y. (2017) Hydrogen isotope fractionation in leaf waxes in the Alaskan Arctic tundra. Geochim. Cosmochim. Acta 213, 216-236. Diefendorf, A.F., Mueller, K.E., Wing, S.L., Koch, P.L. and Freeman, K.H. (2010) Global patterns in leaf 13C discrimination and implications for studies of past and future climate. Proc. Natl. Acad. Sci. U.S.A. 107, 5738-5743. Diefendorf, A.F., Freeman, K.H., Wing, S.L. and Graham, H.V. (2011) Production of n-alkyl lipids in living plants and implications for the geologic past. Geochim. Cosmochim. Acta 75, 7472-7485. Diefendorf, A.F., Sberna, D.T. and Taylor, D.W. (2015) Effect of thermal maturation on plant-derived terpenoids and leaf wax n-alkyl components. Org. Geochem. 89, 61-70. Diefendorf, A.F. and Freimuth, E.J. (2017) Extracting the most from terrestrial plant-derived n-alkyl lipids and their carbon isotopes from the sedimentary record: A review. Org. Geochem. 103, 1- 21. Eglinton, G. and Hamilton, R.J. (1967) Leaf epicuticular waxes. Science 156, 1322-1335.

83 Eley, Y., Dawson, L., Black, S., Andrews, J. and Pedentchouk, N. (2014) Understanding 2H/1H systematics of leaf wax n-alkanes in coastal plants at Stiffkey saltmarsh, Norfolk, UK. Geochim. Cosmochim. Acta 128, 13-28. Eley, Y., Dawson, L. and Pedentchouk, N. (2016) Investigating the carbon isotope composition and leaf wax n-alkane concentration of C3 and C4 plants in Stiffkey saltmarsh, Norfolk, UK. Org. Geochem. 96, 28-42. Feakins, S.J. and Sessions, A.L. (2010) Controls on the D/H ratios of plant leaf waxes from an arid ecosystem. Geochim. Cosmochim. Acta 74, 2128-2141. Feakins, S.J., Kirby, M.E., Cheetham, M.I., Ibarra, Y. and Zimmerman, S.R.H. (2014) Fluctuation in leaf wax D/H ratio from a southern California lake records significant variability in isotopes in precipitation during the late Holocene. Org. Geochem. 66, 48-59. Feakins, S.J., Bentley, L.P., Salinas, N., Shenkin, A., Blonder, B., Goldsmith, G.R., Ponton, C., Arvin, L.J., Wu, M.S. and Peters, T. (2016) Plant leaf wax biomarkers capture gradients in hydrogen isotopes of precipitation from the Andes and Amazon. Geochim. Cosmochim. Acta 182, 155-172. Ficken, K.J., Li, B., Swain, D.L. and Eglinton, G. (2000) An n-alkane proxy for the sedimentary input of submerged/floating freshwater aquatic macrophytes. Org. Geochem. 31, 745-749. Freeman, K.H. and Pancost, R.D. (2014) Biomarkers for Terrestrial Plants and Climate, in: Turekian, H.D.H.K. (Ed.), Treatise on Geochemistry (Second Edition). Elsevier, Oxford, pp. 395-416. Freimuth, E.J., Diefendorf, A.F. and Lowell, T.V. (2017) Hydrogen isotopes of n-alkanes and n-alkanoic acids as tracers of precipitation in a temperate forest and implications for paleorecords. Geochim. Cosmochim. Acta 206, 166-183. Gagosian, R.B. and Peltzer, E.T. (1986) The importance of atmospheric input of terrestrial organic material to deep sea sediments. Org. Geochem. 10, 661-669. Gao, L., Edwards, E.J., Zeng, Y. and Huang, Y. (2014) Major Evolutionary Trends in Hydrogen Isotope Fractionation of Vascular Plant Leaf Waxes. PLoS ONE 9, e112610. Garcin, Y., Schefuß, E., Schwab, V.F., Garreta, V., Gleixner, G., Vincens, A., Todou, G., Séné, O., Onana, J.-M., Achoundong, G. and Sachse, D. (2014) Reconstructing C3 and C4 vegetation cover using n-alkane carbon isotope ratios in recent lake sediments from Cameroon, Western Central Africa. Geochim. Cosmochim. Acta 142, 482-500. Giri, S.J., Diefendorf, A.F. and Lowell, T.V. (2015) Origin and sedimentary fate of plant-derived terpenoids in a small river catchment and implications for terpenoids as quantitative paleovegetation proxies. Org. Geochem. 82, 22-32. Glover, K.C., Lowell, T.V., Wiles, G.C., Pair, D., Applegate, P. and Hajdas, I. (2011) Deglaciation, basin formation and post-glacial climate change from a regional network of sediment core sites in Ohio and eastern Indiana. Quat. Res. 76, 401-410. Heiri, O., Lotter, A.F. and Lemcke, G. (2001) Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results. J. Paleolimn. 25, 101-110. Hou, J., D’Andrea, W.J., MacDonald, D. and Huang, Y. (2007a) Evidence for water use efficiency as an important factor in determining the δD values of tree leaf waxes. Org. Geochem. 38, 1251-1255. Hou, J., D’Andrea, W.J. and Huang, Y. (2008) Can sedimentary leaf waxes record D/H ratios of continental precipitation? Field, model, and experimental assessments. Geochim. Cosmochim. Acta 72, 3503-3517. Hou, J.Z., D'Andrea, W.J., MacDonald, D. and Huang, Y.S. (2007b) Hydrogen isotopic variability in leaf waxes among terrestrial and aquatic plants around Blood Pond, Massachusetts (USA). Org. Geochem. 38, 977-984. IAEA/WMO (2015). Global Network of Isotopes in Precipitation. The GNIP Database. Accessible at: https://nucleus.iaea.org/wiser. Jacob, J., Huang, Y., Disnar, J.-R., Sifeddine, A., Boussafir, M., Albuquerque, A.L.S. and Turcq, B. (2007) Paleohydrological changes during the last deglaciation in Northern Brazil. Quat. Sci. Rev. 26, 1004-1015.

84 Kahmen, A., Hoffmann, B., Schefuß, E., Arndt, S.K., Cernusak, L.A., West, J.B. and Sachse, D. (2013a) Leaf water deuterium enrichment shapes leaf wax n-alkane δD values of angiosperm plants II: Observational evidence and global implications. Geochim. Cosmochim. Acta 111, 50-63. Kahmen, A., Schefuß, E. and Sachse, D. (2013b) Leaf water deuterium enrichment shapes leaf wax n- alkane δD values of angiosperm plants I: Experimental evidence and mechanistic insights. Geochim. Cosmochim. Acta 111, 39-49. Liu, W., Yang, H. and Li, L. (2006) Hydrogen isotopic compositions of n-alkanes from terrestrial plants correlate with their ecological life forms. Oecologia 150, 330-338. Lutz, B., Wiles, G., Lowell, T. and Michaels, J. (2007) The 8.2-ka abrupt climate change event in Brown's Lake, northeast Ohio. Quat. Res. 67, 292-296. McInerney, F.A., Helliker, B.R. and Freeman, K.H. (2011) Hydrogen isotope ratios of leaf wax n-alkanes in grasses are insensitive to transpiration. Geochim. Cosmochim. Acta 75, 541-554. Meyers, P.A. and Kites, R.A. (1982) Extractable organic compounds in midwest rain and snow. Atmos. Environ. 16, 2169-2175. Nelson, D.B., Knohl, A., Sachse, D., Schefuß, E. and Kahmen, A. (2017) Sources and abundances of leaf waxes in aerosols in central Europe. Geochim. Cosmochim. Acta 198, 299-314. Nelson, D.B., Ladd, S.N., Schubert, C.J. and Kahmen, A. (2018) Rapid atmospheric transport and large- scale deposition of recently synthesized plant waxes. Geochim. Cosmochim. Acta 222, 599-617. Newberry, S.L., Kahmen, A., Dennis, P. and Grant, A. (2015) n-Alkane biosynthetic hydrogen isotope fractionation is not constant throughout the growing season in the riparian tree Salix viminalis. Geochim. Cosmochim. Acta 165, 75-85. Nichols, J., Booth, R.K., Jackson, S.T., Pendall, E.G. and Huang, Y. (2010) Differential hydrogen isotopic ratios of Sphagnum and vascular plant biomarkers in ombrotrophic peatlands as a quantitative proxy for precipitation—evaporation balance. Geochim. Cosmochim. Acta 74, 1407- 1416. Pagani, M., Pedentchouk, N., Huber, M., Sluijs, A., Schouten, S., Brinkhuis, H., Sinninghe Damste, J.S., Dickens, G.R. and Scienctists, t.e. (2006) Arctic hydrology during global warming at the Palaeocene/Eocene thermal maximum. Nature 442, 671-675. Pancost, R.D. and Boot, C.S. (2004) The palaeoclimatic utility of terrestrial biomarkers in marine sediments. Marine Chemistry 92, 239-261. Parnell, A.C., Inger, R., Bearhop, S. and Jackson, A.L. (2010) Source partitioning using stable isotopes: coping with too much variation. PloS one 5, e9672. Parnell, A.C., Phillips, D.L., Bearhop, S., Semmens, B.X., Ward, E.J., Moore, J.W., Jackson, A.L., Grey, J., Kelly, D.J. and Inger, R. (2013) Bayesian stable isotope mixing models. Environmetrics 24, 387-399. Parnell, A. (2016). simmr: A Stable Isotope Mixing Model. R package version 0.3. https://CRAN.R- project.org/package=simmr Pedentchouk, N., Sumner, W., Tipple, B. and Pagani, M. (2008) δ13C and δD compositions of n-alkanes from modern angiosperms and conifers: An experimental set up in central Washington State, USA. Org. Geochem. 39, 1066-1071. Polissar, P., Preefer, M. and Liu, C. (2014) Fidelity of leaf-wax n-alkane and n-alkanoic acid D/H ratios in space and time, American Geophysical Union, pp. PP43E-07. Polissar, P.J. and Freeman, K.H. (2010) Effects of aridity and vegetation on plant-wax δD in modern lake sediments. Geochim. Cosmochim. Acta 74, 5785-5797. R Core Team (2017). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. Rach, O., Brauer, A., Wilkes, H. and Sachse, D. (2014) Delayed hydrological response to Greenland cooling at the onset of the Younger Dryas in western Europe. Nature Geosci 7, 109-112. Rommerskirchen, F., Plader, A., Eglinton, G., Chikaraishi, Y. and Rullkötter, J. (2006) Chemotaxonomic significance of distribution and stable carbon isotopic composition of long-chain alkanes and alkan-1-ols in C4 grass waxes. Org. Geochem. 37, 1303-1332.

85 Sachse, D., Radke, J. and Gleixner, G. (2004) Hydrogen isotope ratios of recent lacustrine sedimentary n- alkanes record modern climate variability. Geochim. Cosmochim. Acta 68, 4877-4889. Sachse, D., Radke, J. and Gleixner, G. (2006) δD values of individual n-alkanes from terrestrial plants along a climatic gradient – Implications for the sedimentary biomarker record. Org. Geochem. 37, 469-483. Sachse, D., Kahmen, A. and Gleixner, G. (2009) Significant seasonal variation in the hydrogen isotopic composition of leaf-wax lipids for two deciduous tree ecosystems (Fagus sylvativa and Acer pseudoplatanus). Org. Geochem. 40, 732-742. Sachse, D., Gleixner, G., Wilkes, H. and Kahmen, A. (2010) Leaf wax n-alkane δD values of field-grown barley reflect leaf water δD values at the time of leaf formation. Geochim. Cosmochim. Acta 74, 6741-6750. Sachse, D., Billault, I., Bowen, G.J., Chikaraishi, Y., Dawson, T.E., Feakins, S.J., Freeman, K.H., Magill, C.R., McInerney, F.A., van der Meer, M.T.J., Polissar, P., Robins, R.J., Sachs, J.P., Schmidt, H.- L., Sessions, A.L., White, J.W.C., West, J.B. and Kahmen, A. (2012) Molecular paleohydrology: Interpreting the hydrogen-isotopic composition of lipid biomarkers from photosynthesizing organisms. Annu. Rev. Earth Planet. Sci. 40, 221-249. Schimmelmann, A., Sessions, A.L. and Mastalerz, M. (2006) Hydrogen Isotopic (D/H) Composition of Organic Matter During Diagenesis and Thermal Maturation. Annu. Rev. Earth Planet. Sci. 34, 501-533. Seki, O., Nakatsuka, T., Shibata, H. and Kawamura, K. (2010) A compound-specific n-alkane δ13C and δD approach for assessing source and delivery processes of terrestrial organic matter within a forested watershed in northern Japan. Geochim. Cosmochim. Acta 74, 599-613. Sessions, A.L., Sylva, S.P., Summons, R.E. and Hayes, J.M. (2004) Isotopic exchange of carbon-bound hydrogen over geologic timescales 1. Geochim. Cosmochim. Acta 68, 1545-1559. Simoneit, B.R.T., Chester, R. and Eglinton, G. (1977) Biogenic lipids in particulates from lower atmosphere over eastern Atlantic. Nature 267, 682-685. Smith, F.A. and Freeman, K.H. (2006) Influence of physiology and climate on δD of leaf wax n-alkanes from C3 and C4 grasses. Geochim. Cosmochim. Acta 70, 1172-1187. Tierney, J.E., Russell, J.M., Huang, Y., Damsté, J.S.S., Hopmans, E.C. and Cohen, A.S. (2008) Northern Hemisphere Controls on Tropical Southeast African Climate During the Past 60,000 Years. Science 322, 252-255. Tierney, J.E., Pausata, F.S.R. and deMenocal, P.B. (2017) Rainfall regimes of the Green Sahara. Science Advances 3. Tipple, B.J., Berke, M.A., Doman, C.E., Khachaturyan, S. and Ehleringer, J.R. (2013) Leaf-wax n- alkanes record the plant–water environment at leaf flush. Proc. Natl. Acad. Sci. U.S.A. 110, 2659-2664. Tipple, B.J. and Pagani, M. (2013) Environmental control on eastern broadleaf forest species’ leaf wax distributions and D/H ratios. Geochim. Cosmochim. Acta 111, 64-77. Welker, J. (2000) Isotopic (δ18O) characteristics of weekly precipitation collected across the USA: an initial analysis with application to water source studies. Hydrological Processes 14, 1449-1464. West, A.G., Patrickson, S.J. and Ehleringer, J.R. (2006) Water extraction times for plant and soil materials used in stable isotope analysis. Rapid Communications in Mass Spectrometry 20, 1317- 1321. Wu, M.S., Feakins, S.J., Martin, R.E., Shenkin, A., Bentley, L.P., Blonder, B., Salinas, N., Asner, G.P. and Malhi, Y. (2017) Altitude effect on leaf wax carbon isotopic composition in humid tropical forests. Geochim. Cosmochim. Acta 206, 1-17. Yang, H. and Huang, Y. (2003) Preservation of lipid hydrogen isotope ratios in Miocene lacustrine sediments and plant fossils at Clarkia, northern Idaho, USA. Org. Geochem. 34, 413-423.

86

Figure 1. Maps of Brown’s Lake Bog State Nature Preserve. Panel A shows the northwestern preserve boundary (dashed line), the bog and adjacent lake, and sites of tree sampling (squares), Z. mays sampling (diamond) and precipitation collection (triangle). Panel B shows detail of the area inside the white box (from panel A), including the bog shoreline with the locations of shoreline vegetation sampling (circles) and the site of bog surface core collection (star).

87

Figure 2. Age-depth model for BLB based on 210Pb ages in yr before core collection (2016 CE). Plant lipids analyzed in this study were sampled from the upper 40 cm of sediment (indicated with dashed line), which corresponds to 1998-2015 CE.

88

Figure 3. The hydrogen isotopic composition of plant xylem waters (A) and environmental waters (B) at BLB. Plant waters were sampled in September 2016 from C3 graminoids, herbs, shrubs and woody groundcover in the bog shoreline zone, as well as from trees in the forest. Bog water (squares) and precipitation (diamonds) are plotted by day of year (DOY) of collection in 2014 (white fill), 2015 (red fill) and 2016 (blue fill). The OIPC-modeled monthly and mean annual precipitation δD values are indicated with black circles and the horizontal dashed line, respectively. The duration of the growing season (from bud break to leaf fall) is indicated by the shaded area.

89

Figure 4. Mean lipid abundance for each plant growth form, forest litter, and bog sediments at BLB, for n-C26 to n-C30 alkanoic acids (A) and n-C27 to n-C35 alkanes (B). n-Alkane concentrations are plotted on a log scale for clarity.

90

Figure 5. Distribution of chain lengths for different plant growth forms, forest litter, rain-scavenged aerosols, and bog sediments at BLB, for A) n-alkanoic acids (ACL26-30) and B) n-alkanes (ACL27-35).

91

Figure 6. The δD and δ13C composition of n-alkanes (A) and n-alkanoic acids (B) extracted from each plant growth form, forest litter, rain-scavenged aerosols, and bog sediments at BLB.

92

13 Figure 7. Plots of the δD and δ C values of n-C29 alkane (A) and n-C28 alkanoic acid (B) in sediments and plant growth forms at BLB. Each point represents the mean isotope values for a different species, based on data from replicate individuals sampled from each species. The size of each point is scaled to the lipid concentration in µg/g dry leaf for plants and µg/g dry sediment for bog sediments. Note different bubble size scales for panels A and B. Points are color coded by growth form. Points representing bog sediments are black. Concentration-weighted isotope values for all source vegetation are plotted as open squares. The isotopic composition of pooled rain-scavenged aerosol n-alkanes is indicated with a cross in panel A.

93

13 Figure 8. Regressions of the δD and δ C values of n-C29 alkanes (left panel) and n-C28 alkanoic acids

(right panel) from individual shrubs (pink circles) and trees (green circles).

94

Table 1. Mean n-alkane and n-alkanoic acid ACL and concentrations for plant species and forest litter (in µg/g dry leaf) as well as for the upper 40 cm of bog sediment (in µg/g dry sediment). Unidentified C3 grass and herbaceous plant species are numbered with laboratory identifiers. Compounds that were not detected by GC-MS are indicated with n.d.

ACL n-Alkane (ug/g dry mass) n-Alkanoic acid (ug/g dry mass) Species # Species Sample Type n-C27-35 alkane n-C26-30 acid n-C27 n-C29 n-C31 n-C33 n-C35 n-C26 n-C28 n-C30 10 Rhynochospora capitellata C3 Grass 27.9 28.8 33.9 16.4 3.6 n.d. n.d. n.d. 10.0 6.6 11 Scirpus cyperinus C3 Grass 29.3 28.9 2.1 3.5 2.8 n.d. n.d. n.d. 7.3 6.5 3 C3 Grass 30.0 28.5 2.3 11.0 12.3 2.1 n.d. 1.0 15.5 7.6 NA Zea mays C4 Grass 31.4 29.0 2.6 14.3 26.6 23.8 6.0 n.d. 5.6 6.0 1 Vaccinium corymbosum Shrub 30.2 29.4 2.5 52.5 66.9 5.0 n.d. 0.2 8.9 23.3 7 Toxicodendron vernix Shrub 29.8 28.9 7.5 19.5 17.4 4.7 n.d. 2.4 14.7 17.9 NA Acer saccharinum Tree 29.2 27.6 106.6 289.0 117.4 17.5 n.d. 33.5 34.4 17.2 NA Prunus serotina Tree 31.0 27.3 16.9 731.0 1867.7 692.6 15.3 154.5 28.5 65.1 NA Quercus alba Tree 28.7 27.6 28.7 34.1 13.7 1.6 n.d. 33.1 30.1 18.0 NA Quercus rubra Tree 30.0 28.0 112.2 676.5 807.1 89.3 n.d. 71.3 36.3 30.5 NA Ulmus americana Tree 29.3 27.2 6.3 64.3 15.0 1.4 n.d. 180.5 31.3 64.5 15 Vaccinium macrocarpon Woody groundcover 30.4 29.4 6.1 59.5 123.1 8.5 n.d. n.d. 3.3 7.3 2 Salix petiolaris Woody groundcover 28.6 28.9 6.4 11.9 2.3 n.d. n.d. 1.0 12.7 12.6 13 Herb 29.9 28.7 11.0 25.1 43.6 5.0 n.d. 1.1 32.0 18.5 14 Asclepias syriaca Herb 28.5 28.9 5.2 7.4 1.4 n.d. n.d. n.d. 10.0 7.5 4 Peltandra virginica Herb 27.8 28.6 6.4 4.1 n.d. n.d. n.d. 8.3 42.6 32.0 5 Herb 28.8 29.3 0.3 2.4 n.d. n.d. n.d. 0.3 7.8 14.1 6 Herb 28.9 29.0 4.1 23.1 4.6 n.d. n.d. 9.7 35.2 63.9 8 Sphagnum sp. Herb 28.5 28.2 5.7 3.2 2.8 n.d. n.d. 2.5 3.8 3.3 9 Herb 29.5 29.4 2.5 40.9 19.8 n.d. n.d. n.d. 3.7 8.9 Forest Litter 30.2 27.0 24.4 325.6 733.6 257.2 6.2 87.9 22.5 22.3 Sediment 29.3 27.4 11.9 16.7 9.4 4.1 0.4 11.6 5.4 3.8

95 Table 2. Rain-scavenged particulate wax sampling dates with filtered rainwater volume, ACL and n-alkane concentration (in µg/liter of rainwater) for each sample. Residual samples were from solvent-rinsed carboys (see Section 2.3). Compounds that were not detected by

GC-MS are indicated with n.d.

n-alkane concentration (ug/liter) Date Start Date Stop Collector Filtered vol. (liters) ACL (n-C27-31) n-C21 n-C23 n-C25 n-C27 n-C29 n-C31 7/22/16 8/20/16 A 1.6 29.7 0.24 0.24 n.d. n.d. 0.33 0.18 7/22/16 8/20/16 B 2.3 29.0 n.d. n.d. n.d. n.d. 0.35 n.d. 8/20/16 9/22/16 A 4.3 29.5 n.d. n.d. n.d. 0.03 0.08 0.07 8/20/16 9/22/16 B 3.4 30.1 n.d. n.d. n.d. n.d. 0.08 0.09 9/22/16 11/1/16 A 5.5 29.6 n.d. n.d. 0.02 0.06 0.21 0.21 9/22/16 11/1/16 B 10.8 29.5 n.d. n.d. n.d. 0.05 0.18 0.13 11/1/16 12/31/16 A 4.0 29.4 n.d. n.d. 0.06 0.15 0.42 0.31 11/1/16 12/31/16 B 4.2 29.3 n.d. 0.11 0.09 0.19 0.50 0.32 residual residual A NA 29.5 n.d. n.d. 0.36 0.91 2.75 2.27 residual residual B NA 29.6 n.d. n.d. n.d. 0.50 2.22 1.80

96 Table 3. Summary isotope data (mean, 1s, n) for plant xylem water and wax δD (‰ VSMOW) and δ13C (‰ VPDB), with corresponding fractionation relative to xylem water (εwax-xyl) and mean annual precipitation (εwax-MAP), by species. Mean data for each growth form, particulate wax and bog sediments are in bold text.

δDn-alkane δDn-acid δ13Cn-alkane δ13Cn-acid Ewax-xyl Ewax-MAP Species Sample Type δDxw 1s n n-C29 1s n n-C31 1s n n-C28 1s n n-C29 1s n n-C31 1s n n-C28 1s n n-C29 n-C31 n-C28 n-C29 n-C31 n-C28 3 C3 Grass -18 1 -186 -37.3 -37.7 -37.2 -171 -143 R. capitellata C3 Grass -9 1 -196 -181 -35.4 -36.0 -36.2 -189 -173 -154 -138 S. cyperinus C3 Grass -35 3 2 -180 0.0 -33.2 -150 -137 C3 Grass -24 13 4 -196 -182 3 3 -36.3 1.4 2 -36.9 1.3 2 -35.5 2.1 3 -189 -165 -154 -139 Z. mays C4 Grass -158 -165 -101 -19.8 -20.0 -24.5 -158 -165 -101 -114 -121 -54 5 Herb -34 1 -162 14 2 0.0 -132 -118 6 Herb -18 1 -160 11 2 -152 1 2 -38.2 -37.4 -39.0 0.01 2 -144 -137 -115 -107 9 Herb -160 -151 -41.4 -42.2 -160 -151 -116 -107 13 Herb -178 7 2 -162 7 2 -141 14 2 -38.3 0.4 2 -39.3 0.4 2 -37.5 0.5 2 -178 -162 -141 -135 -118 -96 A. syriaca Herb -163 0.0 -39.4 -163 -119 P. virginica Herb -37 1 -170 -146 6 2 0.0 -40.2 0.9 2 -138 -113 -127 -101 Sphagnum sp. Herb -22 4 3 0.0 Herb -30 8 9 -168 11 6 -158 8 3 -152 11 9 -38.9 1.4 5 -39.6 2.0 4 -39.0 1.2 7 -157 -158 -134 -124 -114 -107 T. vernix Shrub -27 6 6 -167 2 4 -171 8 4 -169 11 4 -33.2 1.8 4 -34.1 1.8 4 -33.8 0.5 4 -143 -147 -146 -123 -127 -125 V. corymbosum Shrub -32 5 5 -185 6 3 -161 3 3 -136 2 3 -36.2 1.3 3 -38.1 0.9 3 -37.7 1.3 3 -158 -133 -107 -142 -117 -90 Shrub -29 3 7 -175 11 7 -166 8 7 -155 20 7 -34.5 2.2 7 -35.8 2.6 7 -35.5 2.3 7 -150 -141 -129 -131 -123 -110 A. saccharinum Tree -40 -168 -137 -34.7 -33.2 -33.7 -133 -101 -124 -92 P. serotina Tree -42 -160 -164 -150 -33.0 -32.6 -31.9 -123 -128 -113 -115 -120 -105 Q. alba Tree -41 -165 -137 -34.7 -35.3 -34.3 -130 -100 -121 -92 Q. rubra Tree -37 1 -155 2 2 -154 3 2 -144 4 2 -32.1 0.5 2 -33.8 -34.8 0.6 2 -123 -122 -111 -111 -110 -99 U. americana Tree -62 -168 -147 -126 5 2 -34.4 2.0 2 -33.6 1.5 2 -38.0 0.7 2 -113 -147 -98 -125 -102 -80 Tree -43 9 6 -162 6 6 -155 7 4 -138 9 7 -33.6 1.5 7 -33.7 1.0 7 -35.1 2.2 7 -124 -130 -104 -118 -110 -92 S. petiolaris Woody groundcover -22 1 -203 -160 12 2 -37.4 0.4 2 -39.6 0.5 2 -38.0 0.3 2 -185 -140 -161 -115 V. macrocarpon Woody groundcover -17 -229 -219 -34.0 -35.0 -216 -205 -188 -178 Woody groundcover -21 3 4 -216 18 2 -219 -160 12 2 -36.3 1.9 3 -38.0 2.7 3 -38.0 0.3 2 -200 -205 -140 -175 -178 -115 Particulate Wax -167 -156 -30.8 -31.1 -123 Bog Sediment -176 3 10 -169 3 10 -148 5 6 -32.9 0.2 11 -32.8 0.2 11 -33.9 0.4 10 -133 -125 -103

97 Chapter 4

Comparative fidelity of n-alkanes and n-alkanoic acids to regional-scale precipitation δD and catchment-scale vegetation: a survey of lake sediments in the Adirondack Mountains,

NY (USA)3

ABSTRACT

Plant wax hydrogen isotopes (δDwax) are important archives of water isotopes in the geologic past. Plant waxes are common in lacustrine sediments, providing widely distributed terrestrial records of hydrologic change in locations where other water isotope proxies (e.g., ice sheets or cave deposits) do not exist.

However, application of δDwax is subject to large uncertainties regarding the biological source of these compounds as well as alterations of the original water isotope signal during their transfer from plant to ecosystem to sediment. Investigating modern controls on this proxy system can address these uncertainties. Here we investigate transfer of the precipitation isotope signal (δDp) to lake sediment δDwax at the regional scale in the Adirondack Mountains of northern New York, USA. The goals of this study are to 1) quantify the variability in sediment δDwax in the absence of major differences in δDp among sites within the same region; 2) identify possible catchment-level drivers of this variability (e.g., vegetation cover, fluvial complexity, depositional setting); and 3) compare the molecular and δD composition of two plant wax compounds (n-alkanes and n-alkanoic acids) to assess potential differences in their sources or apparent fractionation (εapp) relative to δDp. We evaluate plant waxes in surface sediments from 12

Adirondack lakes that have similar δDp but vary in their shoreline vegetation structure, basin size, and

3 Freimuth, E.J., Diefendorf, A.F., Lowell, T.V., Bates, B.R., Schartman, A., Bird, B., Landis, J., Stewart, A.K. Comparative fidelity of n-alkanes and n-alkanoic acids to regional-scale precipitation δD and catchment-scale vegetation: a survey of lake sediments in the Adirondack Mountains, NY (USA). Geochimica et Cosmochimica Acta (in prep.)

98 fluvial complexity. We also measured xylem water from common plants in seven of the 12 lake catchments to confirm that plant source water δD reflects summer-biased δDp and does not differ significantly among sites. We found a larger range in n-alkanoic acid δD than n-alkane δD (35‰ and

22‰, respectively) and larger εapp for n-alkanes than for n-alkanoic acids (-132 ± 6‰ and -106 ± 10‰, respectively). Comparison with comparable studies in multiple biomes suggests that εapp from n-alkanes in forested catchments is stable regardless of catchment-scale aridity, which increases confidence in estimates of εapp for interpretation of older sediments. The δD offset between compound classes was highly variable among sites and strongly correlated with sediment C/N ratios, which may indicate an aquatic source for long-chain n-alkanoic acids. In addition, the relative abundance of evergreen conifer vegetation in lake shorelines correlated with the abundance of long-chain n-alkanes relative to n-alkanoic acids, which may indicate coniferous sources for n-alkanoic acids. These results underscore the influence that site-specific factors can have on δDwax records, and the importance of constraining vegetation and depositional setting, when possible.

99 1. Introduction

13 The hydrogen and carbon isotopic compositions of plant waxes (δDwax and δ Cwax, respectively) have become important tools for reconstructing paleovegetation and paleohydrology (Pancost and Boot, 2004;

Castañeda and Schouten, 2011; Sachse et al., 2012; Freeman and Pancost, 2014; Sessions, 2016). Plants synthesize leaf waxes using hydrogen derived from plant source waters (including soil, stream and lake water), which are fed by precipitation. Therefore, δDwax reflects the δD composition of the precipitation

(δDp) that fed a plant’s source water pool (Sachse et al., 2006; Feakins and Sessions, 2010; Kahmen et al.,

2013a; Kahmen et al., 2013b; Tipple and Pagani, 2013). The relationship between δDp and δDwax persists from the plant-level to both marine and lacustrine sediments, which can serve as tools for reconstructing plant source water (i.e., precipitation) throughout the Cenozoic (e.g., Pagani et al., 2006; Tierney et al.,

2008; Niedermeyer et al., 2014).

Molecular paleohydrology relies upon estimates of the net apparent fractionation (εapp) between plant source water and the resulting leaf wax. While εapp is critical for converting δDwax measured in sediments to ancient δDp, estimating εapp can be challenging because it varies not only at the global level among growth forms and taxonomic groups (McInerney et al., 2011; Sachse et al., 2012), but also locally (by up to ~100‰) among species growing at the same site (Hou et al., 2007; Feakins and Sessions, 2010; Eley et al., 2014). Processes controlling εapp in plants include evaporative D-enrichment of plant source water during soil evaporation (McInerney et al., 2011) and leaf transpiration (Feakins and Sessions, 2010;

Sachse et al., 2010; Kahmen et al., 2013a), as well as fractionation during biosynthesis (εbio), which is likely constant within a species (Sessions et al., 1999; Sachse et al., 2012), but has been observed to vary by 50‰ to 65‰ among species in natural settings and growth chambers (Kahmen et al., 2013a; Kahmen et al., 2013b). While much recent progress has been made in understanding controls on εapp in modern plants, the processes of catchment-wide mixing, integration and transport that determine εapp values and variability in sediments remain poorly understood (Sachse et al., 2012).

100

One of the main advantages of applying the δDwax proxy in lake sediments is that lakes can provide records of paleohydrologic conditions at exceptionally high temporal (~sub-decadal) and spatial (i.e., a single lake catchment) resolution (e.g., Feakins et al., 2014; Rach et al., 2014). However, these same advantages of lacustrine δDwax records may also make them vulnerable to a high degree of uncertainty because these terrestrial plant-derived proxies may be expressed non-uniformly due to sourcing from vegetation assemblages unique to each lake catchment. For example, δDwax values can vary by as much as

60‰ in the sediments of multiple lakes in close proximity as a result of differences in catchment veegtation or local aridity effects (Douglas et al., 2012; Daniels et al., 2017). Therefore, it is difficult to estimate εapp with a high degree of confidence when reconstructing δDp using lake sediments. This is in distinct contrast to the case for marine sediments, where effectively constant εapp values (± ~5‰) are observed in nearshore surface sediments spanning large gradients in aridity and vegetation composition

(Vogts et al., 2016). The stability of εapp in marine sediments may be due to a high degree of signal averaging, both spatially (by large riverine networks and aeolian transport) and temporally (with considerably lower sedimentation rates than those typically observed in lakes).

Therefore, this study seeks to determine the variability in lake sediment δDwax (εapp) that occurs in the absence of major differences in δDp. Our approach was to survey δDwax in multiple lakes in the same region that receive the same δDp but vary markedly in shoreline and catchment vegetation composition, lake to watershed area ratio, fluvial complexity and bulk sediment characteristics (organic and inorganic).

The Adirondack Mountains, NY, USA are uniquely suited to this approach due to the abundance of undeveloped lakes with diverse depositional characteristics and vegetation structure. In recent decades,

Adirondack lakes have been the subject of comprehensive regional water quality monitoring (Driscoll and

Newton, 1985; Driscoll et al., 1995) as well as several paleohydrological and paleovegetation studies

(Charles et al., 1990; Whitehead et al., 1990; Stager et al., 2016). We seek to assess the δDwax signal at the sediment level throughout the Adirondacks and identify any site-specific factors that may account for

101 signal variability. Further, we aim to compare the regional sediment-level εapp values from Adirondack temperate mixed forests to the sediment-level εapp values observed in other biomes. This is the first in a series of studies that together aim to identify catchment- and biome-specific factors that investigators can use to guide site selection to reduce ‘noise’ (resulting from depositional setting) and thereby increase

(statistical) confidence in reconstructions of δDp based on δDwax in lake sediments.

2. Materials and methods

2.1. Sampling locations and sediment collection

This project is focused on the Adirondack Park, NY, USA, which is the largest state park in the conterminous USA (24,000 km2; 37 to 1629 m above sea level) with approximately 2,800 lakes covering

1,030 km2 and 13,000 km of streams (Driscoll et al., 1991). The Adirondacks are a transitional vegetation zone between eastern deciduous angiosperm and mixed angiosperm-conifer forest with the distribution of vegetation assemblages dependent primarily on elevation. Adirondack climate is humid continental.

Long-term (1981-2010) mean temperature and precipitation data obtained from the National Climatic

Data Center (NCDC) for three stations in the Adirondack Park (between 507 and 533 m asl) range from minimum temperatures in January (-10.1 °C) to maximum temperatures in July (17.9 °C). Monthly precipitation rates peak between July and October (106 and 127 mm). February precipitation minima at each station ranged from 41 to 78 mm. The long-term mean annual precipitation amount ranged from 953 to 1269 mm. Across the Adirondack Park, precipitation amount generally increases from the northeast to the southwest, with the amount of spring precipitation dependent on latitude, and summer precipitation dependent primarily on elevation (Ito et al., 2002).

102 The sediment cores used in this study were collected from 12 lakes within the Adirondack Park (Fig. 1).

All samples were collected on public lands designated as wilderness or wild forest areas, with the exception of Heart and Wolf lakes, which are privately owned by the Adirondack Mountain Club and the

State University of New York, respectively. Details of the different sampling sites are reported in Table 1.

The sampled lakes spanned elevations from 426 to 665 m asl, ranged in area from 0.1 to 7.5 km2 and had fluvial inlet/outlet ratios from 0 to 4. Chazy and Raquette lakes are the only dammed sites included in this study. While Chazy was originally a lake, Raquette is a dammed river and therefore the sediment delivery and plant lipid sourcing may not be comparable to other sites, which are all much smaller and have less developed land area within their catchments (Table 1; Fig. 2). Based on data from the National Land

Cover Database (NLCD; Homer et al., 2015), the 12 lake shorelines have fractional evergreen coverage ranging from 0.07 to 0.79 with the remaining vegetation comprised of deciduous or mixed forest and woody wetlands (Fig. 2b). Lake catchments are generally dominated by deciduous angiosperm trees with remaining contributions from evergreen conifer or mixed forest (Fig. 2a).

Sediments were collected in October 2016 from an anchored boat and in January 2017 from lake ice. The upper ~1.0 meter of sediment was recovered from the deepest part of each lake (with the exception of

Chazy and Raquette; Table 1) using a 1.5 m polycarbonate core tube and Bolivia piston corer. An underwater camera was deployed to ensure that the sediment-water contact was captured during core recovery. Headspace water was drained and the core tube was trimmed before capping. Cores were transported and stored vertically at 4 °C until analysis.

A subset of the cores from Heart, Moose, Wolf, Debar, and East Pine were dated using 210Pb by gamma counting at the Dartmouth GeoAnalytical Resource Center, Hanover, NH, USA. Based on these results, the age at 10 cm depth ranged from calendar year 1960 at Wolf to 2004 at Heart. We therefore estimate that the intervals used for plant wax analysis (the upper 10 cm, see Section 2.3) represent approximately

103 the last 50 yr. This is in good agreement with 210Pb results from over 50 lakes throughout the Adirondacks

(Binford, 1990; Binford et al., 1993).

2.2. Bulk sediment characterization

Each core was scanned for magnetic susceptibility using a Barington MS2 meter. Percent total organic matter and total carbonate were determined based on the mass loss-on-ignition (LOI) of 1 cm3 of sediment sampled every 2 cm after drying (60 °C overnight) and heating at 550 °C (4 h) and 1000 °C (2 h; Heiri et al., 2001). Approximately 1 cm3 of sediment was taken from each core at 4, 8, and 12 cm and at 10 cm intervals from 20 cm to the base of the core for grain size analysis. Sample pre-treatment followed Gray et al. (2010) and Bird et al. (2014). Organic matter was removed using 30% H2O2 at room temperature for 24 h followed by three repeat 24 h treatments with 30% H2O2 in a 70 ºC water bath.

Biogenic silica was removed using 1M NaOH at 60 ºC for 6 h. Lastly, carbonates were removed using

1M HCl at room temperature. Sediments were freeze dried and weighed to determine the percent lithics based on the mass loss after pretreatment. Grain size particle distribution was determined using a

Beckman-Coulter LS230 instrument.

A separate aliquot of decarbonated (1M HCl until reaction completion, followed by DI water rinses until neutralized), dried and homogenized sediment was used to determine bulk sediment total organic carbon

13 13 (TOC), total nitrogen (TN), and δ Corg. Bulk δ C analysis was performed via continuous flow (He; 120 ml/min) on a Costech elemental analyzer (EA) interfaced with an IRMS via a Conflo IV. Raw δ13C values were corrected for sample size dependency and normalized to the VPDB scale using a two-point calibration with in-house standards which were calibrated with IAEA standards (NBS-19, L-SVEC) to -

38.26‰ and -11.35‰ VPDB following Coplen et al. (2006). Error was determined by analyzing two additional independent standards with a precision of 0.11‰ (1σ, n = 34) and accuracy of 0.09‰ (n = 34).

Carbon/nitrogen (C/N) ratios were determined based on the weight % of TOC/TN.

104

2.3. Lipid extraction, separation and quantification

The uppermost sediments from each lake were subsampled at three 0.5 cm intervals (1.0 to 1.5, 5.0 to 5.5, and 9.0 to 9.5 cm depth) for a total of 36 samples, which were transferred to glass jars and freeze dried in preparation for lipid analysis. Dried sediment was homogenized with mortar and pestle prior to accelerated solvent extraction (ASE; Dionex 350) with 9:1 (v/v) DCM/MeOH with three extraction cycles at 100 ºC and 10.3 MPa. The total lipid extract was saponified with 3 ml of 0.5 N KOH in MeOH/H2O

(3:1, v/v) for 2 h at 75 ºC. Once cool, 2.5 ml of NaCl in water (5%, w/w) were added and acidified with

6N HCl. The solution was extracted with hexanes/DCM (4:1, v/v), neutralized with NaHCO3/H2O (5%, w/w), and water was removed through addition of Na2SO4. Neutral and acid fractions were separated using DCM/IPA (2:1, v/v) and 4% formic acid in diethyl ether, respectively, over aminopropyl-bonded silica gel. The neutral fraction was then separated into aliphatic and polar fractions using activated alumina oxide column chromatography. The aliphatic fraction was separated over 5% silver impregnated silica gel with hexanes to collect saturated compounds. The acid fraction was evaporated under a gentle stream of nitrogen and methylated by adding ~1.5 ml of 95:5 MeOH/12 N HCl (v/v) of known δD composition and heating at 70 ºC for 12-18 h. HPLC grade water was added and the methylated acid fraction was extracted with hexanes and eluted through Na2SO4 to remove water before separation over

5% deactivated silica gel with DCM to collect fatty acid methyl esters (FAMEs).

n-Alkanes and FAMEs were identified by GC-MS using an Agilent 7890A GC and Agilent 5975C quadrupole mass selective detector system, and quantified using a flame ionization detector (FID).

Compounds were separated on a fused silica column (Agilent J&W DB-5ms) and the oven ramped from an initial temperature of 60 ºC (held 1 min) to 320 ºC (held 15 min) at 6 ºC/min. Compounds were identified using authentic standards, fragmentation patterns and retention times. All samples were diluted in hexanes spiked with 1,1’-binaphthyl, as an internal standard, prior to quantification. Compound peak

105 areas were normalized to those of 1,1’-binaphthyl and converted to concentration using response curves for an in-house mix on-n-alkanes (even-chains from C14 to C18; odd-chains from C25 to C35) and FAMEs

(C16, C18, C24, C28, C30) at concentrations ranging from 0.5 to 100 µg/ml. Precision and accuracy were determined by analyzing external standards at 25 µg/ml as unknowns and were 0.61 (1s, n = 13) and 0.41

(1s, n = 13), respectively. Quantified concentrations were normalized to the mass of dry sediment or leaf material extracted and to the equivalent mass of TOC extracted. FAME concentrations were converted to equivalent n-alkanoic concentrations using respective molar masses.

Average chain length (ACL) is the weighted average concentration of all the long-chain waxes, defined as

( )( ACL = & (Eq. 1) $%& (*$ )( where a-b is the range of chain lengths and Ci is the concentration of each wax compound with i carbon atoms. ACL27-35 is used to indicate ACL values for n-C27 to n-C35 alkanes (odd chain-lengths only).

ACL26-30 is used to indicate ACL values for n-C26 to n-C30 alkanoic acids (even chain-lengths only). The preference for odd n-alkane chain lengths was assessed using the carbon preference index (CPI), determined following (Marzi et al., 1993).

2.4. δD analysis of n-alkanes and n-alkanoic acids

The δD composition of n-alkanes and n-alkanoic acids was determined using a Thermo Trace GC Ultra coupled to an Isolink pyrolysis reactor (1420 ºC) and interfaced to a Thermo Electron Delta V Advantage

IRMS via a Conflo IV. The GC oven program ramped from 80 ºC (held 2 min) to 320 ºC (held 15 min) at

+ a rate of 8 ºC/min. The H3 factor was tested daily and averaged 4.8 ppm/mV during the period of analysis. Data were normalized to the VSMOW/SLAP scale using a two-point calibration with a standard of known δD composition (Mix A5; A. Schimmelmann, Indiana University). The long-term precision of a co-injected n-C38 alkane standard was 2.9‰ (1s, n = 40). The δD value of hydrogen added to n-alkanoic

106 acids during derivatization was determined by mass balance of phthalic acid of known δD composition

(A. Schimmelmann, Indiana University) and derivatized methyl phthalate. Hydrogen isotopic fractionations (εapp, εbio, etc.) are reported as enrichment factors (in units of per mil, ‰), using the following equation:

(-./01) +$%& = − 1 (Eq. 2) (-.301) where a is the product and b is the substrate.

2.5. Water sampling and isotope analysis

Surface water samples (Table 2) were collected from lakes and rivers throughout the Adirondacks in July and October 2016 and in January 2017 using 30 ml Nalgene bottles. At the time of sampling, water temperature and conductivity were measured and the sampling location was determined using GPS.

Collection of plant material for stem water isotope analysis (n = 58) occurred in May 2017. Samples were collected from deciduous angiosperm (n = 4) and evergreen gymnosperm (all conifers; n = 7) species at seven of the 12 sites where sediment was collected (Table 3). Plant sampling methods followed Freimuth et al. (2017). Two stems (<1.0 cm diameter) were clipped from each individual, stripped of outer bark, and sealed in separate pre-ashed glass exetainer vials with airtight septa screwcaps. All samples were kept in a cooler in the field and frozen (-5 ºC) in the lab until water extraction. Each sample was treated and analyzed separately.

Stem (xylem) water from each sample was extracted using cryogenic vacuum distillation following West et al. (2006). Exetainer vials containing frozen stems were evacuated to a pressure <8 Pa (<60 mTorr), isolated from the vacuum pump, and heated to 100 ºC. Water vapor was collected in borosilicate test tubes immersed in liquid nitrogen for a minimum of 60 minutes. To verify extraction completion, samples were weighed following cryogenic vacuum distillation and then again after freeze drying. Based on the

107 mass difference, the recovery of plant water was >99%. Collected water was thawed and pipetted into 2 ml crimp-top vials and refrigerated at 4 ºC until analysis.

Analysis of water d18O and dD was made by headspace equilibration. For d18O analysis, 500 µl of water were transferred to an exetainer vial, stoppered, and purged using an offline purging manifold system with 0.3% CO2 in He for 10 minutes at 120 ml/min (Paul and Skrzypek, 2006). Samples were then equilibrated in a water bath at 25ºC for 18 hours. For dD analysis, 200 µl of water were transferred to an exetainer vial, and a Pt catalyst was added. Samples were purged using 2% H2 in He for 10 minutes at 120 ml/min and equilibrated at 25ºC for 1 hour. The isotopic composition of equilibrated headspace gases was analyzed on a Thermo Delta V Advantage isotope ratio mass spectrometer (IRMS) with a Thermo

Gasbench II connected to the IRMS via a Conflo IV interface. Data were normalized to the

VSMOW/SLAP scale (Coplen, 1988) using three in-house standards. For surface waters, precision and accuracy were 2.2‰ (1s, n = 22) and 2.3‰ (n = 2) for δD, respectively, and 0.15‰ (1s, n = 19) and

0.3‰ (n = 19), respectively, for δ18O. For plant waters, precision and accuracy were 1.6‰ (1s, n = 20) and 2.5 ‰ (n = 20), respectively, for δD and 0.3‰ (1s, n = 22) and -0.1‰ (n = 22), respectively, for

δ18O.

We calculated deuterium excess (d-excess) as a measure of the extent of evaporation that has affected the isotopic composition of each water sample using the following equation (Dansgaard, 1964):

d-excess = δD - 8 * δ18O (Eq. 3)

The average d-excess value for precipitation is 10. As environmental waters undergo evaporation (kinetic fractionation), d-excess values typically decrease to values below 10 (Fritz and Clark, 1997). We also calculated line-conditioned excess (lc-excess*) values for Adirondack water samples to characterize their

108 offset from the local meteoric water line (LMWL) based on monthly precipitation collection at a GNIP station in Ottawa ~100 km from the Adirondack Park (IAEA/WMO, 2018). The equation for lc-excess* follows Landwehr and Coplen (2006):

lc-excess* = [δD – a δ18O – b] / S (Eq. 4)

where a and b are the slope and intercept of the LMWL (7.57 and 6.72, respectively). S is the one standard deviation measurement uncertainty based on measurement uncertainties at the University of

Cincinnati (1.67 for δD and 0.07 for δ18O), where:

S = [1.672 + (a * 0.07)2]0.5 (Eq. 5)

Data reduction and analysis was completed using JMP Pro 12.0 (SAS, Cary, USA). We assign significance for alpha levels at or below 0.05 and parametric tests are performed on normally distributed data.

3. Results

3.1. Bulk sediment characteristics

The average total organic matter (%TOM) determined by loss-on-ignition from the upper 50 cm of sediment ranged from 14 to 68% among all 12 sites, with most sites between 20 to 40%. The total carbonate (%TC) was less than 6% in all cores. Magnetic susceptibility was consistently low (< 5.3 SI units) across all cores. Grain size distributions in the upper 50 cm of sediments varied widely, with most sites dominated by clay or silt (Fig. 3).

109

The C/N ratio in all of the sediments ranged from 10 to 14, indicating the dominance of aquatic sources of sedimentary organic matter, with varying contributions from vascular land plants (Meyers, 2003; Table

1). Across all sites, we observed a 25% range in TOC (from 10% at Chazy to 35% at Quiver), and a ~5‰

13 range in δ Corg (from -31.9‰ at Quiver to -26.8‰ at Chazy; Table 1). There was an inverse relationship

13 2 13 between TOC and δ Corg values among lakes (R = 0.42, p < 0.0001; Fig. 4). Particularly low δ Corg at several sites (< -30‰; Fig. 4) may reflect either reduced algal productivity or incorporation of 13C- depleted DIC from the heterotrophic respiration of sedimentary organic matter (Meyers, 2003; Diefendorf et al., 2008). Lower TOC in Chazy and Raquette lakes, which are dammed, may reflect an apparent dilution of organics with relatively high proportions of sand and silt in the sediment grain size distribution at those sites (Fig. 3).

3.2. Adirondack isotope hydrology

3.2.1. Surface and meteoric water isotopes

Adirondack lake waters sampled between July 2016 and January 2017 had δD values ranging from -93 to

-43‰ and δ18O values from -12.0 to -5.1‰ (Fig. 5). The mean δD and δ18O values of lakes sampled in

August and October of 2016 (-61.9 and -8.0‰, respectively) were higher than in January 2017 (-76.7‰ and -9.0‰, respectively). In the summer and autumn, there was a small but significant D- and 18O- enrichment in closed basins relative to open basins (t-test, p = 0.026 and 0.001, respectively; Fig. 6). By contrast, in the winter closed basins were significantly D-depleted compared to open basins, but there was no difference in δ18O between basin types (t-test, p = 0.022 and 0.25, respectively; Fig. 6). Lake waters sampled in this project were comparable to those reported for different Adirondack lakes sampled in

September 2007 (-69‰, δD and -9.2‰, δ18O; Brooks et al., 2014). The lowest d-excess values we observed were from lakes sampled in the summer/autumn (1.7) and winter (-4.3). Summer/autumn lake

110 waters (combined due to similar climate and precipitation isotopes in July and October) were in close agreement with mean d-excess values measured by Brooks et al. (2014) from Adirondack lakes in

September (4.7; p = 0.3). These low (< 10) values indicate Adirondack lakes undergo significant evaporation (kinetic fractionation) during summer months.

The summer and autumn river waters we sampled ranged from -83 ‰ to -48‰ for δD and from -11.6‰ to -5.6‰ for δ18O, overlapping with lake water values across all seasons (Fig. 6). Six different rivers sampled within 60 km west and southeast of Adirondack Park in the 1980s had peak δD and δ18O values in August and October; lowest values in March and April (Kendall and Coplen, 2001). When the data are limited to August through October, the mean isotope values for the rivers we sampled in this project (n =

10) and those reported by Kendall and Coplen (2001; n = 34) are indistinguishable (δD, -65 ± 11‰ and -

66 ± 4‰) and (δ18O, -9.0 ± 1.5‰ and -9.4 ± 0.5‰). The mean d-excess value from rivers measured in this study (9.2) was significantly higher than lakes (p = 0.0001), and was in close agreement with d- excess measured in rivers within 60 km of the Adirondack Park (11.3; p = 0.28; Kendall and Coplen,

2001). Overall, river water d-excess values in and near the Adirondacks were close to the expected value for meteoric water (10) and also had lc-excess* values close to zero. This suggests that, consistent with a synoptic survey of rivers throughout the conterminous USA (Kendall and Coplen, 2001), rivers in the

Adirondacks integrate the δ18O and δD values of meteoric waters and are a reliable surface water proxy for regional meteoric water.

3.2.2. Plant water isotopes

Plant xylem water δD values (δDxw) were variable among individual plants, differing by 10 to 28‰ at each site (Fig. 5). Among sites, δDxw was indistinguishable (one-way ANOVA, p = 0.25). Across sites, the mean δDxw value (-65 ± 6‰, n = 57) was close to the inferred δDMAP based on the intersection of

111 Adirondack surface waters with local meteoric water lines (LMWL; -78‰) from precipitation collected at the Ottawa GNIP station (~100 km northwest of the Adirondack Park; IAEA/WMO, 2018) and in

Tompkins County, NY (~160 km southwest of the Adirondack Park; Coplen and Huang, 2000). Mean

δDxw may be higher than δDMAP due to soil water evaporation, or they may reflect incorporation of rainfall at the time of sampling (May/June 2017 Ottawa δDp = -54‰). Xylem water d-excess ranged from -0.4 to

29.7‰, with a mean of 13.6‰ (Fig. 6), indicating that some plant waters were evaporatively-enriched (d- excess < 10) while others were not, however there were no trends based on species or phylogeny. These plant waters were collected in May/June, when precipitation at Ottawa from 2007-2017 has d-excess values of 9.7‰. The lc-excess* values for xylem water ranged from -5.8 to 10.5‰ with a mean of 1.5 ±

3.1‰. These values are variable, but are relatively close to zero; an lc-excess* value of zero indicates no offset from the isotopic signature of precipitation. Globally, xylem water in temperate forests tends to have lc-excess* values ranging from -15 to +5 (Evaristo et al., 2015). Together, this is evidence that plant source waters across basins that span a range of elevations and vegetation assemblages mainly reflect annually integrated rainwater, with variable extents of evaporative enrichment.

Gymnosperm waters (-62 ± 7‰, n = 25) had modest but significant D-enrichment compared to angiosperms (-67 ± 6‰, n = 28; t-test, p = 0.009), perhaps due to contrasting hydraulic anatomy and freeze-thaw cavitation resistance between angiosperm xylem vessels and gymnosperm tracheids

(Brodribb et al., 2012). Additionally, groundwater contributions to gymnosperms and angiosperms may differ, although which group has greater groundwater contributions may depend on local conditions

(Evaristo et al., 2017; Evaristo and McDonnell, 2017).

The weighted mean annual isotope value for monthly precipitation collected during 2016 at the GNIP station in Ottawa was -76‰ for δD and -11.0‰ for δ18O (IAEA/WMO, 2018). The 2016 values are representative of the long term means measured at the site over the last 40 years (-75 ± 7‰ and -10.9 ±

112 1‰). The long-term LMWL at Ottawa (δD = 7.57 * δ18O + 6.72) is comparable to the LMWL from 1994 precipitation events sampled in Tompkins County, NY (δD = 7.68 * δ18O + 9.09; Coplen and Huang,

2000). The intersection of waters we sampled in the Adirondacks with the LMWLs from Ottawa and

Tompkins County, NY are comparable (within 5‰ of each other on average), and the mean of these two values is within 1‰ of the intersection of the same waters with the modeled LMWL for the Adirondacks from the Online Isotopes in Precipitation Calculator (OIPC; Bowen and Revenaugh, 2003; IAEA/WMO,

2015; Bowen, 2017; Bowen et al., 2005). Therefore, we use the intersection of Adirondack waters with the OIPC LMWL to estimate the precipitation source for summer/autumn lake water (-78‰), river water

(-79‰) and xylem water (-61‰) collected in the Adirondacks (Fig. 7). The intersection points for the lake and river waters we collected are comparable to that for lakes sampled by Brooks et al. (2014; -76‰) and nearby rivers sampled by Kendall and Coplen (2001; -73‰).

3.3. Sedimentary plant lipid abundance and distribution

Molecular distributions for lake sediment n-alkanes and n- alkanoic acids are reported in Table 4 and Fig.

8 and 9. All lake sediments had detectable amounts of n-C16 to n-C30 alkanoic acids with a strong even- over-odd preference and n-C17 to n-C35 alkanes with an odd-over-even chain length preference (CPI ranged from 1.9 to 5.8), which is indicative of terrestrial plant sources (Freeman and Pancost, 2014;

Diefendorf and Freimuth, 2017).

Across all lake sites, total n-alkanoic acid concentrations (∑C16-C30, µ = 3070 µg/g TOC) were between 2 and 16 times higher than total n-alkane concentrations (∑C17-C35, µ = 575 µg/g TOC), owing to especially high concentrations of n-C16, n-C20, n-C22, and n-C26 acid. The longer chain-lengths targeted for δD analysis (n-C28-30 acid and n-C27-31 alkane) had more comparable abundances to one another (Fig. 9); for these compounds, the compound in highest abundance (Cmax) for n-alkanoic acids and n-alkanes was generally n-C28 and n-C29, respectively in all lakes. The mean ACL27-35 across all sites (29.5 ± 0.2) also

113 reflects the dominance of n-C29 alkane. The mean ACL26-30 (26.5 ± 0.07) was much less variable and reflects the dominance of n-C26 alkanoic acid. Most lakes had n-alkane concentrations that were either higher than, or comparable to, long-chain than n-alkanoic acid concentrations, with the exception of four

(Chazy, East Pine, Little Green, and Hope) in which the n-C28 acid was considerably higher than n- alkanes (Fig. 9). The ratio of n-C28 alkanoic acid to n-C29 alkane ranged from 0.2 to 6 across all sites, and was strongly correlated with the total n-alkanoic acid to n-alkane ratio (R2 = 0.88; p < 0.0001, n = 36), which ranged from 1.3 to 19.

3.4. Sedimentary plant lipid δD and εapp composition

Sedimentary δDwax data are reported in Table 5 and Fig. 10. The mean n-C28 and n-C30 alkanoic acid δD values in the upper 9 cm of sediment across all sites were -164 ± 9‰ and -159 ± 9‰, respectively. In comparison, n-alkanes were significantly D-depleted across all long-chain homologues (mean n-C27, n-

C29 and n-C31 alkane δD: -191 ± 7‰, -188 ± 6‰ and -176 ± 6‰, respectively, t-test, p < 0.0001 for all n- alkane and n-alkanoic acid pairs). The overall δDwax variability was significantly larger for n-C28 alkanoic acid (1s = 9‰, n = 34) than for n-C29 alkane (1s = 6‰; Levene test, p = 0.004; Fig. 10). There were significant differences in δDwax values among lakes, for both n-C28 alkanoic acid and n-C29 alkane

(Kruskal-Wallis test, H = 30, 11 d.f., p = 0.0014; H = 31, 11 d.f., p = 0.0012, respectively).

Within the n-alkane class, the δD of long-chain homologues were positively correlated with one another

2 and had significant shared variance (n-C29 with n-C27 and n-C31, R = 0.63 and 0.38, respectively, p <

0.0001 for both). Stronger shared variance for n-C29 and n-C27 alkane suggests that these two compounds likely have a similar water source, while n-C31 is either influenced by different source water or is derived from different plants. A similar phenomenon was observed by Garcin et al. (2012) in the West African tropics. The δD of long-chain n-alkanoic acid homologues were positively correlated with shared variance

114 2 indicating common biological or environmental influences (n-C28 with n-C26 and n-C30, R = 0.53 and

0.67, respectively, p < 0.0001 for both).

Comparing between compound classes, n-alkanoic acid and n-alkane δD values were not significantly correlated with one another for all long-chain homologues except n-C27 alkane and n-C30 acid (p = 0.02).

This indicates that, while controls on δDwax are largely shared within compound classes, there are distinct controls on n-alkanes and n-alkanoic acids. The weighted mean δD values for n-alkanes (n-C27 to n-C31) and n-alkanoic acids (n-C28 to n-C30) are within 1-2‰ of the values for n-C29 alkane and n-C28 acid due to the dominance of these homologues (Fig. 9). Therefore, we focus the following discussion on these individual compounds.

Within individual lakes, n-alkane δD was always more negative than n-alkanoic acid δD, but the magnitude of the δD offset ranged among sites (Fig. 10). Offsets between n-C29 alkane and n-C30 acid

(εC30-C29) ranged from -8‰ (Raquette) to -58‰ (Little Green), and offsets between n-C27 alkane and n-C28 acid (εC28-C27) were comparable. Across all sites, mean εC30-C29 was -35‰ and εC28-C27 was -33‰. The sign and magnitude of δD offsets between compound classes in Adirondack lakes is similar to (but more variable than) that observed previously in temperate North American plants or Arctic lake sediments (-20 to -30‰; Chikaraishi and Naraoka, 2007; Hou et al., 2007; Daniels et al., 2017; Freimuth et al., 2017).

Lastly, we determined εapp values based on various source water estimates (Table 5). First, based on mean plant stem water (δDxw = -65‰), we determined εapp for n-C29 alkane (εC29-xw; -132 ± 6‰, n = 36) and n-

C28 alkanoic acid (εC28-xw; -106 ± 10‰, n = 34). We also compared this to εapp estimates based on OIPC- modeled mean annual precipitation δD at each site (ranged from -76 to -72‰ across all 12 sites) for n-C29 alkane (εC29-MAP; -123 ± 6‰, n = 36) and n-C28 alkanoic acid (εC28-MAP; -97 ± 9‰, n = 34).

115

4. Discussion

4.1. Drivers of plant lipid distributions and δD among lakes

The overall range of n-C29 alkane and n-C28 alkanoic acid δD in the 12 Adirondack lakes included in this study (22‰ and 35‰, respectively) was relatively low compared to prior surveys of multiple lakes in a single region with minimal δDp variation (~44-60‰; Douglas et al., 2012; Daniels et al., 2017). We investigated the influence of local depositional setting (including fluvial complexity and catchment vegetation) on sedimentary δDwax.

Basin fluvial complexity (here summarized by the ratio of fluvial inlets to outlets) represents how hydrologically open a lake basin is and therefore how vulnerable to evaporative lake water D-enrichment.

Indeed, mean lake water δD an δ18O values had significant yet relatively weak negative relationships with fluvial complexity (R2 = 0.23, p = 0.005 and R2 = 0.36, p = 0.0002, respectively), indicating that lakes that are more hydrologically open have more D-depleted water values. Fluvial complexity had no relationship to n-C29 alkane δD (p = 0.6, n = 36) but was negatively correlated with n-C28 alkanoic acid

δD (n = 34) and can explain up to approximately 27% of its variation (R2 = 0.27, p = 0.0017, n = 34). This suggests that sedimentary n-C29 alkane δD values are insensitive to lake water, while n-C28 alkanoic acid

δD may derive in part from plant sources that are affected by lake water evaporation and fluvial complexity.

While long-chain n-alkanes and n-alkanoic acids are likely derived mainly from terrestrial higher plants

(Ficken et al., 2000; Bush and McInerney, 2013), it may be the case that they are produced in different abundances in different plant species or growth forms, and that therefore the two compounds could have distinct sources (Polissar et al., 2014). While all of the sampled lake basins are forested, the proportion of

116 evergreen to deciduous vegetation varies considerably (Fig. 2), as does the extent of woody and emergent wetland vegetation in the shoreline. This may lead to a high degree of variability in lipid abundance and molecular composition, biosynthetic δD fractionation, and timing of leaf (and wax) synthesis in the source vegetation contributing to the sediments in each basin.

The NLCD data for a 30 m shoreline buffer around each lake was used to assess the influence of intra- catchment vegetation on sedimentary plant wax composition. Chazy and Raquette are excluded from this because they are dammed, which may lead to organic matter sourcing and flooded shoreline vegetation that diverge considerably from other sites (Section 2.1). The fraction of the shoreline that is comprised of evergreen vegetation is strongly correlated with the abundance ratio of long-chain (n-C27 to n-C35) alkanes

2 to long-chain (n-C28 and n-C30) acids in sediments (R = 0.67, n= 36, p < 0.0001; Fig. 11). This suggests evergreen trees are a dominant source of long-chain fatty acids to sediments. Based on the relationship with shoreline vegetation, the most likely mechanism of delivery would be direct leaf fall to the sediments; the correlation is slightly weaker when considering the entire lake drainage area (R2 = 0.52, n

= 36, p < 0.0001). By contrast, no relationship exists between the same long-chain n-alkane/n-alkanoic acid abundance ratio and the shoreline % deciduous forest (R2 = 0.02; p = 0.5). This suggests that while evergreen shoreline vegetation is a major source of n-alkanoic acids, the n-alkanes are more generally derived, both from deciduous and evergreen vegetation. It has been previously established that the relative abundance of n-alkanoic acids and n-alkanes in evergreen conifers varies among families and at the family level n-alkanoic acids are often rarer than n-alkanes (Diefendorf et al., 2015). However, these broad trends can vary considerably at the species- and family-level. For instance, the two compounds are comparably abundant or n-alkanoic acids are more abundant in the Pinaceae, and this family is well represented in the evergreen conifer species common in the Adirondacks (genera include Tsuga, Pinus,

Picea and Abies). While this study did not include a full vegetation survey, Thuja occidentalis (in a subclade of Cupressaceae, which tends to have higher alkane/acid ratios) and Tsuga canadensis on the shoreline of Moose Pond both had higher or comparable n-alkanoic acid to n-alkane composition (Fig. 9).

117

4.2. Do n-alkanes and n-alkanoic acids record the same water signal?

While the shoreline evergreen abundance is closely related to the relative abundance of n-alkanoic acids and n-alkanes in sediments, it has no significant relationship to sediment δDwax. This is consistent with prior studies that did not find any systematic differences in εapp between gymnosperms and angiosperms growing in the same environment (Chikaraishi and Naraoka, 2003; Bi et al., 2005; Liu et al., 2006; Gao et al., 2014). One contrasting study found that evergreen conifers had distinctly larger εapp than coexisting deciduous angiosperms (Pedentchouk et al., 2008). However, the latter was observed in a semi-desert setting and may therefore not be applicable to the Adirondacks. Therefore, we would not necessarily expect the abundance of evergreen vegetation to impact sediment δDwax.

We did find a relationship between wetland cover and sedimentary δDwax. The relative abundance of wetlands (including woody wetlands and emergent aquatic plants) was correlated with both n-C28 and n-

2 C30 acid δD (R = 0.39 and 0.44, respectively, p < 0.0001). The range of fractional wetland cover is small

(0.004 to 0.07), but nevertheless the basins with more wetland coverage had significantly lower n- alkanoic acid δD values. The negative correlation may reflect more D-depleted source water accessed by wetland vegetation than trees growing in well-drained soils. The same relationship did not exist for n- alkanes, which had no relationship with the abundance of wetland cover. This may further indicate that the two compounds have distinct plant sources and possibly access different water pools.

In addition, n-C29 alkane and n-C28 alkanoic acid δD were negatively correlated with latitude (p ≤ 0.005) and positively correlated with elevation (p ≤ 0.0018). There is a decreasing gradient in precipitation amount from the southwest to northeast across the Adirondacks (Ito et al., 2002) as well as an increasing gradient with elevation. This is manifested in a latitudinal gradient in surface waters with greater lc- excess* values in lower latitudes. While δDwax from both compounds was approximately equally affected

118 by regional-scale factors such as latitude and elevation, the basin-specific vegetation differences seemed only to affect n-alkanoic acids. In addition, the abundance and δD of n-alkanoic acids varied more widely than n-alkanes (Fig. 9). Together, this suggests greater sensitivity of n-alkanoic acid δD to localized factors at the catchment level, such as vegetation assemblage. If this is the case, then n-alkanes may more faithfully record regional signals such as δDp. If n-alkanoic acids are in fact more sensitive to intra- catchment factors, then including δDwax from both compounds may have potential to yield additional environmental information. The δD offset between n-alkanes and n-alkanoic acids (εacid-alkane) varied widely (from 8‰ to 55‰) among lakes and affords an opportunity to investigate the underlying drivers of this feature.

We found a strong correlation between sediment C/N and the δD offset between n-C30 alkanoic acid and

2 n-C29 alkane (εC30-C29; R = 0.7, p < 0.0001), wherein the magnitude of εC30-C29 was largest in lakes with low C/N and thus a greater aquatic organic matter sourcing (Fig. 12). The same was not true for the

2 comparable homologues n-C28 alkanoic acid and n-C27 alkane (R = 0.17, p = 0.02) indicating that the relationship is specific to the larger chain-length homologues. Emergent and submerged aquatic plants tend to produce higher concentrations of total n-alkanoic acids (n-C14-34) than total n-alkanes (n-C19-35), and while submerged aquatic plants generally have n-alkanoic acid Cmax between C20-24, emergent aquatics tend to have higher and more variable Cmax, ranging from C22-32 (Ficken et al., 2000). Therefore, it is possible that the εC30-C29 relationship with C/N is linked to emergent aquatic plants and their use of lake water. However, further investigation of the potential aquatic sources of n-C30 alkanoic acid is needed.

4.3. εapp in the Adirondacks compared with other biomes

εapp estimates for the Adirondacks based on xylem water (-132 ± 6‰, εC29-xw and -106 ± 10‰, εC28-xw) are comparable to εapp estimates from lakes in other temperate forested sites in Europe, North America and

119 Japan (-137‰ to -131‰; Sachse et al., 2004; Polissar and Freeman, 2010; Seki et al., 2010). This suggests an emerging consistency (within ~10‰) in εapp at the sediment-level for the temperate forested biome. In order to compare our εapp results more broadly with other biomes, we use an approach from

Polissar and Freeman (2010). In that study, εapp in the uppermost sediments of multiple lakes within several different biomes was evaluated relative to the 18O-enrichment of lake water relative to

18 precipitation δ O (εlakewater-ppt), which is an indicator of local aridity. This allows for comparison of εapp values across biomes while controlling for the inclusion of more evaporative/arid sites in the dataset from each biome. They found that the slopes in this comparison for forested catchments were near zero (Fig.

13), indicating limited influence of local aridity on soil water or plant water evaporative enrichment. By contrast, larger positive slopes for shrub- and grass-dominated regions suggest greater evaporative effects on εapp at various sites within these ecosystems/biomes (Fig. 13).

Sites in the Adirondack dataset had εlakewater-ppt ranging from 0.6 to 5.2, and these values were negatively correlated with the ratio of fluvial inlets to outlets (R2 = 0.4, p = 0.0001) indicating that more hydrologically open basins tend to have smaller evaporative 18O-enrichemnt relative to precipitation.

Evaluating the Adirondack data using the approach from Polissar and Freeman (2010), the intercept of the

Adirondacks dataset (-132‰) is nearly identical to the intercept from an oak forest nearly 400 km to the southwest (Fig. 13; Polissar and Freeman, 2010). This is also close (within 1s) to the intercept for sediments from a tropical angiosperm forest biome in South America (-125‰; Polissar and Freeman,

2010), demonstrating common εapp values for forested biomes in both temperate and tropical climates, regardless of differing species composition and climates. We note that the slope of the Adirondack data is near zero, which is further indication that forested catchments are relatively insensitive to local aridity (as indicated by εlakewater-ppt values). This increases our confidence in applying εapp values derived by this approach for downcore interpretation when we can confirm the presence of forests through other proxy data.

120

5. Conclusions

This study measured the δDwax signal in lake surface sediments throughout the Adirondack Mountains to determine the degree to which catchment-specific factors such as vegetation cover and fluvial complexity affect the fidelity of plant waxes as tracers of water isotopes. Based on plant water and surface water isotope measurements in the Adirondacks, plant source waters among basins are biased toward summer precipitation values but do not vary systematically among sites. Provided sedimentary plant wax primarily reflects plant source water, then we would also expect δDwax to be similar among the 12 lakes sampled. However, n-alkane and n-alkanoic acid δD values were significantly different among sites, with higher variability in n-alkanoic acid δD than n-alkane δD. We also found that across basins with diverse vegetation assemblages, the abundance of long-chain n-alkanes was remarkably consistent while long- chain n-alkanoic acid abundance varied widely. In addition, the extent of evergreen vegetation in lake shorelines was correlated with the abundance of long-chain n-alkanes relative to n-alkanoic acids, which may indicate evergreen conifer sources for n-alkanoic acids. Together, these data suggest that sedimentary n-alkanoic acids are more responsive to site-specific conditions than n-alkanes.

The δD offset between compound classes was highly variable among sites and was strongly correlated with sediment C/N ratios, which may indicate an aquatic source for long-chain n-alkanoic acids. Overall, we found larger εapp for n-alkanes (-132 ± 6‰) than for n-alkanoic acids (-106 ± 10‰). Comparison with similar studies across a range of biomes suggests that εapp from n-alkanes in forested catchments is stable regardless of catchment-scale aridity. This increases confidence in applying εapp estimates based on surface sediments to interpret older sediments. Overall, this study provides evidence for a significant vegetation effect on sedimentary δDwax (especially n-alkanoic acids) in the absence of a gradient in δDp over the sampled region. This suggests that sedimentary plant waxes are sourced from within a lake catchment rather than mixed at a regional scale, and that sedimentary wax composition may even be

121 sensitive to the vegetation composition in the immediate shoreline area. Site-specific factors had a stronger apparent influence on n-alkanoic acid molecular and δD variability among sites. By contrast, n- alkanes were relatively insensitive to site-specific factors, which may make them more appropriate candidates for δDp reconstruction.

Acknowledgements:

We thank the New York State Department of Environmental Conservation, especially to Jonathan

DeSantis and Barbara Lucas-Wilson, for sampling permits (TRP #2349/10310). Daphne Taylor and

Natasha Karniski at the SUNY-ESF Newcomb campus provided access to Wolf Pond. Thanks go to

Helen Eifert and Elliot Boyd for field sampling assistance, and to Andrew Diefendorf for designing and building refrigerated core storage and an essential core splitter. Paul Mckenzie provided temporary core storage and Kelly Grogan assisted with sample preparation. This research was supported by the US

National Science Foundation (EAR-1229114 to AFD and EAR-1636740 to AFD and TVL) and by a

University of Cincinnati Research Council award to EJF.

122 REFERENCES

Bi, X., Sheng, G., Liu, X., Li, C. and Fu, J. (2005) Molecular and carbon and hydrogen isotopic composition of n-alkanes in plant leaf waxes. Org. Geochem. 36, 1405-1417. Binford, M.W. (1990) Calculation and uncertainty analysis of 210 Pb dates for PIRLA project lake sediment cores. J. Paleolimn. 3, 253-267. Binford, M.W., Kahl, J.S. and Norton, S.A. (1993) Interpretation of 210 Pb profiles and verification of the CRS dating model in PIRLA project lake sediment cores. J. Paleolimn. 9, 275-296. Bird, B.W., Polisar, P.J., Lei, Y., Thompson, L.G., Yao, T., Finney, B.P., Bain, D.J., Pompeani, D.P. and Steinman, B.A. (2014) A Tibetan lake sediment record of Holocene Indian summer monsoon variability. Earth Planet. Sci. Lett. 399, 92-102. Bowen, G.J. (2018) The Online Isotopes in Precipitation Calculator, version 3.1. http://www.waterisotopes.org. Bowen, G.J. and Revenaugh, J. (2003) Interpolating the isotopic composition of modern meteoric precipitation. Water Resources Research 39. Bowen, G.J., Wassenaar, L.I. and Hobson, K.A. (2005) Global application of stable hydrogen and oxygen isotopes to wildlife forensics. Oecologia 143, 337-348. Brodribb, T.J., Pittermann, J. and Coomes, D.A. (2012) Elegance versus Speed: Examining the Competition between Conifer and Angiosperm Trees. International Journal of Plant Sciences 173, 673-694. Brooks, J.R., Gibson, J.J., Birks, S.J., Weber, M.H., Rodecap, K.D. and Stoddard, J.L. (2014) Stable isotope estimates of evaporation: inflow and water residence time for lakes across the United States as a tool for national lake water quality assessments. Limnol. Oceanogr. 59, 2150-2165. Bush, R.T. and McInerney, F.A. (2013) Leaf wax n-alkane distributions in and across modern plants: Implications for paleoecology and chemotaxonomy. Geochim. Cosmochim. Acta 117, 161-179. Castañeda, I.S. and Schouten, S. (2011) A review of molecular organic proxies for examining modern and ancient lacustrine environments. Quat. Sci. Rev. 30, 2851-2891. Charles, D.F., Binford, M.W., Furlong, E.T., Hites, R.A., Mitchell, M.J., Norton, S.A., Oldfield, F., Paterson, M.J., Smol, J.P. and Uutala, A.J. (1990) Paleoecological investigation of recent lake acidification in the Adirondack Mountains, NY. J. Paleolimn. 3, 195-241. Chikaraishi, Y. and Naraoka, H. (2003) Compound-specific δD-δ13C analyses of n-alkanes extracted from terrestrial and aquatic plants. Phytochemistry 63, 361-371. Chikaraishi, Y. and Naraoka, H. (2007) δ13C and δD relationships among three n-alkyl compound classes (n-alkanoic acid, n-alkane and n-alkanol) of terrestrial higher plants. Org. Geochem. 38, 198-215. Coplen, T.B. (1988) Normalization of oxygen and hydrogen isotope data. Chemical Geology: Isotope Geoscience section 72, 293-297. Coplen TB, Huang R. 2000. Stable hydrogen and oxygen isotope ratios for selected sites of the National Oceanic and Atmospheric Administration’s Atmospheric Intergrated Research Monitoring Network (AIRMoN). US Geological Survey Open-File Report 00-279: Reston; 54 pp. Coplen, T.B., Brand, W.A., Gehre, M., Gröning, M., Meijer, H.A.J., Toman, B. and Verkouteren, R.M. (2006) New guidelines for δ13C measurements. Anal. Chem. 78, 2439-2441. Daniels, W.C., Russell, J.M., Giblin, A.E., Welker, J.M., Klein, E.S. and Huang, Y. (2017) Hydrogen isotope fractionation in leaf waxes in the Alaskan Arctic tundra. Geochim. Cosmochim. Acta 213, 216-236. Dansgaard, W. (1964) Stable isotopes in precipitation. Tellus 16, 436-468. Diefendorf, A.F., Patterson, W.P., Holmden, C. and Mullins, H.T. (2008) Carbon isotopes of marl and lake sediment organic matter reflect terrestrial landscape change during the late Glacial and early Holocene (16,800 to 5,540 cal yr BP): a multiproxy study of lacustrine sediments at Lough Inchiquin, western Ireland. J. Paleolimn. 39, 101-115. Diefendorf, A.F., Leslie, A.B. and Wing, S.L. (2015) Leaf wax composition and carbon isotopes vary among major conifer groups. Geochim. Cosmochim. Acta 170, 145-156.

123 Diefendorf, A.F. and Freimuth, E.J. (2017) Extracting the most from terrestrial plant-derived n-alkyl lipids and their carbon isotopes from the sedimentary record: A review. Org. Geochem. 103, 1- 21. Douglas, P.M.J., Pagani, M., Brenner, M., Hodell, D.A. and Curtis, J.H. (2012) Aridity and vegetation composition are important determinants of leaf-wax δD values in southeastern Mexico and Central America. Geochim. Cosmochim. Acta 97, 24-45. Driscoll, C., Poster, K., Kretser, W. and Raynal, D. (1995) Long-term trends in the chemistry of precipitation and lake water in the Adirondack region of New York, USA. Water, Air, and Soil Pollution 85, 583-588. Driscoll, C.T. and Newton, R.M. (1985) Chemical characteristics of Adirondack lakes. Environmental science & technology 19, 1018-1024. Driscoll, C.T., Newton, R.M., Gubala, C.P., Baker, J.P. and Christensen, S.W. (1991) Adirondack mountains, Acidic Deposition and Aquatic Ecosystems. Springer, pp. 133-202. Eley, Y., Dawson, L., Black, S., Andrews, J. and Pedentchouk, N. (2014) Understanding 2H/1H systematics of leaf wax n-alkanes in coastal plants at Stiffkey saltmarsh, Norfolk, UK. Geochim. Cosmochim. Acta 128, 13-28. Evaristo, J., Jasechko, S. and McDonnell, J.J. (2015) Global separation of plant transpiration from groundwater and streamflow. Nature 525, 91. Evaristo, J. and McDonnell, J.J. (2017) Prevalence and magnitude of groundwater use by vegetation: a global stable isotope meta-analysis. Scientific reports 7, 44110. Evaristo, J., McDonnell, J.J. and Clemens, J. (2017) Plant source water apportionment using stable isotopes: A comparison of simple linear, two-compartment mixing model approaches (revision 3). Hydrological Processes. Feakins, S.J. and Sessions, A.L. (2010) Controls on the D/H ratios of plant leaf waxes from an arid ecosystem. Geochim. Cosmochim. Acta 74, 2128-2141. Feakins, S.J., Kirby, M.E., Cheetham, M.I., Ibarra, Y. and Zimmerman, S.R.H. (2014) Fluctuation in leaf wax D/H ratio from a southern California lake records significant variability in isotopes in precipitation during the late Holocene. Org. Geochem. 66, 48-59. Ficken, K.J., Li, B., Swain, D.L. and Eglinton, G. (2000) An n-alkane proxy for the sedimentary input of submerged/floating freshwater aquatic macrophytes. Org. Geochem. 31, 745-749. Freeman, K.H. and Pancost, R.D. (2014) Biomarkers for Terrestrial Plants and Climate, in: Turekian, H.D.H.K. (Ed.), Treatise on Geochemistry (Second Edition). Elsevier, Oxford, pp. 395-416. Freimuth, E.J., Diefendorf, A.F. and Lowell, T.V. (2017) Hydrogen isotopes of n-alkanes and n-alkanoic acids as tracers of precipitation in a temperate forest and implications for paleorecords. Geochim. Cosmochim. Acta 206, 166-183. Fritz, P. and Clark, I. (1997) Environmental isotopes in hydrogeology. CRC press. Gao, L., Edwards, E.J., Zeng, Y. and Huang, Y. (2014) Major Evolutionary Trends in Hydrogen Isotope Fractionation of Vascular Plant Leaf Waxes. PLoS ONE 9, e112610. Garcin, Y., Schwab, V.F., Gleixner, G., Kahmen, A., Todou, G., Séné, O., Onana, J.-M., Achoundong, G. and Sachse, D. (2012) Hydrogen isotope ratios of lacustrine sedimentary n-alkanes as proxies of tropical African hydrology: insights from a calibration transect across Cameroon. Geochim. Cosmochim. Acta 79, 106-126. Gray, A.B., Pasternack, G.B. and Watson, E.B. (2010) Hydrogen peroxide treatment effects on the particle size distribution of alluvial and marsh sediments. The Holocene 20, 293-301. Heiri, O., Lotter, A.F. and Lemcke, G. (2001) Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results. J. Paleolimn. 25, 101-110. Homer, C., Dewitz, J., Yang, L., Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N., Wickham, J. and Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States–representing a decade of land cover change information. Photogrammetric Engineering & Remote Sensing 81, 345-354.

124 Hou, J.Z., D'Andrea, W.J., MacDonald, D. and Huang, Y.S. (2007) Hydrogen isotopic variability in leaf waxes among terrestrial and aquatic plants around Blood Pond, Massachusetts (USA). Org. Geochem. 38, 977-984. IAEA/WMO (2018). Global Network of Isotopes in Precipitation. The GNIP Database. Accessible at: https://nucleus.iaea.org/wiser. Kahmen, A., Hoffmann, B., Schefuß, E., Arndt, S.K., Cernusak, L.A., West, J.B. and Sachse, D. (2013a) Leaf water deuterium enrichment shapes leaf wax n-alkane δD values of angiosperm plants II: Observational evidence and global implications. Geochim. Cosmochim. Acta 111, 50-63. Kahmen, A., Schefuß, E. and Sachse, D. (2013b) Leaf water deuterium enrichment shapes leaf wax n- alkane δD values of angiosperm plants I: Experimental evidence and mechanistic insights. Geochim. Cosmochim. Acta 111, 39-49. Kendall, C. and Coplen, T.B. (2001) Distribution of oxygen-18 and deuterium in river waters across the United States. Hydrological Processes 15, 1363-1393. Landwehr J, Coplen T. 2006. in Isotopes in environmental studies 132– 135 (IAEA-CN-118/56, International Atomic Energy Agency) Liu, W., Yang, H. and Li, L. (2006) Hydrogen isotopic compositions of n-alkanes from terrestrial plants correlate with their ecological life forms. Oecologia 150, 330-338. Marzi, R., Torkelson, B.E. and Olson, R.K. (1993) A revised carbon preference index. Org. Geochem. 20, 1303-1306. McInerney, F.A., Helliker, B.R. and Freeman, K.H. (2011) Hydrogen isotope ratios of leaf wax n-alkanes in grasses are insensitive to transpiration. Geochim. Cosmochim. Acta 75, 541-554. Meyers, P.A. (2003) Applications of organic geochemistry to paleolimnological reconstructions: a summary of examples from the Laurentian Great Lakes. Org. Geochem. 34, 261-289. Niedermeyer, E.M., Sessions, A.L., Feakins, S.J. and Mohtadi, M. (2014) Hydroclimate of the western Indo-Pacific Warm Pool during the past 24,000 years. Proc. Natl. Acad. Sci. U.S.A. 111, 9402- 9406. Pagani, M., Pedentchouk, N., Huber, M., Sluijs, A., Schouten, S., Brinkhuis, H., Sinninghe Damste, J.S., Dickens, G.R. and Scienctists, t.e. (2006) Arctic hydrology during global warming at the Palaeocene/Eocene thermal maximum. Nature 442, 671-675. Pancost, R.D. and Boot, C.S. (2004) The palaeoclimatic utility of terrestrial biomarkers in marine sediments. Marine Chemistry 92, 239-261. Paul, D. and Skrzypek, G. (2006) Flushing time and storage effects on the accuracy and precision of carbon and oxygen isotope ratios of sample using the Gasbench II technique. Rapid Communications in Mass Spectrometry 20, 2033-2040. Pedentchouk, N., Sumner, W., Tipple, B. and Pagani, M. (2008) δ13C and δD compositions of n-alkanes from modern angiosperms and conifers: An experimental set up in central Washington State, USA. Org. Geochem. 39, 1066-1071. Polissar, P., Preefer, M. and Liu, C. (2014) Fidelity of leaf-wax n-alkane and n-alkanoic acid D/H ratios in space and time, American Geophysical Union, pp. PP43E-07. Polissar, P.J. and Freeman, K.H. (2010) Effects of aridity and vegetation on plant-wax δD in modern lake sediments. Geochim. Cosmochim. Acta 74, 5785-5797. Rach, O., Brauer, A., Wilkes, H. and Sachse, D. (2014) Delayed hydrological response to Greenland cooling at the onset of the Younger Dryas in western Europe. Nature Geosci 7, 109-112. Sachse, D., Radke, J. and Gleixner, G. (2004) Hydrogen isotope ratios of recent lacustrine sedimentary n- alkanes record modern climate variability. Geochim. Cosmochim. Acta 68, 4877-4889. Sachse, D., Radke, J. and Gleixner, G. (2006) δD values of individual n-alkanes from terrestrial plants along a climatic gradient – Implications for the sedimentary biomarker record. Org. Geochem. 37, 469-483. Sachse, D., Gleixner, G., Wilkes, H. and Kahmen, A. (2010) Leaf wax n-alkane δD values of field-grown barley reflect leaf water δD values at the time of leaf formation. Geochim. Cosmochim. Acta 74, 6741-6750.

125 Sachse, D., Billault, I., Bowen, G.J., Chikaraishi, Y., Dawson, T.E., Feakins, S.J., Freeman, K.H., Magill, C.R., McInerney, F.A., van der Meer, M.T.J., Polissar, P., Robins, R.J., Sachs, J.P., Schmidt, H.- L., Sessions, A.L., White, J.W.C., West, J.B. and Kahmen, A. (2012) Molecular paleohydrology: Interpreting the hydrogen-isotopic composition of lipid biomarkers from photosynthesizing organisms. Annu. Rev. Earth Planet. Sci. 40, 221-249. Seki, O., Nakatsuka, T., Shibata, H. and Kawamura, K. (2010) A compound-specific n-alkane δ13C and δD approach for assessing source and delivery processes of terrestrial organic matter within a forested watershed in northern Japan. Geochim. Cosmochim. Acta 74, 599-613. Sessions, A.L., Burgoyne, T.W., Schimmelmann, A. and Hayes, J.M. (1999) Fractionation of hydrogen isotopes in lipid biosynthesis. Org. Geochem. 30, 1193-1200. Sessions, A.L. (2016) Factors controlling the deuterium contents of sedimentary hydrocarbons. Org. Geochem. 96, 43-64. Stager, J.C., Cumming, B.F., Laird, K.R., Garrigan-Piela, A., Pederson, N., Wiltse, B., Lane, C.S., Nester, J. and Ruzmaikin, A. (2016) A 1600-year diatom record of hydroclimate variability from Wolf Lake, New York. The Holocene 27, 246-257. Tierney, J.E., Russell, J.M., Huang, Y., Damsté, J.S.S., Hopmans, E.C. and Cohen, A.S. (2008) Northern hemisphere controls on tropical southeast African climate during the past 60,000 years. Science 322, 252-255. Tipple, B.J. and Pagani, M. (2013) Environmental control on eastern broadleaf forest species’ leaf wax distributions and D/H ratios. Geochim. Cosmochim. Acta 111, 64-77. Vogts, A., Badewien, T., Rullkötter, J. and Schefuß, E. (2016) Near-constant apparent hydrogen isotope fractionation between leaf wax n-alkanes and precipitation in tropical regions: evidence from a marine sediment transect off SW Africa. Org. Geochem. 96, 18-27. Welker, J. (2000) Isotopic (δ18O) characteristics of weekly precipitation collected across the USA: an initial analysis with application to water source studies. Hydrological Processes 14, 1449-1464. West, A.G., Patrickson, S.J. and Ehleringer, J.R. (2006) Water extraction times for plant and soil materials used in stable isotope analysis. Rapid Communications in Mass Spectrometry 20, 1317- 1321. Whitehead, D.R., Charles, D.F., Jackson, S.T., Smol, J.P. and Engstrom, D.R. (1990) The developmental history of Adirondack (NY) lakes, Paleolimnology and the Reconstruction of Ancient Environments. Springer, pp. 169-190.

126

Figure 1. Elevation map showing the Adirondack Park border (blue line) and locations of the 12 lake sampling sites (circles) as well as locations of precipitation isotope data used in this study (triangles), to the north in Ottawa (IAEA/WMO, 2018) and to the southwest in Tompkins County, NY (Coplen and

Huang, 2000).

127

Figure 2. Vegetation assemblages for each lake catchment (A) and for a 30 m lake shoreline buffer (B) based on National Land Cover Data (NLCD) from 2011.

128

Figure 3. Lake sediment grain-size distribution as percent clay, silt and sand.

129

13 Figure 4. Crossplot of the TOC and δ Corg of sediments sampled at 1, 5, and 9 cm depth in the sediment core from each lake for closed basins (filled symbols) and open basins (open symbols). The data are significantly correlated (R2 = 0.4, p < 0.0001, n = 36).

130

Figure 5. Adirondack δD and δ18O composition of surface waters, including lake water (black symbols, closed for closed basins) and river water (yellow symbol), snow (light blue symbols), and plant xylem waters from 7 of the 12 sites included in this study.

131

Figure 6. Deuterium-excess (d-excess) values for waters collected in this study (bold text) and others, including [1] precipitation from the Ottawa GNIP station (IAEA/WMO, 2018); [2] OIPC-modeled monthly precipitation for the 12 lakes in this study (Welker, 2000; Bowen and Revenaugh, 2003;

Bowen, 2017); [3] precipitation collected in Tompkins County, NY, USA (Coplen and Hunag, 2000); river waters collected ~60 km west of Adirondack Park (Kendall and Coplen, 2001); and [5] September lake water sampled in the Adirondacks (Brooks et al., 2014). The dashed line represents a d-excess value typical of unevaporated meteoric water.

132

Figure 7. The isotopic composition of lake water (summer and autumn), river water, xylem water and snow collected in the Adirondacks relative to the OPIC-based LMWL (black).

133

Figure 8. Molecular distributions for n-C17 to n-C35 alkanes from each lake. Data are in µg n-alkane/g

TOC. Dark bars represent odd carbon numbered compounds. Error bars represent the 1s standard deviation for three samples (at 1, 5, and 9 cm) from each sediment core.

134

Figure 9. Plant wax concentrations in lake sediments (left panel, in µg/g TOC) and in two evergreen conifers at Moose Lake (right panel, in µg/g dry leaf) for long-chain n-alkanoic acids (brown) and n- alkanes (blue).

135

Figure 10. The δD composition of n-C29 alkane (blue) and n-C28 alkanoic acid (brown) from

Adirondack lake surface sediments.

136

Figure 11. Crossplot showing the correlation between the percent of shoreline with evergreen land cover and the abundance ratio of long-chain n-C27 to n-C35 alkanes to long-chain n-C26 to n-C30 alknaoic acids in Adirondack lake sediments.

137

Figure 12. Relationship between the C/N ratio in sediments and the δD offset between n-C29 alkane and n-C30 acid (εC30-C29).

138

18 Figure 13. The relationship between catchment aridity (εlakewater-ppt δ O) and εapp for n-C29 alkane in different North and South American biomes. The dashed line indicates εlakewater-ppt values where lakewater 18O-enrichment relative to precipitation (i.e., local aridity) is at a minimum. Data are from

Douglas et al. (2012) [1], Polissar and Freeman (2010) [2], and this study (black points). Figure is modified from Polissar and Freeman (2010).

139

Table 1. Location and hydrologic characteristics of each of the 12 sampled lakes, as well as organic properties of the surface sediments.

Elevation Lake Area Catchment Lake/Catchment Water depth at εlakewater-ppt δ13Corg TOC Site Latitude Longitude (m) (km^2) Area (km^2) Area Ratio core site (m) Inlets/Outlets (δ18O, ‰) (‰) (Wt%) C/N Chazy 44.72583 -73.82077 466 7.5 58 0.13 3.1 4 0.6 -26.8 10.0 12.5 Debar 44.62189 -74.19359 483 0.4 6.4 0.06 9.2 3 1.5 -31.1 18.9 10.5 East Pine 44.33827 -74.41497 513 0.3 1.1 0.24 10.2 1 3 -31.3 29.7 10.6 Heart 44.18251 -73.96928 665 0.1 0.8 0.14 13.12 0 3.2 -28.5 18.9 12.2 Hope 44.51320 -74.12534 523 0.1 0.4 0.28 9.9 0 3.8 -29.6 29.3 13.2 Horseshoe 44.13558 -74.62190 524 1.6 11.5 0.14 4 1 4 -30.5 15.5 10.3 Little Green 44.35744 -74.29887 521 0.3 1.4 0.21 14.05 2 3.4 -29.3 23.6 12.2 Moose 44.37172 -74.06197 475 0.7 17.4 0.04 20 2 0.9 -28.6 17.7 12.1 Quiver 43.73786 -74.87087 539 0.1 0.6 0.13 4 2 3.5 -32.0 35.2 13.4 Raquette 44.22676 -74.46995 470 4.1 17.1 0.24 3.8 4 1.2 -28.7 15.4 14.1 Sucker 44.18081 -75.12080 426 0.4 2.6 0.14 1.8 2 5.2 -28.1 25.2 11.6 Wolf 44.01785 -74.22170 558 0.6 4.7 0.13 13 1 -28.8 17.7 12.2

140 Table 2. Summary δD, δ18O, d-excess, and lc-excess* values (mean, 1s, n; ‰ VSMOW) for surface waters, snow and xylem water collected in the Adirondacks.

Sample Type δD 1s n δ18O 1s n d-excess 1s n lc-excess* 1s n Lake -61.9 8.0 46 -8.0 1.2 46 1.8 6.8 46 -3.4 2.6 47 Lake - Winter 2017 -76.7 11.7 11 -9.0 1.4 11 -4.3 17.4 11 -6.0 6.8 11 River -67.3 9.4 9 -9.4 1.1 9 7.7 6.4 9 -1.2 2.5 9 Snow -105.6 17.7 3 -15.4 2.6 3 17.8 3.4 3 1.8 0.9 3 Xylem -73.5 6.8 58 -10.2 1.0 57 8.5 9.2 56 -1.5 5.1 56

141 Table 3. Xylem water δD and δ18O values (‰ VSMOW) for plants collected at seven of the 12 sites of lake sediment analysis.

Site Species Angio/Gymno δDxw 1s n δ18Oxw 1s n East Pine Abies balsamea Gymnosperm -70.6 0.5 2 -10.4 0.5 2 East Pine Acer rubrum Angiosperm -76.3 2.0 2 -10.1 0.6 2 East Pine Betula papyrifera Angiosperm -72.3 0.1 2 -9.9 0.3 2 East Pine Pinus strobus Gymnosperm -76.9 0.1 2 -10.7 0.1 2 Heart Acer rubrum Angiosperm -74.1 0.0 2 -10.3 1.0 2 Heart Betula papyrifera Angiosperm -79.7 1 -11.2 1 Heart Betula papyrifera Angiosperm -68.9 8.5 2 -10.2 1 Heart Pinus resinosa Gymnosperm -79.9 1.7 2 -12.5 1.0 2 Heart Shrub Angiosperm -65.7 11.2 2 -8.9 0.5 2 Heart Thuja occidentalis Gymnosperm -74.0 4.0 2 -9.9 0.9 2 Heart Tsuga canadensis Gymnosperm -74.9 2.2 2 -10.9 1.0 2 Hope Tsuga canadensis Gymnosperm -71.9 0.6 2 -9.6 0.1 2 Horseshoe Betula papyrifera Angiosperm -86.6 1 -11.5 1 Little Green Acer rubrum Angiosperm -72.6 1 -9.3 1 Little Green Betula papyrifera Angiosperm -69.1 1.2 2 -9.6 0.2 2 Little Green Fagus grandifolia Angiosperm -74.5 3.0 2 -9.9 0.2 2 Little Green Picea rubens Gymnosperm -73.4 6.3 2 -10.6 0.7 2 Little Green Pinus strobus Gymnosperm -57.7 14.6 2 -8.3 0.1 2 Little Green Tsuga canadensis Gymnosperm -67.1 5.5 3 -9.2 0.6 3 Moose Acer rubrum Angiosperm -75.0 3.2 3 Moose Betula papyrifera Angiosperm -86.3 3.5 3 -12.0 0.4 3 Moose Fagus grandifolia Angiosperm -72.9 10.7 3 -10.7 0.5 3 Moose Shrub Angiosperm -74.0 1.2 2 -10.0 0.1 2 Moose Thuja occidentalis Gymnosperm -72.6 2.0 2 -9.9 0.9 2 Moose Tsuga canadensis Gymnosperm -75.1 7.9 2 -11.2 1.2 2 Raquette Pinus sylvestris Gymnosperm -71.9 0.4 2 -10.0 0.7 2

142 Table 4. Summary molecular abundance and distribution data for long-chain n-alkanes and n-alkanoic acids in surface sediments at each site.

ACL (n-C27-35 alkane)/ n-alkane abundance (µg/g TOC) n-alkanoic acid abundance (µg/g TOC) Site n-C27-35 alkane n-C26-30 acid (n-C28-30 acid) n-C27 n-C29 n-C31 n-C33 n-C35 n-C26 n-C28 n-C30 Chazy 29.7 26.5 1.4 54.2 61.7 53.5 17.4 13.1 973.8 154.2 84.1 Debar 29.3 26.5 5.8 77.0 75.7 57.5 19.5 5.6 175.2 28.1 12.8 East Pine 29.4 26.4 0.5 27.1 30.6 27.3 9.1 1.1 1052.6 145.5 50.3 Heart 29.6 26.6 7.4 71.9 90.9 75.7 24.6 10.8 144.9 23.8 14.0 Hope 29.3 26.5 1.1 61.6 71.0 60.4 15.1 2.1 1009.0 141.7 79.5 Horseshoe 29.1 26.5 4.1 52.0 46.5 29.4 12.0 1.9 192.5 30.4 12.1 Little Green 29.6 26.4 0.8 42.9 45.6 48.5 14.7 2.3 1371.5 161.3 69.5 Moose 29.7 26.5 1.6 64.7 57.9 48.0 17.4 19.4 618.6 86.1 47.0 Quiver 29.7 26.5 4.1 33.0 48.9 52.7 13.1 2.3 187.1 27.4 14.5 Raquette 29.7 26.6 3.4 59.5 82.3 67.8 20.2 10.8 444.5 85.2 54.7 Sucker 29.4 26.6 4.5 49.5 63.1 42.8 16.7 3.5 177.4 31.0 13.5 Wolf 29.9 26.5 6.5 51.5 68.3 57.5 16.3 20.3 156.1 21.6 13.5

143 Table 5. Mean δDwax and εapp with 1s standard deviation (italicized) for lake surface sediments (n = 3 samples per site, at 1, 5, and 9 cm depth). εapp values are based on OIPC-based mean annual precipitation (εwax/MAP) or on mean xylem water δD values across all sites

(εwax/xyl).

n- alkane δD (‰) n- alkanoic acid δD (‰) εacid-alkane (‰) εwax/MAP (‰) εwax/xyl (‰) Site n-C27 n-C29 n-C31 n-C28 n-C30 εC30-C29 εC28-C27 n-C29 n-C28 n-C29 n-C28 Chazy -188.1 1.8 -187.9 1.4 -178.8 10.5 -174.1 3.2 -164.5 2.7 -28.0 -17.0 -121.1 1.5 -106.2 3.5 -131.5 1.5 -116.2 3.5 Debar -194.5 1.8 -195.7 1.9 -178.8 1.9 -167.1 1.4 -158.5 1.4 -44.1 -33.0 -129.5 2.1 -98.6 1.5 -139.8 2.0 -108.7 1.5 East Pine -198.6 4.2 -193.2 2.8 -180.3 7.6 -158.0 2.3 -155.1 1.1 -45.1 -48.3 -127.8 3.0 -89.7 2.6 -137.2 3.0 -98.9 2.5 Heart -187.4 1.3 -185.5 1.5 -174.5 1.3 -154.5 0.7 -153.6 1.4 -37.7 -38.9 -119.5 1.7 -85.9 0.8 -128.9 1.6 -95.2 0.8 Hope -201.8 3.0 -189.7 3.5 -176.2 4.2 -171.4 3.8 -174.6 3.1 -18.3 -36.7 -123.0 3.7 -103.2 4.1 -133.4 3.7 -113.3 4.1 Horseshoe -187.4 5.6 -188.2 1.3 -176.4 5.0 -166.8 3.9 -152.4 4.8 -42.2 -24.7 -123.3 1.4 -100.2 4.2 -131.7 1.4 -108.4 4.2 Little Green -189.4 1.1 -190.0 3.3 -178.8 1.4 -152.1 1.9 -146.2 2.6 -52.3 -44.7 -124.3 3.6 -83.3 2.1 -133.7 3.6 -92.7 2.1 Moose -193.7 3.5 -190.3 1.3 -175.9 2.3 -161.6 3.9 -160.9 1.9 -35.0 -38.2 -125.6 1.4 -94.6 4.2 -134.0 1.3 -102.9 4.2 Quiver -180.5 2.0 -178.9 0.8 -176.2 2.4 -153.2 3.5 -149.3 0.8 -35.0 -32.7 -115.2 0.9 -87.5 3.7 -121.8 0.9 -93.9 3.7 Raquette -189.5 0.7 -186.0 0.6 -166.9 0.6 -180.0 3.9 -175.5 4.0 -12.7 -11.6 -120.9 0.7 -114.4 4.2 -129.4 0.7 -122.5 4.2 Sucker -199.6 2.4 -197.3 1.6 -181.4 0.7 -166.9 2.4 -158.9 7.2 -45.6 -39.2 -133.2 1.7 -100.3 2.6 -141.5 1.7 -108.5 2.6 Wolf -183.7 0.1 -178.9 1.6 -168.0 0.9 -156.0 1.8 -153.6 5.0 -29.9 -32.9 -114.2 1.8 -89.6 1.9 -121.8 1.8 -96.9 1.9

144 Chapter 5

Conclusion

The hydrogen isotope values of terrestrial plant waxes are primarily controlled by the isotopic composition of precipitation. Therefore, these molecular biomarkers contain important insight into the magnitude and mechanisms of hydrologic responses to climate perturbations in the past. However, the relationship between precipitation hydrogen isotopes (δDp) and plant wax hydrogen isotopes (δDwax) is modified by processes during wax production in plants and integration in sediments, which limits quantitative reconstruction of δDp based on sedimentary waxes. This dissertation increases our level of understanding of this proxy system from wax production in individual plants to integration of catchment- wide δDwax in sediments. This research also applies these insights to understand controls on sedimentary

δDwax at the regional scale. Specifically, the collective studies constrain: 1) the seasonal timing of leaf wax production and implications for δDp reconstruction from temperate forested settings; 2) sediment bias toward trees and woody plants, indicating that closed basins integrate plant wax primarily via direct leaf fall; and 3) the influence of site-specific vegetation and depositional setting on regional variability in lake sediment δDwax records.

Results from the first study (Chapter 2) indicate that long chain n-alkanes and n-alkanoic acids produced in the same plant differ in key ways, including concentration, timing of synthesis and δD composition.

Specifically, in four out of five tree species, n-alkane concentrations in mature leaves were 2 to 29 times higher than n-alkanoic acid concentrations in the same individuals. Further, while the majority of de novo n-C29 alkane synthesis occurs during a discrete interval as leaves reach full expansion, n-C28 alkanoic acids are produced throughout the growing season. During leaf emergence and expansion, exposure of leaf tissues with undeveloped cuticle drives transpirational D-enrichment of leaf water which, in turn, acts as a primary control on δDwax values as leaves and the cuticle reach maturity. Despite differences in the

145 timing of n-alkane and n-alkanoic acid synthesis, δDwax values for both compound classes are primarily determined during the period of leaf emergence and expansion. This suggests that the δDwax signal from trees in temperate settings may be biased toward the local timing of spring leaf emergence. Prior to autumn leaf fall, interspecies δDn-C29 alkane variability (1σ = 4‰) was lower than for δDn-C28 acid (1 σ = 9‰), suggesting that n-alkanes may be more sensitive tracers of changes in source water δD. During the mature leaf stage, we observed a consistent ~-19‰ offset of εn-C29/xw values relative to εn-C28/xw, underscoring that the two compound classes fractionate shared source water to different extents, and should be interpreted from sedimentary records accordingly.

The second study (Chapter 3) found that long-chain n-alkane concentrations varied by two orders of magnitude among plant species growing in the BLB catchment, while long-chain n-alkanoic acid concentrations were generally lower and more evenly distributed among species. The total range in sedimentary δDwax (6 to 11‰) was far lower than in plants at BLB (77 to 84‰), indicating reduction of biologically-driven δDwax variability in sediments. At the plant level, source water δD could only explain a portion of the overall δDwax differences, which are more likely explained by species-specific differences in biosynthetic fractionation. Based on the molecular distribution and isotopic composition (δD and δ13C) of n-alkanes and n-alkanoic acids, bog sediments are reflective mainly of trees and other woody plants growing in the shoreline area. For both compound classes, direct leaf fall from trees and woody plants growing near the bog was the dominant source of sedimentary lipids. Our observations suggest that sedimentary n-C28 alkanoic acids may reflect waxes from trees more exclusively than n-C29 alkanes, which may also reflect certain woody shoreline species. Therefore, n-C28 alkanoic acid δD in sediments may be more reflective of soil water (i.e., precipitation) accessed by trees, while n-C29 alkane may reflect more mixed water sources including bog water due to the influence of shoreline vegetation on sedimentary n-alkanes. We found a ~30‰ offset in εapp values for n-C29 alkane (-133‰) and n-C28 alkanoic acid (-103‰) in sediments, which is comparable to previous observations in temperate settings.

146 The third study (Chapter 4) measured the δDwax signal in lake surface sediments throughout the

Adirondack Mountains to determine whether catchment-specific factors such as vegetation cover and fluvial complexity affect the fidelity of plant waxes as tracers of water isotopes. We found that n-C28 alkanoic acid molecular and δD composition were significantly correlated with the relative abundance of evergreen vegetation in lake shorelines and fluvial complexity, respectively. By contrast, n-C29 alkanes were insensitive to the same factors. We also found that across basins with diverse vegetation assemblages, the abundance of long-chain n-alkanes was remarkably consistent, while long-chain n- alkanoic acid abundance varied widely. Together, these data suggest that sedimentary n-alkanoic acids are more responsive to site-specific conditions than n-alkanes. The δD offset between compound classes was highly variable among sites and was strongly correlated with sediment C/N ratios, which may indicate an aquatic source for long-chain n-alkanoic acids. Overall, we found larger εapp for n-alkanes (-132 ± 6‰) than for n-alkanoic acids (-106 ± 10‰). Comparison with comparable studies in similar and contrasting biomes suggests that εapp from n-alkanes in forested catchments is stable regardless of catchment-scale aridity, increasing confidence in εapp estimates for locations or periods with forest cover. Overall, this study provides evidence for a significant vegetation effect on sedimentary δDwax (especially n-alkanoic acids) in the absence of a gradient in δDp over the sampled region. This suggests that sedimentary plant waxes are sourced from within a lake catchment rather than mixed at a regional scale, and that sedimentary wax composition may even be sensitive to the vegetation composition in the immediate shoreline area.

This dissertation identifies factors and processes that influence the biological production, spatial integration and regional variation in δDwax signals from temperate settings. The results of this dissertation address areas of lingering uncertainty in this proxy system, identify promising directions for future proxy calibration research, and ultimately will contribute to reducing uncertainties inherent in proxy-based observations of the hydrologic response of the Earth system to climate perturbations in the geologic past.

147