University of Southern California Dissertations and Theses

Isotopic fractionations in plant biomarker molecules with application to paleoclimate

(USC Thesis Other)

Isotopic fractionations in plant biomarker molecules with application to paleoclimate

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ISOTOPIC FRACTIONATIONS IN PLANT BIOMARKER MOLECULES WITH APPLICATION TO PALEOCLIMATE by Hyejung Lee _________________________________________________ A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (GEOLOGICAL SCIENCES) May 2019 Copyright 2019 Hyejung Lee i Abstract The carbon and hydrogen isotopic compositions of plant compounds are important proxies used to reconstruct past vegetation and precipitation. However, many stages between biosynthesis and sample analysis can potentially alter the original environmental isotopic signal of interest and obscure the geological implications. Experimental approaches are used in this thesis to constrain isotopic fractionations in plant wax compounds caused by changes in leaf physiology and laboratory protocols. In chapter 2, I use a genetically modified model organism to isolate the effect of changing stomatal density on the carbon and hydrogen isotopic compositions of plant n- alkanes. Chapter 3 investigates the rate of hydrogen isotope exchange during phthalic acid methylation, which is an important step in the isotopic analysis of plant n-alkanoic acids. Chapters 4 and 5 focus on development and application of a relatively new biomarker approach: isotopic analysis and quantification of methoxyl groups within lignin. In chapter 4, a liquid injection method is used to quantify and determine hydrogen isotopic composition of iodomethane evolved from a suite of lignin-bearing samples. Results show that concentrations of methoxyl groups capture changes in wood diagenesis and catagenesis. Phenol and wood samples were surveyed with potential to serve as isotopic standards. Chapter 5 analyzes lignin characteristics and the carbon and hydrogen isotopic composition of bulk wood and lignin methoxyl groups of wood fragments found in the Bengal Fan spanning the last 19 million years. This thesis presents fundamental isotope biogeochemistry efforts to constrain plant biomarkers as proxies for paleoclimate and to study an overlooked wood contribution to the organic carbon burial in the global carbon cycle. ii Acknowledgements I thank the key people that made completing this thesis not only possible but a rich experience. First of all, Sarah thank you for your guidance throughout the years. You have never failed to respond to my questions, updates, late night emails, etc. and I thank you for your commitment to me as an advisor. Even when projects didn‘t go as planned, you still challenged me to push forward, think of new ideas, find new instrumental parts and did everything you can to support me along this process. I have been and will always be thankful and inspired to mimic your level of work ethic and engagement. Next, I thank all of the members of the leaf wax lab, past and present, for being the best lab mates a PhD student could ask for. The camaraderie in this group in and out of the cheese grater building has been the main source of support and encouragement throughout my PhD. Thank you Christine, Hannah, Mark, Emily, Patrick, Camilo, Efrain, Annie, Elias, and all of the long/short term leaf wax lab mates for refining my thought process and going through all of the highs and lows together. I also thank my committee and the USC Earth Science department for mentorship and genuine care. Travis (and the Williams lab members, especially Yao) thank you for guiding me in the chemistry aspects of my research and taking in a geochemist neighbor to your lab. Frank, thank you for serving both as my committee member (qual and defense) and as the departmental grad student advisor. Josh and Dave, thank you for your rigorous input towards shaping my projects. Also, I thank the department staff for all of the logistical help over the years. I thank my coauthors outside of USC who provided countless inputs during data analyses and manuscript revisions. Valier, thank you for hosting me at WHOI and for your guidance in analyzing an extraordinary set of buried wood samples in the Bengal Fan. Yvonne, Christian, Albert, Arndt, Alex, Jess, and Leo, thank you for your collaboration and valuable feedback toward publications. Finally I‘d like to thank my family and friends near and far for their support during PhD. Joe, thank you for being an amazing boyfriend, now husband, and for always knowing how to make me laugh especially during these PhD years. Umma, appa and unni, thank you for supporting me throughout this process so well even from afar. Also, Tapestry LA church and friends thank you for being a loving community of people that made LA feel like home away from home. iii Table of Contents ABSTRACT…………………………………………………………………………………. i ACKNOWLEDGEMENTS…………………………………………………………………….. ii LIST OF FIGURES……………………………………………………………………………. vii LIST OF TABLES……………………………………………………………………………. x CHAPTER 1: INTRODUCTION………………………………………………………….. 1 1.1 Two classes of plant biochemicals of geological relevance……………… 2 1.1.1 Plant wax n-alkyl lipids…………………………………………… 2 1.1.2 Lignin ……………………………………………………………… 3 1.2 Plant isotope biogeochemistry ……………………………………...…….. 5 1.2.1 Carbon isotope biogeochemistry ………………………………… 5 1.2.2 Hydrogen isotope biogeochemistry……………………………… 6 1.3 Dissertation chapters and research objectives………………………………7 References………………………………………………………………………… 9 CHAPTER 2: CARBON AND HYDROGEN ISOTOPIC EFFECTS OF STOMATAL DENSITY IN ARABIDOPSIS THALIANA ………….................................................................................. 10 Abstract…………………………………………………………………………… 10 2.1 Introduction………………………………………………………………… 11 2.2 Methods…………………………………………………………………… 17 2.2.1 Variant selection…………………………………………………. 17 2.2.2 Growth conditions………………………………………………….. 17 2.2.3 Plant water extractions and isotopic measurement………………… 18 2.2.4 Lipid extraction …………………………………………………… 19 2.2.5 Compound-specific analytical methods…………………………… 19 2.2.5.1 Compound specific molecular abundance distributions…… 19 2.2.5.2 Compound specific hydrogen and carbon isotopic analysis 20 2.2.6 Bulk carbon isotope analysis………………………………………. 22 2.3 Results…………………………………………………………………….. 22 2.3.1 Plant waters……………………………………………………….. 22 2.3.2 Molecular abundance distribution………………………………… 23 2.3.3 Hydrogen isotopic compositions…………………………………. 24 2.3.4 Carbon isotopic compositions……………………………………. 26 2.4 Discussions………………………………………………………………. 29 2.4.1 Implications of STOMAGEN mutations on leaf water enrichment 29 iv 2.4.2 Implications of STOMAGEN for plant wax biosynthesis and isotopic composition ……………………………………………………………… 30 2.4.3 Carbon isotopic composition………………………………………. 32 2.4.4 Implications for paleoenvironmental reconstructions using the plant wax isotopic compositions ……………………………………………….….…. 34 2.5 Conclusions……………………………………………………………….. 37 Acknowledgements……………………………………………………………….. 38 References………………………………………………………………………… 39 CHAPTER 3: COMPARISON OF THREE METHODS FOR THE METHYLATION OF ALIPHATIC AND AROMATIC COMPOUNDS ………….……….……….……….……….……….…. 44 Abstract……………………………………………………………………………. 44 3.1 Introduction……………………………………………………………….. 45 3.1.1 Acid catalyzed esterification ……………………………………… 46 3.1.2 Non-acidic coupling reagents to promote esterification…………… 48 3.1.3 Method comparison……………………………………………… 49 3.2 Experimental………………………………………………………………. 50 3.2.1 Phthalic acid and methanol of known isotopic composition ……… 50 3.2.2 The acidic methods: HCl and CH 3 COCl………………………… 51 3.2.3 The non-acidic method: EDCI & DMAP…………………………. 52 3.2.4 GC/MS/FID analysis ……………………………………………. 52 3.2.5 GC/IRMS analysis ……………………………………………. 53 3.2.6 1 H/ 2 H exchange rate using 1 H-NMR……………………………… 55 3.3 Results and discussion…………………………………………………….. 56 3.3.1 Product yield……………………………………………………… 56 3.3.1.1 Yield for acidic methods…………………………………. 56 3.3.1.2 Yield for carbodiimide method…………………………… 57 3.3.2 Isotopic analyses…………………………………………………. 59 3.3.2.1 Is acid-catalyzed exchange a cause of 2 H-enrichment?…… 60 3.3.2.2 Is evaporation a cause of 2 H-fractionation?…………….… 60 3.3.2.3 Is acid-catalyzed exchange a cause of D-enrichment? …… 61 3.3.3 Implications for aromatic and aliphatic acid esterification ……….. 62 3.4 Conclusions………………………………………………………………… 65 Acknowledgements………………………………………………………………… 66 References…………………………………………………………………………. 67 v CHAPTER 4: QUANTITATIVE DETERMINATION OF LIGNIN METHOXYL CONCENTRATION AND ISOTOPIC COMPOSITION ……………………...………………...……………….. 70 Abstract…………………………………………………………………………… 70 4.1 Introduction………………………………………………………………… 71 4.1.1 Analytical considerations…………………………………………. 72 4.2 Materials and methods……………………………………………………. 73 4.2.1 Materials…………………………………………………………… 73 4.2.2 Iodomethane isolation and injection……………………………… 74 4.2.3 Identification and quantification by GC – MS/FID………………. 75 4.2.4 Hydrogen isotopic analyses……………………………………… 75 4.2.5 Carbon isotopic analyses……………………………………………77 4.2.6 Lignin phenol analyses…………………………………………… 78 4.3 Results……………………………………………………………………… 79 4.3.1 Calibration for iodomethane quantification……………………… 79 4.3.2 Calibration for methoxy group quantification……………………. 81 4.3.3 Iodomethane quantification in heterogeneous wood samples…… 82 4.3.4 Methoxyl quantification in lignin-bearing sedimentary deposits… 83 4.3.5 Lignin phenol concentrations and ratios of the USC Standards … 85 4.3.6 Hydrogen isotopic composition of the USC Standards…………… 86 4.3.7 Carbon isotopic composition of the USC Standards ……………… 88 4.4 Discussion………………………………………………………………… 89 4.4.1 Recommended practices for quantification of iodomethane……… 89 4.4.2 Lignin methoxyl δD and δ 13 C analyses…………………………… 90 4.4.3 Fate of methoxyl during early diagenesis and maturation………… 91 4.4.4 Methoxyl concentrations in peats under different ecosystems …… 92 4.5 Conclusions………………………………………………………………… 93 Acknowledgements………………………………………………………………… 95 References…………………………………………………………………………. 96 CHAPTER 5: SUSTAINED WOOD BURIAL IN THE BENGAL FAN OVER THE LAST 19 MILLION YEARS …………...……………..……………..……………...……………….. 100 Abstract…………………………………………………………………………… 100 5.1 Introduction……………………………………………………………….. 101 5.2 Results …………………………………………………………………… 104 5.2.1 Abundance of wood in the Bengal Fan…………………………… 104 5.2.2 Lignin evidence for plant type and degradation…………………… 106 5.2.3 Bulk wood carbon concentration and isotopic composition……… 107 vi 5.2.4 Lignin methoxyl hydrogen isotopic composition………………… 110 5.3 Discussion………………………………………………………………….. 111 5.3.1 Wood burial in the Bengal Fan…………………………………… 111 5.3.2 Lowland dominated, fresh wood export…………………………. 114 5.3.3 Vegetation change in the G-B catchment………………………… 115 5.3.4. Constraints on the isotopic composition of precipitation…………. 119 5.4 Conclusion………………………………………………………………… 120 5.5 Materials and methods…………………………………………………… 122 Acknowledgements………………………………………………………………… 123 References…………………………………………………………………………. 124 CHAPTER 6: DISSERTATION CONCLUSIONS…………………………………. 129 APPENDIX………………………………………………………………………… 134 A.1 Supplementary information for Chapter 3……………………………….. 134 A.1.1 Protocol for phthalic acid methylation…………………………… 134 A.1.1.1 Purpose…………………………………………………… 134 A.1.1.2 Necessary materials……………………………………….. 134 A.1.1.3 Procedure………………………………………………….. 135 A.1.1.4 Isotope mass balance……………………………………… 136 A.1.1.5 Error propagation………………………………………….. 137 A.1.2 References………………………………………………….……… 142 A.2 Supplementary information for Chapter 4……………………………….. 143 A.2.1 References…………………………………………………………. 144 A.3 Supplementary information for Chapter 5……………………………….. 145 A.3.1 Extended methods………………………………………………… 145 A.3.1.1 Bulk wood elemental and isotopic analyses……………… 145 A.3.1.2 Lignin methoxyl hydrogen isotopic analyses………………145 A.3.1.3 Lignin phenol characterization…………………………… 147 A.3.2 Extended supplementary data discussion…………………………. 148 A.3.2.1 Organic carbon content in wood………………………….. 148 A.3.2.2 Variations in atmospheric carbon isotopic composition…. 149 A.3.3 References………………………………………………………….. 173 vii List of Figures Figure 1.1 Lignin structure and methoxyl groups…………………………………… 4 Figure 1.2 Hydrogen isotope fractionation scheme for leaf wax lipids and lignin methoxyl group ……………………………………………………………………… 7 Figure 2.1 Scanning Electron Microscope image of leaf abaxial surface for STOMAGEN mutated A. thaliana plants………………………………………………… 17 Figure 2.2 Abundances of n-alkanes and n-alkanoic acids by chain length for A. thaliana variants ………………………………………………………………….… 24 Figure 2.3 D of leaf water and wax for A. thaliana variants ………………………… 25 Figure 2.4 Leaf water enrichment, net fractionation and biosynthetic fractionations in A. thaliana variants …………………………………………………………... 25 Figure 2.5  13 C values of bulk and plant wax n-alkanes from A. thaliana variants…... 26 Figure 2.6 Comparison of stomatal densities of A. thaliana variants with those of Cenozoic fossil leaves………………………………………………………………… 36 Figure 2.7 Comparison of stomatal densities of A. thaliana variants with those of Cenozoic fossil leaves with respect to pCO 2 ………………………………………… 37 Figure 3.1 Acid-catalyzed 1 H/ 2 H exchange in dimethyl phthalate …………….…… 48 Figure 3.2 Structures of EDCI and DMAP ………………………………………….. 49 Figure 3.3 Dimethyl phthalate isolated yield and isotopic results ……………………. 62 Figure 3.4 HCl-mediated 1 H/ 2 H exchange rate in dimethyl phthalate ……………….. 64 Figure 3.5 Methyl octacosanoate isolated yield, carbon and hydrogen isotopic composition ……………………………………………………………… 65 Figure 4.1 Comparison of the headspace and liquid injection methods of iodomethane quantification by GC–FID………………………………………………… 80 viii Figure 4.2 Comparison of the iodomethane standard with the quantification calibration of iodomethane produced via the Ziesel method …………………………….. 81 Figure 4.3 The methyl concentrations of Miocene age Polish lignites analyzed using the liquid method ……………………………………………………………… 84 Figure 4.4 Cross plot of hydrogen and carbon isotopic composition of iodomethane, and the methoxyl groups released from phenolic compounds, wood and peat…….. 87 Figure 5.1 Examples of wood deposits in IODP Exp 354 cores ……………………… 103 Figure 5.2 Mass and age-depth plot of buried wood deposits in IODP Expedition 354 Sites U1450, U1451, U1452, U1453, U1454, U1455 in the Bengal Fan ……… 105 Figure 5.3 Isotopic composition and phenol ratio of wood fragments in the Bengal Fan from the Expedition 354………………………………………………………… 107 Figure 5.4 Comparison of δD methoxyl and δ 13 C wood for paired analyses within sedimentary horizons …………………………………………………………………... 111 Figure A.1 Maps of Exp354 IODP cores analyzed in this study ……………..………. 150 Figure A.2 Wood fragments and barren samples displayed in approximate sampling location on the Exp 354 stratigraphic columns . .… .……………………………… 151 Figure A.3 Occurrence and isotopic composition of wood in IODP Expedition 354 Sites U1450, U1451, U1452, U1453, U1454, U1455 in the Bengal Fan ……….. 152 Figure A.4 Ratios and concentrations of major lignin phenol groups in Bengal Fan buried wood fragments. …………………………………………………………. 153 Figure A.5 Bulk organic carbon properties of Bengal Fan wood fragments …………. 154 Figure A.6 Monthly δD precip plotted against elevation from a selection of Global Network of Isotopes in Precipitation (GNIP) stations in the G – B catchment ………. 154 Figure A.7 Modern C 3 plant survey of woody branch tissue δ13C values across an elevation transect in Arunachal Pradesh, India and Central Nepal …………………. 155 ix Figure A.8 Box plots comparing δ13C of modern plant survey across an elevation transect in Arunachal Pradesh, India and Central Nepal and GNIP δD precip data with Exp 354 buried wood isotope results.………………………………………………. 156 x List of Tables Table 2.1 Hydrogen and carbon isotopic compositions of A.thaliana variants and calculated fractionations. ………………………………………………………………27 Table 2.2 Plant wax n-alkane and n-alkanoic acid molecular abundance distribution of A.thaliana variants ………………………………………………………… 28 Table 3.1 Hydrogen and carbon isotope values calculated for C 28 and C 16 FAMEs and propagated uncertainties based on dimethyl phthalate……………………. 55 Table 4.1 Materials with known and unknown methyl mass analyzed using the liquid method .………………………………………………………………….… 82 Table 4.2 Hydrogen isotope results of the standard suite in both headspace and liquid method and carbon isotope results in headspace method ………………… 86 Table A.1 Comparison of dimethyl phthalate yield and carbon and hydrogen isotopic composition by three methylation approaches.…………………………… 139 Table A.2 Comparison of methyl octacosanoate yield and carbon and hydrogen isotopic composition by three methylation approaches .…………………………… 139 Table A.3 Individual replicate tests of methylation of phthalic acid ………………… 140 Table A.4 Yields of dimethyl phthalate with variations to the stated experimental protocol for EDCI/DMAP method.………………………………………………… 141 Table A.5 Miocene age lignite samples from Belchatow deposit in Poland from Drobniak and Mastalerz, 2006 ……………………………………………………… 143 Table A.6 Bengal Fan wood mass recovered.………………………………………… 157 Table A.7 Bengal Fan wood lignin phenol concentrations…………………………… 159 Table A.8 Bengal Fan wood organic carbon concentrations and carbon isotopic composition………………………………………………………………… 162 Table A.9 Bengal Fan wood lignin methoxyl hydrogen isotopic composition……… 166 Table A.10 Kml spreadsheet of locations referenced in this study. …………………… 167 xi Table A.11 Arunachal Pradesh (India) and Central Nepal elevation transects to survey of woody plant carbon isotopic composition. ..……………………………… 170 Table A.12 Compilation of published bulk wood carbon isotopic compositions in tropical and temperate forests....…… ………………………………………………… 172 1 Chapter 1 Introduction This dissertation describes the development of plant molecular proxies used to reconstruct precipitation and vegetation in the geological past. I study two classes of plant biomarkers: the plant wax n-alkyl lipids found on the soft tissues of plants, and lignin, a major structural component for support in trunks and other tissues and the second most common biochemical made by plants after cellulose. The second chapter takes an experimental approach – with a genetic mutant of the model organism Arabidopsis – to constrain a major uncertainty in the plant leaf wax proxy: how changes in plant physiology, specifically the density of stomatal pores on the leaf, affect the carbon and hydrogen isotopic composition of plant wax n-alkanes. The third chapter considers whether laboratory procedure affects the stable carbon and hydrogen isotopic composition of long chain n-alkanoic acids. The fourth chapter describes a new proxy based on the abundance and isotopic composition of lignin methoxyl groups and introduces a new method for analysis to improve precision in both quantitative recovery and hydrogen isotopic determinations. I then apply this method to various geological applications including a modern elevation gradient from the Andes-Amazon, and fossil wood from the Miocene of Antarctica. Finally, I study the hydrogen and carbon isotopic composition and the lignin composition of wood fragments found in the sediments of the Bengal Fan, to study the origins of wood transported thousands of kilometers from growth location to deposition in the fan. I use these wood fragments to reconstruct changes in sourcing and changes in monsoon precipitation over the last 19 million years encompassing the global climatic changes of the Miocene, Pliocene and Pleistocene. 2 1.1 Two classes of plant biochemicals of geological relevance 1.1.1 Plant wax n-alkyl lipids Plant n-alkyl lipids such as long chain n-alkanes and n-alkanoic acids form the cuticular wax of terrestrial vascular plants. These biomarkers are useful for paleoclimate reconstructions because their stable isotopic composition reflects information about the environment and they have high preservation potential in terrestrial, lacustrine and marine sediments over geologic timescales. The stable carbon isotopic composition ( 13 C) of lipids is primarily shaped by the  13 C of atmospheric carbon dioxide and isotopic fractionation related to photosynthetic pathway, with applications including the reconstruction of canopy closure. The stable hydrogen isotopic composition (D) of lipids reflects that of source water with large fractionations during biosynthesis. For terrestrial biomarkers, the source water is meteoric water providing a valuable tracer of isotopes in precipitation. Large differences between biosynthesis pathways for different organisms and different metabolisms mean that compound specific approaches (rather than bulk analyses) are necessary to extract information about source water. Another challenge with some compounds is the presence of exchangeable hydrogens (such as on -OH or -NH bonds), whereas C–H bonds are strong covalent bonds and do not exchange with available hydrogens in the environment under 150 C (Schimmelmann et al., 2006). Thus, selecting hydrocarbon molecules (e.g. plant waxes), or intramolecular positions (-OCH 3 within lignin) avoids the problem of exchange with surrounding water that would otherwise erase the original source water signal of interest. 3 From synthesis of the lipids in plants, to diagenesis and catagenesis in geological storage, to sample preparation and analysis in the laboratory, there are many opportunities for isotopic fractionations, but some processes have not yet been quantified. In this thesis I take a fundamental isotope biogeochemistry approach and experimentally determine the magnitudes of fractionations under controlled conditions in the laboratory. In the first case, I investigate how the number of stomata on the plant leaf, affect the exchange of CO 2 and H 2 O to influence the  13 C and D of plant leaf wax n-alkanes. In the second case I test whether common laboratory preparative procedures that involve hot acidic reactions modify the isotopic compositions of the compound of interest. I compare methods, and I test both reasonable conditions as well as experimentally constrain the rates of exchange, contributing a fundamental rate constant for n- alkanoic acids. I also make practical recommendations for organic geochemistry sample preparation protocols. Each of these experimental studies helps in different ways to constrain potential isotopic effects in the environment or the laboratory. 1.1.2 Lignin Lignin is the second most abundant plant molecule after cellulose. It provides structural support to plants and is resistant to degradation. Lignin is a large, heterogeneous polymer and it contains both exchangeable and non-exchangeable hydrogen within its structure. The non-exchangeable H are found within the methoxyl groups (-OCH 3 ; circled in red on Figure 1). Lignin methoxyl hydrogen isotopic composition represents an intramolecular target within a plant biomarker that has promise for application to reconstruct precipitation isotopic composition from a variety of archives including tree rings on living trees, fossil wood fragments and buried logs in sedimentary deposits, and the lignin in peat, lignite and coal along the spectrum of geological alteration of ancient buried forest wood. 4 Fig. 1.1. Example lignin structure with methoxyl groups circled in red. Initial lignin methoxyl studies found a high correlation between lignin methoxyl D and source water D over a latitudinal transect (Keppler et al., 2008) as well as across a freshwater to saltwater gradient in a coastal forest (Feakins et al., 2013) with several species represented in those surveys. Additional studies have investigated the different biosynthetic fractionations between species (Anhauser et al., 2017) and degradation over time (Anhauser et al., 2015). As this proxy is still in its early stages, method development and calibration efforts are needed to more comprehensively test and develop the proxy towards developing the potential and pitfalls for various possible paleoclimate reconstruction applications. 5 1.2 Plant isotope biogeochemistry Since the advent of mass spectrometry in the mid-20 th Century, and the development of compound specific isotopic analyses of plant biochemicals in the following decades, it has been possible to study plant biochemistry based on carbon and hydrogen isotopic fractionations. Each of these reveal aspects of plant biosynthesis and environment and thus have found application in the field of isotope biogeochemistry. 1.2.1 Carbon isotope biogeochemistry The substrate for carbon in plants is atmospheric CO 2 . This means that the carbon isotopic composition of plants reflects that of the atmosphere, whether microscale variations under the canopy or large-scale atmospheric variations over geological time. Fractionation occurs during diffusion into the leaf (a) and a larger step during photosynthesis (b), modified by the gradient between pCO 2 inside the leaf relative to the atmosphere:  13 C = a + (b-a) C internal /C atmos (1) (Farquhar et al., 1989). This theory is well-established for modern plants, yet one caveat to geological application is that as pCO 2 varies, plants adapt by varying their numbers of stomata. I experimentally test how the density of stomata affect the carbon and hydrogen isotopic composition of plant wax n-alkanes in Chapter 2. These carbon isotopic fractionations (eq. 1) also get carried into other plant biochemicals and we conducted some of the first analyses of lignin methoxyl  13 C and did so in a large-scale, multi-species, Andes-Amazon forest transect. We find a trend in  13 C of lignin methoxyl as fractionations vary with altitude, not previously reported in tree wood, but echoing patterns in bulk leaves and leaf waxes in the same transect (Wu et al., 2017). This  13 C trend is a consequence of changing atmospheric pressure upslope 6 and is also relevant to geological atmospheric pCO 2 excursions (Foster et al., 2017). Intriguingly, we see a bigger signal between canopy and understory reflecting respired CO 2 uptake in the understory, which potentially enables the use of wood fragments as a proxy for canopy closure, similar to that proposed for pedogenic carbonates (Cerling et al., 2011). For paleoclimate applications, carbon isotopes of plants are mostly used to reconstruct changes in the  13 C of atmospheric carbon dioxide (so-called carbon isotope excursions such as that of the Paleocene- Eocene Thermal Maximum) or widespread changes in the photosynthetic pathway with C 4 grasses expanding in the late Miocene (Feakins et al., 2013) with suggestions that plant  13 C can reconstruct pCO 2 remaining controversial (Schubert and Jahren, 2018). The fundamental calibration efforts in my thesis help to resolve how large the respective signals can be. 1.2.2 Hydrogen isotope biogeochemistry The ultimate substrate of hydrogen to the plant is water, typically meteoric water. Water contains both O and H, and I focus on H which is incorporated into plant lipids as well as into lignin. We seek H that is in C–H bonds where that H is non-exchangeable whether in the n-alkyl lipids in plant waxes or in the methoxyl (–OCH 3 groups) on lignin. H atoms in these positions, contain the signatures of source water, taken up through roots without fractionation, transported through the xylem up to the leaves. The long chain n-alkyl lipids are synthesized in the leaf where the isotopic composition of water can be altered prior to synthesis. During photosynthesis when stomata are open to obtain CO 2 , plants can lose their leaf water through evapotranspiration (Chapter 2). The degree of water loss and the resulting change in the isotopic composition of leaf water vary with relative humidity (Kahmen et al., 2013; Feakins et al., 2010) and is a step in the fractionation scheme for the leaf wax proxy (Figure 1.2). As lignin is synthesized in the xylem, it 7 is not subject to leaf water transpiration (Figure 1.2). Thus lignin has potential in the quest for a quantitative proxy for isotopes in precipitation (Chapter 4). Fig. 1.2. Hydrogen isotope fractionation scheme for leaf wax lipids and lignin methoxyl group. Modified from Sachse et al., 2012 with addition of the lignin methoxyl approach. 1.3 Dissertation chapters and research objectives Chapter 2. Stomatal density of plants changes with pCO 2 which swings significantly over the geologic timescale. However, its effects on isotopic compositions of plant waxes have not been investigated. Here I used genetic mutations on a plant model organism to provide the first controlled tests of stomatal density and explore the implications of using plant wax D and  13 C as paleoenvironmental proxies. Chapter 3. All n-alkanoic acids from plant waxes must be derivatized for gas chromatography analyses. A common procedure is to methylate the carboxylic group in the fatty acid using acidic catalysts, which have potential to cause hydrogen isotopic fractionation. I compare three 8 methylation protocols for aliphatic and aromatic acids and analyze the hydrogen and carbon isotopic fractionations using isotope ratio mass spectrometry and nuclear magnetic resonance. Chapter 4. Quantification and isotopic composition of methoxyl groups in lignin can be potentially used to study lignin maturation and precipitation isotope. Previously, lignin methoxyl groups were analyzed as iodomethane, a volatile gas, and I suggest a liquid injection method to securely quantify and analyze the isotopic compositions of methoxyl groups in a variety of standards and samples including lignin phenols, kraft lignin, wood, peat, lignite and coal. Chapter 5. Fragments of woody debris were recovered in the IODP Expedition 354 cores located mid Bengal Fan spanning the last 19 million years. I use lignin characteristics and the carbon and hydrogen isotopic compositions of wood and lignin methoxyl groups to constrain the provenance of wood and study the coarse organic carbon export regime over the Neogene. 9 References Anhäuser T., Greule M. and Keppler F. (2017) Stable hydrogen isotope values of lignin methoxyl groups of four tree species across Germany and their implication for temperature reconstruction. Science of the Total Environment 579, 263–271. Anhäuser T., Greule M., Zech M., Kalbitz K., McRoberts C. and Keppler F. (2015) Stable hydrogen and carbon isotope ratios of methoxyl groups during plant litter degradation. Isotopes in Environmental and Health Studies 51, 143–154. Cerling T. E., Wynn J. G., Andanje S. A., Bird M. I., Korir D. K., Levin N. E., Mace W., Macharia A. N., Quade J. and Remien C. H. (2011) Woody cover and hominin environments in the past 6 million years. Nature 476, 51–56. Farquhar G. D., Ehleringer L. 1 R. and Hubick K. T. (1989) CARBON ISOTOPE DISCRIMINATION AND PHOTOSYNTHESIS., Feakins S. J., Ellsworth P. V. and Sternberg L. da S. L. (2013) Lignin methoxyl hydrogen isotope ratios in a coastal ecosystem. Geochimica et Cosmochimica Acta 121, 54–66. Feakins S. and Sessions A. (2010) Controls on the D/H ratios of plant leaf waxes in an arid ecosystem. Geochimica et Cosmochimica Acta. Foster G. L., Royer D. L. and Lunt D. J. (2017) Future climate forcing potentially without precedent in the last 420 million years. Nature Communications 8, 14845. Kahmen A., Hoffmann B., Schefuß E., Arndt S. K., Cernusak L. A., West J. B. and Sachse D. (2013) Leaf water deuterium enrichment shapes leaf wax n-alkane δD values of angiosperm plants II: Observational evidence and global implications. Geochimica et Cosmochimica Acta 111, 50–63. Keppler F. and Hamilton J. T. G. (2008) Tracing the geographical origin of early potato tubers using stable hydrogen isotope ratios of methoxyl groups†. Isotopes in Environmental and Health Studies 44, 337–347. Schimmelmann A., Sessions A. L. and Mastalerz M. (2006) Hydrogen isotopic (D/H) composition of organic matter during diagenesis and thermal maturation. Annual Review of Earth and Planetary Sciences 34, 501–533. Schubert B. A. and Jahren A. H. (2018) Incorporating the effects of photorespiration into terrestrial paleoclimate reconstruction. Earth-Science Reviews 177, 637–642. 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. Geochimica et Cosmochimica Acta 206, 1–17. 10 Chapter 2 Carbon and hydrogen isotopic effects of stomatal density in Arabidopsis thaliana This chapter was published in 2016 as: Lee H, Feakins SJ, Sternberg L da SL (2016) Carbon and hydrogen isotopic effects of stomatal density in Arabidopsis thaliana. Geochimica et Cosmochimica Acta 179:275–286. Abstract Stomata are key gateways mediating carbon uptake and water loss from plants. Varied stomatal densities in fossil leaves raise the possibility that isotope effects associated with the openness of exchange may have mediated plant wax biomarker isotopic proxies for paleovegetation and paleoclimate in the geological record. Here we use Arabidopsis thaliana, a widely used model organism, to provide the first controlled tests of stomatal density on carbon and hydrogen isotopic compositions of cuticular waxes. Laboratory grown wildtype and mutants with suppressed and overexpressed stomatal densities allow us to directly test the isotope effects of stomatal densities independent of most other environmental or biological variables. Direct hydrogen isotope (H/D) measurements of plant waters and plant wax n-alkanes allow us to directly constrain the isotopic effects of leaf water isotopic enrichment via transpiration and biosynthetic fractionations, which together determine the net fractionation between irrigation water and n-alkane hydrogen isotopic composition. We also measure carbon isotopic fractionations of n-alkanes and bulk leaf tissue associated with different stomatal densities. We 11 find offsets of +15‰ for D and −3‰ for  13 C for the overexpressed mutant compared to the suppressed mutant. Since the range of stomatal densities expressed is comparable to that found in extant plants and the Cenozoic fossil record, the results allow us to consider the magnitude of isotope effects that may be incurred by these plant adaptive responses. This study highlights the potential of genetic mutants to isolate individual isotope effects and add to our fundamental understanding of how genetics and physiology influence plant biochemicals including plant wax biomarkers. Keywords: plant wax; carbon isotopes; hydrogen isotopes: Arabidopsis; STOMAGEN. 2.1 Introduction Stomata on plant leaves mediate both the uptake of CO 2 for assimilation in photosynthesis and loss of water via transpiration. In order to maximize CO 2 gas exchange per unit water loss, plants adjust the number of stomatal pores per epidermal cell during early leaf ontogeny and then in the full-size leaf they regulate the degree of opening of stomata on a diurnal basis according to the changing environment (Woodward, 1987; Raven, 2002). Maximum stomatal size and density on mature leaves are inversely correlated due to the biophysical limit of number of stomata at a certain size that can fit within a unit area (Franks and Beerling, 2009). This trade-off must be balanced, and over geological time, changes in atmospheric pCO 2 are the major variable controlling the large scale variations in stomatal distribution (Beerling et al., 1998; Hetherington and Woodward, 2003). Within the last 100 Myr, the long-term decrease in atmospheric pCO 2 has led to a reduction in stomatal size (Franks and Beerling, 2009), which made it physiologically possible for plants to increase stomatal density in response to low pCO 2 during glacial periods 12 (Rundgren and Bjorck, 2003). Changes in pCO 2 during the Cenozoic (last 65 Myr) have been reconstructed from analysis of stomatal densities, normalized for the number of epidermal cells, (i.e. the stomatal index; Woodward, 1987; Wagner et al., 1996; Beerling et al., 2009; Beerling and Royer, 2011). These changing stomatal densities over the Cenozoic, would likely have changed the transpiration and photosynthetic capacity of the leaf (Farquhar and Sharkey, 1982). By controlling the fluxes of CO 2 and water exchange between the atmosphere and plant, stomata are predicted to influence the isotopic effects associated with the openness of these internal pools. All else being equal, denser stomata should allow for a higher concentration of CO 2 within the leaf as the demand for photosynthesis can be more readily met in the more open exchange with the atmosphere. With a higher concentration of CO 2 inside the leaf, Rubisco can be more discriminative against the heavier carbon isotope (Farquhar et al., 1989), as shown by the following equation  (1) where  13 C is the positive discrimination shown by the plant against 13 C, a is the fractionation due to diffusion (4.4‰), b is the discrimination by Rubisco (27‰), C a is the ambient CO 2 concentration of the atmosphere and C i that within the leaf. The C i /C a ratio modulates the carbon isotopic fractionation between the ambient CO 2 and bulk leaf tissue, and stomatal density should be a key factor in regulating that. Further isotope effects take place within biosynthesis steps specific to individual biochemical components of bulk tissue, including those relevant to the plant wax n-alkanes and n-alkanoic acids (DeNiro and Epstein, 1977). Stomata also regulate transpiration. During transpiration, there is enrichment of leaf water 18 O and D at the site of evaporation near stomata (Cernusak et al., 2005; Barbour, 2007; Kahmen et 13 al., 2013). Evaporative enrichment at the site of evaporation ( e ) is a function of the equilibrium fractionation (*), kinetic fractionation ( k ) and the difference between the isotopic composition of vapor and xylem water and the vapor pressure inside (e i ) and outside (e a ) the leaf and for oxygen isotopes may be approximated by: (2) as a variation of the Craig-Gordon model (Farquhar and Lloyd, 1993). For hydrogen isotopes, which incur larger isotopic fractionations because of the greater proportional mass difference of the isotopes, this equation is more completely formulated as: * ( ) + (3) (Farquhar et al., 2007; Cernusak et al., 2015). With more sites of evaporation per unit lamina area, greater evaporative enrichment in lamina water is expected in the overexpressed mutant in this study. However, measured bulk leaf water isotopic composition is only partially affected by evaporation, as it also contains unenriched water in veins (Roden and Ehleringer, 1999). The Craig-Gordon model overestimates leaf water enrichment in many studies and has therefore been modified (Péclet modified Craig-Gordon model) in order to account for leaf water gradients and flows, and in particular the ratio of advection of unenriched water into the lamina from veins versus diffusion of evaporatively enriched water from near stomata into the lamina, in order to better fit observations of bulk leaf water isotopic compositions (Barbour et al., 2003; Kahmen et al., 2011; Larcher et al., 2014). Stomatal densities influence many interacting leaf water fractionation factors as discussed by Larcher et al., (2014) based on theoretical and empirical measurements. In particular, Larcher et al. found decreased distances between stomata to shorten 14 the effective diffusion path length leading to leaf water isotopic enrichment, but with the potential to be counteracted by the tendency to increase transpiration rates. Stable isotopes of carbon ( 13 C), hydrogen (D) and oxygen ( 18 O) have long been used to study plant biochemistry, because these measurements provide important information about photosynthetic pathway (Farquhar et al., 1989; Ehleringer et al., 1997), biosynthesis pathway (Sternberg et al., 1984) and water regulation (Farquhar and Richards, 1984; Farquhar and Cernusak, 2005). Recently, plant waxes have become powerful targets of stable isotope analysis in the study of biogeochemistry and paleoclimate because certain wax components, namely the n-alkanes and n-alkanoic acids, persist beyond the lifetime of the plant and make their way into the geological record (Huang et al., 2002; Schefuss et al., 2005; Tierney et al., 2008). Plant wax D values have been shown to record those of source water, usually precipitation, with a large biological fractionation (Huang et al., 2004; Hou et al., 2008). As precipitation isotopes vary greatly around the world, this dominates the isotopic signals that are recorded (Sachse et al., 2012). Plant wax  13 C values primarily reflect the carbon fixation pathway, with  13 C of atmospheric CO2 as a starting point (which is globally relatively uniform). (Schefuss et al., 2005; Diefendorf et al., 2011). The hydrology and vegetation of many regions and timescales have been interpreted and studied through analysis of D and  13 C plant waxes (Pagani et al., 2006; Tipple and Pagani, 2010; Feakins et al., 2012). However, there remains scope to improve our understanding of plant wax isotopic proxies and to test how variables other than the primary controls of substrate isotopic composition and biosynthetic fractionation might come into play over geologic time. Given the well-known variations in stomatal density in extant plants and in fossil plants across the Cenozoic (Beerling et al., 1993; Vandewater et al., 1994), we seek to test whether these stomatal density need to be taken into account when interpreting the isotopic 15 composition of plant waxes, and if stomatal densities do exert significant controls whether these isotope effects could be accounted for through corrections based on fossil leaf evidence for stomatal densities. Here we use a model organism Arabidopsis thaliana, to test the effect of stomatal density on the isotopic composition of plant waxes. A. thaliana is a small flowering plant in the mustard family that has served as a model organism for research in plant biology for over seventy years (Meyerowitz, 1987). Due to its small genome size and a short generation time, it is the most widely-studied plant today with a rich database of genetic and molecular biology data. There are many wildtypes of this plant around the world and now numerous mutants for an array of experiments including the study of plant development and insertional mutagenesis (Coen and Meyerowitz, 1991; Alonso et al., 2003). Recently, A. thaliana stomatal density was varied by the expression of a positive regulator of stomatal development in leaves called STOMAGEN (AT4gl2970) (Sugano et al., 2010). The STOMAGEN overexpressed mutant (ST-OX) showed high density of stomata and the amiRNA suppressed mutant (ST-RNAi) showed low density and this provides us with 3 lineages, including the wildtype (WT) with which to compare the isotopic effects of stomatal densities. Experiments with these lineages of A. thaliana have shown that stomatal densities influence both carbon and water fluxes (Tanaka et al., 2013; Larcher et al., 2014). Higher stomatal density increases the water flux out of the leaf (i.e., transpiration rates) but shortens the effective path length between xylem and stomatal cavity making counteracting impacts on the 18 O enrichment of bulk leaf water (Larcher et al., 2014). Likewise these evaporative processes are expected to influence D-enrichment of bulk leaf water with a mass dependent isotopic effect converting to ~5‰ D for every 1‰ 18 O with the slope depending on relative humidity (Dansgaard, 1964), which 16 we predict will also be encoded in D values of plant waxes. More stomata also increase the carbon dioxide flux into the leaf, increasing the stomatal conductance in ST-OX (Tanaka et al., 2013). Based on theory (Farquhar et al., 1989), we predict higher stomatal density will increase the internal CO 2 partial pressure within the leaf which will allow greater 13 C discrimination and thus lower  13 C values recorded in plant waxes. Given the predicted effects on hydrogen and carbon isotopic fractionation, between the atmosphere and the plant, we seek to test whether variations in stomatal densities in A. thaliana produce measurable isotopic effects in the hydrocarbons present in cuticular plant waxes. Plant wax n-alkanes and n-alkanoic acids are the target compounds because they are preserved in the geological record, unlike most other plant biochemicals. As far as we are aware, this is the first study of the isotopic composition of plant wax compounds produced by A. thaliana. Extending beyond the leaf water enrichment and oxygen isotope studies of plant water by Larcher et al., (2014) we add studies of hydrogen isotopes in plant waters, and calculate the fractionations of plant wax n-alkanes relative to source waters. Building upon the insights into varied stomatal densities and carbon assimilation rates by Tanaka et al., (2013) we analyze whether there are measurable carbon isotope effects on both plant wax n-alkanes and bulk leaf tissue. Since STOMAGEN may also influence plant wax concentrations (Bourdenx et al., 2011), we test the molecular abundance distributions of the n-alkanes and n-alkanoic acids. Finally, we relate these experimental results to geological observations of variations in stomatal density and plant wax isotopic records from sediment cores to assess the implications for these paleoenvironmental proxies. The ultimate motivating research question for this experimental work is as follows: given variations in plant stomatal densities across the Cenozoic, might plant wax records of past 17 climate and vegetation require corrections for the isotope effects induced by varying stomatal densities? 2.2 Methods 2.2.1 Variant selection Arabidopsis thaliana (L.) Heynh, Columbia-0 (CS60000) was used as the wildtype (WT) and representative lines of ST-OX and with high stomatal density and ST-RNAi with low stomatal density were selected (Figure 2.1). These STOMAGEN mutated A. thaliana plants were cultivated by Sugano et al., and were grown from seed at the Laboratory of Stable Isotope Ecology in Tropical Ecosystems at the at the University of Miami. Fig. 2.1. Scanning Electron Microscope image of leaf abaxial surface for A. thaliana variants with suppressed (ST- RNAi), wildtype (WT), and overexpressed (ST-OX) stomatal density from left to right. 2.2.2 Growth conditions The plants were hydroponically grown under laboratory conditions at room temperature, with WT and transgenic lines randomly distributed (Larcher et al., 2014). This implies that the plants 18 received above normal pCO 2 relative to the free atmosphere and the  13 C values are expected to be variable with the occupation of the building and ventilation than would be the case for plants in natural environments. Plants were grown in identical environment relative to one another. Plants were irrigated with Miami tap water of known isotopic composition, reported as irrigation water. Five plants were combined for one leaf water measurement by (Larcher et al., 2014), with five sample replicates per line. The plants developed their flower stalk approximately 2 months after planting with the suppressed line taking a couple more days relative to the other lines. All plants were harvested on the same day after flower stalk development. The plant sizes between lineages were indistinguishable except the overexpressed line, ST-OX, which attained slightly smaller final size and mass. The samples also had similar leaf to stem ratio, but measurements of plant dimensions and masses were not made. In Tanaka et al., (2013), where they grew the same plants over 6 weeks, the overexpressed line also had the smallest biomass with wildtype being the largest. They also measured flattened leaf projection areas per plant and wildtype had the biggest leaf area compared to the mutant lines. 2.2.3 Plant water extractions and isotopic measurement As reported by Larcher et al., (2014), plant waters were extracted cryogenically and stem and leaf waters were analyzed for 18 O/ 16 O and D/H isotopic composition by methods as reported by Vendramini and Sternberg, (2007). Extracted water was analyzed at the Laboratory of Stable Isotope Ecology in Tropical Ecosystems at the University of Miami for hydrogen isotope ratios by equilibration with an Isoprime isotope ratio mass spectrometry via a Multiflow system (Elementar, Germany). Water isotope ratios are reported relative to Vienna Standard Mean Ocean Water (VSMOW) and the precision of analysis is ±2‰ (1σ). Five A. thaliana plants were combined for one leaf water sample, and a total of five replicate leaf water samples were 19 analyzed per Arabidopsis transgenic line.  18 O measurements of leaf water were reported by Larcher et al., (2014). 2.2.4 Lipid extractions We studied the lipid extracts of 3 samples: a wildtype and two transgenic lines. Each sample comprises 25 A. thaliana plants (pooling the 5 samples of 5 plants each used for water isotopic analyses) to obtain sufficient biomass for wax analysis. Plant waxes were extracted from the dried, chopped, entire plant (leaf and stem) using 9:1 dichloromethane (DCM)/methanol (MeOH) to extract the epicuticular and intracuticular waxes. The extract was separated using column chromatography (5 cm ⅹ 40mm Pasteur pipette, NH 2 sepra bulk packing, 60 Å ), eluting with 2:1 DCM/isopropanol, followed by 4% formic acid in diethyl ether, yielding neutral and acid fractions respectively. The acid fraction was methylated to provide the corresponding fatty acid methyl esters (FAMEs) with 5% HCl and 95% MeOH (of known isotopic composition) at 70°C for 12 hours. Excess milliQ water was added to the hydrolyzed products and the lipids were partitioned into hexane, and dried by passage through a column of anhydrous Na2SO4. They were purified using column chromatography (5 cm ⅹ40mm Pasteur pipette, 5% water- deactivated silica gel, 100-200 mesh), eluting with hexane, followed by FAMEs eluted with DCM. 2.2.5 Compound-specific analytical methods 2.2.5.1 Compound specific molecular abundance distributions The n-alkane fraction was identified and quantified using gas chromatography coupled with both mass-selective detector and flame ionization detection (GC-MSD/FID Agilent). The GC-MS/FID 20 was equipped with a Rxi-5 ms column (30 m ⅹ 0.25 mm, film thickness 0.25 m) with a 1 mL min−1 constant column flow. Helium was used as the carrier gas and a 26 min GC method was used. Initial temperature of 50°C was held at 3.5 min, followed by a temperature ramp of 20°C min−1 until 300°C for remaining 10 min. 1 L of sample was injected via a split/splitless inlet (S/SL). Absolute abundances for individual chain lengths of n-alkanes and n-alkanoic acids were calculated relative to a standard mixture of n-alkane and n-alkanoic acids with individual calibration curves for each compound. Results are reported in microgram per gram dry leaf tissue (g g −1 ). 2.2.5.2 Compound specific hydrogen and carbon isotopic analysis Compound specific hydrogen and carbon isotopic values were obtained using gas chromatography isotope ratio mass spectrometry (GC-IRMS) at University of Southern California. We used a Thermo Scientific® Trace gas chromatrograph equipped with a Rxi-5 ms column (30 m ⅹ 0.25 mm, film thickness 0.25 m) and a programmable temperature vaporizing (PTV) injector, connected via a GC Isolink with pyrolysis furnace at 1400 °C for D/H analysis and via a combustion furnace at 1000 °C for  13 C via a Conflo IV interface to a DeltaVPlus isotope ratio mass spectrometer. We operated the PTV injector in solvent-split mode, with the evaporation temperature at 60°C to exclude the solvent. The GC was operated with a flow of 1mL min−1 and an initial temperature of 50°C with a ramp of 20°C min−1 to 250°C and then 3.5°C min−1 to 310°C. For hydrogen, H3 factor was measured to check for linearity in isotopic determination across a range of peak amplitude (1-8 V), the mean value for the measurements here was 5.4 ppm mV−1. For carbon, standard deviation of linearity peaks across 1-8 V was also determined to be 0.07. 21 Reference peaks of H¬2 or CO2 as appropriate were co-injected bracketing n-alkane peaks during the course of a GC-IRMS run. Two of these peaks were used for standardization of the isotopic analysis, while the rest were treated as unknowns to assess precision. Samples were bracketed by runs of a standard containing a mixture of compounds of known isotopic composition. Data were normalized to the Vienna Standard Mean Ocean Water (VSMOW)/Standard Light Antarctic Precipitation (SLAP) hydrogen isotopic scale and to the Vienna PeeDee Belemnite (VPDB) for carbon via comparison with an external standard with 15 n-alkanes (Amix isotopic standard from A. Schimmelmann, University of Indiana) with D values ranging from −9 to −254‰ and  13 C values ranging from −33.3 to −28.6‰. The RMS error (a combined measure of accuracy and precision) of replicate analysis of the external standard was in average 4‰ for hydrogen and 0.2‰ for carbon. The results are reported using conventional delta notation (‰). Here we report isotopic fractionations between two measured variables D a and D b , as enrichment factors ( a/b ) defined as: ⁄ ⁄ 1 (4) following (Sessions and Hayes, 2005). Enrichment factors are reported in permil notation which implies a factor of 1000 (Coplen, 2011). From the measured values of irrigation water (D w ), bulk leaf water (D lw ) and leaf wax (D wax ), we calculated fractionations for leaf water enrichment ( 2  lw/w ), biosynthetic fractionation ( 2  wax/lw ), and net fractionation ( 2  wax/w ) including compound uncertainties (Table 2.1). The empirically determined fractionations reported in this paper are calculated as follows: 22 leaf water enrichment: ⁄ (5) net or apparent fractionation: ⁄ (6) and biosynthetic fractionation: ⁄ (7) 2.2.6 Bulk carbon isotopic analysis Samples of ground leaf tissues were prepared for bulk  13 C values were measured using elemental analyzer (EA) coupled to a Picarro. The  13 C values were evaluated by comparison to 2 isotopically known standards: bulk CaCO 3 (−37.84‰, 12%C) and USGS40 (−26.39‰, 40.8%C). Analytical precision was 0.04‰ for bulk CaCO 3 (n = 8) and 0.03‰ for USGS40 (n = 3). We calculated the isotopic offset between bulk tissue and leaf wax ( 13  wax/bulk ) as: ⁄ (8) 2.3 Results 2.3.1 Plant waters We found leaf water D values of +17 to +20‰, (Table 2.1). We predicted that leaf water would be more D-enriched with increasing numbers of stomata and while that was observed in the mean values, the difference of 3‰ between lineages was smaller than the analytical uncertainties on water isotopic measurements (2-5‰), thus the difference is statistically insignificant. While xylem water normally reflects source water (Lin and Sternberg, 1994; Ellsworth and Williams, 2007), here the measured stem water values are slightly enriched above irrigation 23 water D values. This is not a surprise as the stems of A. thaliana are small (< 20 cm), green and not suberized (not bark covered), which means that the xylem water is subject to evaporative fractionation and falls between source water and leaf water in isotopic composition. We thus focus on comparisons between leaf water and irrigation water when assessing the enrichment of leaf water above source water. 2.3.2 Molecular abundance distribution We found that the A. thaliana lineages make n-alkanes from C 16 to C 33 chain length and n- alkanoic acids from 16 to 34 carbon chain length. The odd over even carbon preference index (CPI) for n-alkanes between 23 and 33 long chain length ranged from 6 to 9 (Table 2.2). For n- alkanoic acids the even over odd CPI from long chain length 24 to 34 ranged from 5 to 6 (Table 2.2). The change in stomatal density did not have a significant effect on the average chain length (ACL = Σ(C n •n)/ ΣC n , where n is from 23 to 33 for n-alkanes and 24 to 34 for n-alkanoic acids). The ACL was approximately 30 for both n-alkanes and n-alkanoic acids (Table 2.2). The mutant lines had higher wax concentrations per gram of biomass compared to the wildtype. For n- alkanes, the overexpressed line had the highest concentrations while the suppressed line was the highest in n-alkanoic acids. The molecular distribution of long chain n-alkanes show that nC 31 is the most abundant followed by nC 29 (Figure 2.2) in all the lineages, this allows their isotopic results to be compared without concern for bias in relative abundance. We were only able to determine the isotopic compositions of the n-alkanes because of their abundance. n-Alkanoic acids were at lower abundance and would require >0.6 g of dry leaves for isotopic determination. 24 Fig. 2.2. Abundances of a) n-alkanes and b) n-alkanoic acids by chain length for A. thaliana variants. Note y-axis scale change in b). 2.3.3 Hydrogen isotopic compositions We found D-enrichment in the plant wax n-alkanes with increasing stomatal density (Figure 2.3). D values of the two most abundant chain lengths are reported here (Table 2.1). From suppressed (ST-RNAi) to overexpressed (ST-OX) mutants, the difference in D values was +13‰ which is far larger than analytical uncertainties on individual measurements (1 ~3‰). For biosynthetic and net fractionation enrichment calculations, we used C 31 n-alkane results for D wax values. The uncertainties consist of plant water replicate samples (5) and analytical uncertainties and for D wax analytical uncertainties only. We found leaf water enrichment (2 lw/w ) of +16 to +18‰ to be statistically indistinguishable between wildtype and variants within uncertainties (<5‰). Biosynthetic fractionations ( 2  wax/lw ) average −159 ± 5‰ (1). The suppressed line had a more negative biosynthetic fractionation value of −165‰ compared to −155‰ of the overexpressed line. Net fractionation ( 2  wax/w ) spanned from −152 to −140‰ for the low to high density transgenic lines with cumulative uncertainties of 6 to 2‰ (1), including 25 those from water and wax measurements. The difference between ST-RNAi and ST-OX in the net and biosynthetic fractionations was larger than that of leaf water enrichment, i.e. measured leaf water was insufficient to explain the observed net fractionation (Figure 2.4, Table 2.1). Fig. 2.3. D of leaf water and wax for A. thaliana variants with differing stomatal density. For plant water isotopic measurements mean and 1 standard deviation (error bars) refer to replicate samples for plant water isotopic measurements, including growth and instrument uncertainties. For plant wax isotopic composition mean and 1 standard deviation (error bars) refer to replicate GC-IRMS measurements of the same pooled sample. Fig. 2.4. Leaf water enrichment, net fractionation and biosynthetic fractionations in A. thaliana variants with differing stomatal density. Fractionations shown as mean and 1 compound standard deviation (error bars). 26 2.3.4 Carbon isotopic compositions We report the carbon isotopic composition of the n-alkanes and bulk leaf tissue (Table 2.1, Figure 2.5). We note the carbon isotope results are 13C-depleted relative to most environmental plant samples, reflecting growth conditions in a laboratory with likely elevated pCO 2 and 13 C- depleted, respired carbon dioxide, but the composition of laboratory atmosphere was not monitored. High stomatal density resulted in more depleted  13 C values in bulk leaf tissue and plant wax n-alkanes as predicted (except in the chain length nC 29 n-alkane). We calculated the fractionation between bulk tissue and plant wax ( 13  wax/bulk ), taking ‗wax‘ as nC 31 , the most abundant n-alkane, yielding a mean of +8.4 ± 0.1‰, that was consistent across all variants. Fig. 2.5.  13 C values of bulk and plant wax n-alkanes from A. thaliana variants with differing stomatal density. 27 Table 2.1. Measured hydrogen and carbon isotopic compositions and calculated fractionations. SD (mm -2 ) SI (%) Hydrogen Isotopic Results (‰) xw 8 ±4 3 ±4 5 ±2 lw 17 ±3 18 ±2 20 ±5 C 29 -149 ±4 -147 ±4 -132 ±1 C 31 -151 ±6 -143 ±2 -138 ±2 2  lw/w 16 ±3 17 ±2 18 ±5 2  wax/w -152 ±6 -144 ±2 -140 ±2 2  wax/lw -165 ±7 -158 ±3 -155 ±6 Carbon Isotopic Results (‰) C 29 -41.2 ±0.1 -41.6 ±0.1 -40.7 ±0.1 C 31 -41.1 ±0.1 -41.7 ±0.0 -42.5 ±0.0 C 33 -40.5 ±0.1 -42.1 ±0.0 -43.5 ±0.4 bulk -32.9 ±0.0 -33.5 ±0.0 -34.5 ±0.0 13  wax/bulk 8.5 ±0.1 8.4 ±0.3 8.4 ±0.0 SD stomatal density, SI stomatal index, ED epidermal density. SI (%) = [SD/(SD + ED)] × 100%. xw xylem water, lw leaf water, uncertainty is 1 of 5 replicate samples (5 plants each), including growth and instrument uncertainties. w irrigation water, 1.4‰ with 0.5‰ 1 analytical uncertainty. wax refers to C 31 n -alkane. Leaf wax 1 uncertainty, refers to instrument uncertainty only on triplicate measurements of pooled sample (25 plants). Fractionation uncertainty is the compound 1 uncertainty. WT wildtype ST-OX overexpressed ST-RNAi suppressed 43 90 297 50.9 32.9 17.4 28 Table 2.2. Plant wax n-alkane and n-alkanoic acid molecular abundance distribution. ST-RNAi WT ST-OX suppressed wildtype overexpressed SD (mm -2 ) 43 90 297 SI (%) 17.4 32.9 50.9 n -Alkane concentrations (g g -1 ) C 23 0.86 0.61 0.65 C 24 1.81 1.38 1.59 C 25 1.88 1.57 2.54 C 26 1.59 1.45 3.35 C 27 2.79 2.29 4.20 C 28 1.11 1.04 2.53 C 29 25.69 18.85 23.91 C 30 1.66 1.28 2.38 C 31 29.47 21.98 37.80 C 32 1.00 0.89 1.61 C 33 7.07 5.44 11.15 C 34 nd nd nd total 75 57 92 CPI 8.9 7.9 6.5 ACL 29.8 29.8 29.9 n -Alkanoic acid concentrations (g g -1 ) C 23 nd nd nd C 24 0.12 0.04 0.10 C 25 0.02 0.01 0.02 C 26 0.18 0.08 0.16 C 27 0.03 0.01 0.03 C 28 0.19 0.09 0.16 C 29 0.03 0.02 0.03 C 30 0.15 0.09 0.15 C 31 0.04 0.03 0.04 C 32 0.29 0.20 0.20 C 33 0.05 0.04 0.04 C 34 0.26 0.21 0.15 total 1.35 0.84 1.09 CPI 6.3 4.9 5.0 ACL 29.8 30.6 29.4 SD and SI as for Table 1. Only data for >C 23 reported, C 16 and C 18 n -alkanoic acids are highly abundant components of cells but are not reported here. CPI, Carbon Preference Index. For n -alkanes, CPI is calculated as (0.5[C23]+[C25]+[C27]+[C29]+[C31]+0.5[C33]) /([C24]+[C26]+[C28]+[C30]+ [C32]). For n-alkanoic acids, CPI is calculated as (0.5[C24]+[C26]+[C28]+[C30]+[C32] ACL, Average Chain Length. nd, not determined. 29 2.4 Discussions 2.4.1 Implications of STOMAGEN mutations on leaf water enrichment We found no significant correlation between increased stomatal densities and bulk leaf water hydrogen isotopic composition (Figure 2.3). In leaf water, two of the variables that define the Peclét ratio, which predicts enrichment, changed significantly with varying stomatal density. First, the effective path length L decreased with more stomata due to an increase in total cross- sectional area and an increase in the probability of stomatal pore being closer to the vein (Larcher et al., 2014). This decrease in L predicts enriched leaf water. On the contrary, the increase in transpiration rate in the same overexpressed mutant, decreased bulk leaf water enrichment by constant loss of water and replacement of leaf water by depleted stem water (Sternberg, 2009). These two competing factors, explain the observation that bulk leaf water composition recorded no observable signal in the A. thaliana samples for oxygen isotopic composition (Larcher et al., 2014) and here for hydrogen isotopic composition. However, isotopic enrichment is expected at the site of evaporation, close to the stomata (Barbour et al., 2003) and with more stomata, the ST-OX line will have greater enrichment cumulatively in the leaf tissue (lamina). This enriched lamina water cannot be directly measured because the leaves of A. thaliana were so small we could not separate the veins from the lamina. Because waxes are synthesized in the lamina, lamina water may be more relevant as a substrate than measured bulk leaf water. If lamina waters are enriched above bulk leaf water then this may explain the observation that measured wax isotopic compositions follow the prediction for more enrichment associated with greater stomatal densities; alternatively, the difference must be driven by isotope effects associated with biosynthesis (discussed in Section 4.2). Unfortunately, we cannot further distinguish between these two possibilities (lamina water enrichments above bulk leaf water or 30 biosynthetic differences), as we cannot empirically determine the isotopic composition of substrate waters used in the synthesis of waxes. 2.4.2 Implications of STOMAGEN for plant wax biosynthesis and isotopic composition Despite the lack of significant enrichment in bulk leaf water D values, we found a significant D-enrichment of plant wax as predicted with increasing stomatal densities being more prone to evaporation (Table 2.1). This means that either A) lamina water enrichment, exclusive of the vein water, is sufficient to drive the wax enrichment (despite little change in bulk leaf water), and/or B) that the enrichment derives from the fractionation between water and wax changing, perhaps the role of STOMAGEN on biosynthesis. The first (A), is quite likely as waxes would have been synthesized from limited water ‗pools‘ within the leaf at an earlier time in leaf ontogeny than the time at which bulk leaf water was sampled. However, the spatial and temporal variations in water in the lamina close to sites of evaporation cannot be sampled, and the water relevant to wax synthesis carries many unknowns. Thus, we cannot elucidate whether ‗local‘ variations in lamina water would explain the D wax results (as discussed in Section 4.1). The latter option, biosynthesis (B), is also a viable explanation as genetic changes often incur more than one effect. In addition to stomatal density variations, we found higher wax loading in both ST-RNAi (75 g g −1 ) and ST-OX mutants (92 g g −1 ), relative to wildtype (57 g g −1 ) reflecting an increase in wax loading with STOMAGEN mutation (Table 2.2). We also observe an increase in nC 31 /nC 29 ratio in the overexpressed mutant, perhaps indicating different elongation and/or decarboxylation during lipid synthesis with extreme numbers of stomata, whereas the ST-RNAi mutant displayed a similar nC 31 /nC 29 ratio to that in the wildtype. Why STOMAGEN mutation leads to a shift in wax concentrations is unknown, but links between genes that control stomatal characteristics and wax loading have been previously examined. While there are no prior studies 31 of the effects of STOMAGEN on plant wax biosynthesis and isotopic fractionation, others have reported that plants with lesions in wax biosynthesis genes show abnormal stomatal densities compared to wildtype plants (Gray et al., 2000). Moreover, in a study varying the A. thaliana cer3 gene, which controls the cuticle membrane and cuticular waxes, a cascade of changes on stomatal density and trichome development was reported (Bourdenx et al., 2011). Since genes that control stomatal densities can also influence wax concentrations (Gray et al., 2000; Bourdenx et al., 2011), then it is likely that this shift in wax production will be accompanied by a shift in the isotopic fractionation associated with biosynthesis ( 2  wax/lw ). In our study, both suppressed and overexpressed mutants showed higher wax concentrations relative to wildtype but biosynthetic isotopic effects were in opposing directions from that of wildtype, thus biosynthetic wax yield and isotopic fractionation are not correlated. Although we cannot quantify the isotope effects involved in fatty acid elongation of STOMAGEN mutants, it is known that the hydrogen isotopic composition of NADPH significantly contributes to that of fatty acids, and then of alkanes (Schmidt et al., 2003; Zhang et al., 2009). Overexpression entails faster photosynthetic rates compared to suppressed types (Tanaka et al., 2013), and this may lead to kinetic isotopic effects as well as increased use of fresh photosynthate versus hydrogen supplied by NADPH, but we speculate this effect would entail a D-depletion in the faster photosynthesizers, which is opposite to the observed trend in 2  wax/lw leaving us without a mechanism to explain how (B) biosynthesis may drive D-enrichment in the overexpressed mutant. At present, although the D wax variations appear to derive from the biosynthesis step, we do not find support from theory for the direction of the observed shift. Based on changing 2  wax/lw , our review of mechanistic theory instead suggests instead that measured bulk leaf water did not adequately capture the substrate used in biosynthesis, leading us to question the explanatory 32 power of D lw and 2  wax/lw measured here. If D lw is not truly reflective of the water used in wax synthesis, then 2  wax/lw is also affected by the same concern. Further isotopic investigation of biosynthesis with STOMAGEN mutants might be able to elucidate isotope effects associated with changing stomatal densities, if they can better constrain leaf water isotopic compositions. Experimentation with different environmental conditions, biosynthesis pathways and compounds may also help elucidate biosynthetic hydrogen isotopic effects. 2.4.3 Carbon isotopic composition Stomata are known to exert controls on the  13 C compositions of plant tissues (Park and Epstein, 1960; Farquhar et al., 1989). Here we confirm that stomatal density alone influences the  13 C value of bulk and plant wax n-alkanes. We find the expected pattern of 13 C-depletion with higher stomatal density in both bulk and n-alkanes (Figure 2.5). We note that the isotopic result for the C 29 n-alkane appears does not follow that of nC 31 or nC 33 .While we do not have a clear explanation for this aberrant isotopic result for nC 29 , it may be related to the decrease in nC 29 relative to nC 31 , therefore we consider it an outlier and do not discuss further here. A previous study of A. thaliana STOMAGEN mutants found that increased stomatal density results in higher stomatal conductance (Tanaka et al., 2013). Higher stomatal conductance also implies increased rates of diffusion of CO 2 into the leaf, which would increase the partial pressure of CO 2 inside the stomatal cavity (C i ). Higher C i in turn would support faster photosynthetic rates, which may increase kinetic fractionations associated with biosynthesis, such as that involving a partitioning of pyruvate to acetyl-CoA (DeNiro and Epstein, 1977) resulting in 13 C-depletion of lipid products. Faster photosynthetic rates remove CO 2 from inside the stomatal cavity, lowering C i , therefore faster photosynthetic rates are only supported in as 33 much as diffusion can maintain the high C i . If high diffusion rates can maintain fast photosynthesis then 13 C-discrimination should increase because of kinetic effects. We found that the overexpressed mutant had a more 13 C-depleted bulk and plant wax composition than suppressed mutants in these experiments. Tanaka et al., (2013) report a doubling of stomatal conductance and a 30% increase in photosynthesis rate in ST-OX over the suppressed mutant (Tanaka et al., 2013). Together, the results of these two studies demonstrate the mechanistic steps and isotopic evidence supporting theory: elevated stomatal densities lead to an increase in Ci and kinetic fractionations associated with biosynthesis in turn lead to 13C depletion in plant bulk and wax isotopic composition. The fractionation between bulk tissue and plant wax n-alkanes 13  wax/bulk averaged −8.4‰ consistent with reports for other species reported elsewhere (Conte et al., 2003). Plant wax n- alkane is a small part of the bulk leaf, in this case less than 100 g g−1 of biomass (0.01%; Table 2.2). The remaining biochemicals in the leaf including cellulose, lignin and sugars can have widely varying isotopic compositions with unique biosynthetic fractionations that have not been measured here, other than as summarized by bulk leaf isotopic composition. Since the 13  wax/bulk did not vary across the lineages studied, we can conclude that the effects of STOMAGEN and stomatal densities did not extend to varying the carbon isotopic offset between the wax compounds and the bulk tissue, and that either bulk tissue or plant wax molecules can equally resolve the carbon isotopic discriminatory effect associated with varying stomatal densities. This also suggests that stomatal density will have common effects on plant wax and bulk plant carbon isotopic compositions in geological archives. 34 2.4.4 Implications for paleoenvironmental reconstructions using the plant wax isotopic compositions This large-scale study of carbon isotope systematics in tropical lowland and montane rainforests Our study has found that the C and H isotopic composition of plant waxes vary with stomatal expression in 3 lineages of A. thaliana in uniform laboratory growth conditions. The question is then whether paleoenvironmental reconstructions using plant wax C or H isotopic compositions need to account for variations in stomatal densities between species or between climate states with varying pCO 2 , where such information is available from fossil leaves. The A. thaliana experiments isolate the effects of SD that may be generalizable to other species. However, fossil leaves used for pCO 2 reconstructions are usually from higher vascular trees such as Ginkos or Metasequoia (both gymnosperms), whose leaf physiology is very different from the small angiosperm A. thaliana. In particular, differences between needles and broadleaf species, venation and connectivity between vein and lamina may be important for leaf water flow and enrichment (Cernusak et al., 2015). We note that there are two common ways to describe the amount of stomata on extant and fossil plants. One is stomatal density (SD) which is the number of stomata in a given area (mm−2) and the SD on a mature leaf is relevant to plant isotope effects as discussed previously. SD on the mature leaf is affected by both the initiation of stomata and the expansion of epidermal cells which is a function of many factors such as light, temperature, water status, position of leaf crown and intra-leaf position during leaf development (Royer, 2001). In reconstructions of pCO 2 from fossil leaves, measured stomatal densities are typically normalized for epidermal cell 35 densities (ED mm −2 ) to yield a measure of stomatal initiation which best reflects pCO 2 in early leaf ontogeny (Royer et al., 2001). That measure is called the stomatal index (SI), where (9) We compare results from A. thaliana in this study, to fossil leaf data from the literature, in order to estimate the possible isotope effects from varying stomatal densities in ancient ecosystems. We report both SD and SI results to facilitate comparisons across the literature with SI being more relevant to pCO 2 reconstructions and SD being the variable of interest for isotopic fractionations. In our study SD of A. thaliana ranged from 43 to 297 mm −2 (Table 2.1), this compares to part of the range observed in extant plants up to 900 mm −2 in current pCO 2 conditions in the Cenozoic (Figure 2.6) (Franks and Beerling, 2009). Variations in SI between lineages tested here range from 17.4 to 50.9% (Table 2.1; as originally reported by Sugano et al; 2010) which is much higher than those reported in the fossil record, e.g., 6 to 19%, for Ginkgo, Metasequoia and laurel species (L. pseudoprinceps, L. nobilis, O. foeten) (Royer et al., 2001; Kuerschner et al., 2008; Beerling et al., 2009) across pCO 2 in the geological time (Figure 2.7). This reflects the overexpression intentionally engineered in these organisms, however once the leaf reaches full size SD is the more appropriate measure of comparison of stomatal coverage and the mutants fall within natural range of variability. Thus our experiments with various SD provide a guide to the isotopic fractionations due to varying SD in extant and fossil plants. If the results from A. thaliana are transferrable, across the geological record, SD variations of 250 ± 10 mm −2 may be expected to similarly yield +15 ± 3‰ in D and −3 ± 0.4‰ in  13 C, all other factors being equal. However, it must be remembered that geological variations in SD are adaptive responses to major environmental changes that more directly influence plant wax 36 isotopes. For example, at times of low atmospheric pCO 2 (low C a ), we would expect lower C i and an enrichment of leaf tissue  13 C. Plant adaptive increases in stomatal density might cancel out the direct effects of a decrease in C a , allowing the plant to maintain C i . Thus increasing SD would tend to counteract the effects of lowered C a , perhaps leading to a null net isotope effect if stomatal density adaptive responses keep pace with atmospheric chemistry changes, although additional considerations include photosynthetic rate which might slow at cooler temperatures. For hydrogen isotopes, the D enrichment predicted for plant waters with higher SD could be offset at times of low atmospheric pCO 2 and cooler climates by the effects of lower temperatures on transpiration rates and fractionations. We note that our growth experiment only varies SD genetically and not environmental conditions, thus a valuable follow-up experiment could take the same lineages and test their isotopic responses under varying pCO 2 , temperature and other environmental factors to make our results more comparable to past climates. Additional experiments might extend these approaches to consider isotope effects in plants with other leaf morphologies especially those represented in the fossil record. Fig. 2.6. Comparison of stomatal densities of A. thaliana lineages (black circles) in this study (data from Sugano et al., 2010) with those of Cenozoic fossil leaves (data compiled by Franks and Beerling, 2009). 37 Fig. 2.7. Comparison of stomatal indices of A. thaliana lineages used in this study (data from Sugano et al., 2010) with those of Cenozoic fossil leaves of Ginko biloba, Metasequoia glyptostroboides (Beerling et al., 2009) and of Laurus nobilis, Ocotea foetens, Ginkgo biloba, Laurophyllum pseudoprinceps (Kuerschner et al., 2008) with respect to pCO 2 . 2.5 Conclusions We experimentally determined the isotope effects of stomatal density on plant wax carbon and hydrogen isotopic composition through controlled growth experiments with wildtype and genetic mutants of A. thaliana with three different stomatal densities. We found that the overexpressed mutant had offsets of +15‰ and −3‰ for D and  13 C respectively, compared to the suppressed mutant. We found the plant wax D-enrichment was larger than that which could be explained by measured bulk leaf water values. We infer that measured bulk leaf water may not represent the composition of water used for wax synthesis. Alternatively, isotope effects involved in biosynthesis of n-alkanes in STOMAGEN mutants may add to the D-enrichment of plant wax, for example via demonstrated additional controls on wax synthesis, including greater waxiness. While the mechanism remains unclear, the direction of hydrogen isotopic effects in plant wax with stomatal density matches predictions. For carbon isotopes, overexpressed mutants had more 38 depleted  13 C values, representing greater discrimination against the heavier isotope given the more open exchange of CO 2 with higher stomatal densities, matching predictions from theory. These isotopic shifts with stomatal density changes could have implications for paleoclimate studies, as stomatal densities are well known to change as a result of pCO 2 variations across the geologic timescale. However, climatic factors associated with pCO 2 , such as temperature and relative humidity, would also directly affect stomatal conductance and transpiration, thus plant isotopic compositions. We anticipate that these additional climatic variables will modulate the isotopic effects from stomatal densities alone. Additional experiments with these lineages under varying environmental conditions would help quantify those trade-offs further. Acknowledgements This work was supported by the USC Provost‘s Fellowship and the USC Department of Earth Sciences‘ graduate student research fund to HL and from USC to SF. We thank Ikuko Hara- Nishimura and his collaborators for providing STOMAGEN mutant A. thaliana seeds. We also thank Nick Rollins and Camilo Ponton for laboratory assistance at USC and L. Larcher for sharing the plant samples from her doctoral work funded by a fellowship from CAPES-PDSE at U. Miami. SF wishes to thanks to Kate Freeman, Francesca McInerney, David Bottjer, Alex Sessions and Elliot Meyerowitz for helpful discussions over the years. This manuscript was improved with the comments of Associate Editor Alex Sessions, Dirk Sachse and two anonymous reviewers. 39 References Alonso J. M., Stepanova A. N., Leisse T. J., Kim C. J., Chen H. M., Shinn P., Stevenson D. K., Zimmerman J., Barajas P., Cheuk R., Gadrinab C., Heller C., Jeske A., Koesema E., Meyers C. C., Parker H., Prednis L., Ansari Y., Choy N., Deen H., Geralt M., Hazari N., Hom E., Karnes M., Mulholland C., Ndubaku R., Schmidt I., Guzman P., Aguilar-Henonin L., Schmid M., Weigel D., Carter D. E., Marchand T., Risseeuw E., Brogden D., Zeko A., Crosby W. L., Berry C. C. and Ecker J. R. (2003) Genome-wide Insertional mutagenesis of Arabidopsis thaliana. Science 301, 653–657. Barbour M. M. (2007) Stable oxygen isotope composition of plant tissue: a review. Funct. Plant Biol. 34, 83–94. Barbour M. M., Roden J.S., Farquhar G. D. and Ehleringer J. R. (2003) Expressing leaf water and cellulose oxygen isotope ratios as enrichment above source water reveals evidence of a Péclet effect. Oecologia 138, 426–435. Beerling D., Chaloner W., Huntley B., Pearson J. and Tooley M. (1993) Stomatal Density Responds to the Glacial Cycle of Environmental-Change. Proc. R. Soc. B-Biol. Sci. 251, 133– 138. Beerling D.J., Fox A. and Anderson C. W. (2009) Quantitative Uncertainty Analyses of Ancient Atmospheric Co2 Estimates from Fossil Leaves. Am. J. Sci. 309, 775–787. Beerling D.J., McElwain J. C. and Osborne C. P. (1998) Stomatal responses of the ―living fossil‖ Ginkgo biloba L. to changes in atmospheric CO2 concentrations. J. Exp. Bot. 49, 1603–1607. Beerling D.J. and Royer D. L. (2011) Convergent Cenozoic CO2 history. Nat. Geosci. 4, 418– 420. Bourdenx B., Bernard A., Domergue F., Pascal S., Leger A., Roby D., Pervent M., Vile D., Haslam R. P., Napier J. A., Lessire R. and Joubes J. (2011) Overexpression of Arabidopsis ECERIFERUM1 Promotes Wax Very-Long-Chain Alkane Biosynthesis and Influences Plant Response to Biotic and Abiotic Stresses. Plant Physiol. 156, 29–45. Cernusak L. A., Barbour M. M., Arndt S. K., Cheesman A. W., English N. B., Feild T. S., Helliker B. R., Holloway-Phillips M. M., Holtum J. A. M., Kahmen A., McInerney F. A., Munksgaard N. C., Simonin K. A., Song X., Stuart-Williams H., West J. B. and Farquhar G. D. (2015) Stable isotopes in leaf water of terrestrial plants. Plant, Cell & Environment, n/a–n/a. Cernusak L. A., Farquhar G. D. and Pate J. S. (2005) Environmental and physiological controls over oxygen and carbon isotope composition of Tasmanian blue gum, Eucalyptus globulus. Tree Physiol 25, 129–146. Coen E. and Meyerowitz E. (1991) The War of the Whorls - Genetic Interactions Controlling Flower Development. Nature 353, 31–37. 40 Conte M. H., Weber J. C., Carlson P. J. and Flanagan L. B. (2003) Molecular and carbon isotopic composition of leaf wax in vegetation and aerosols in a northern prairie ecosystem. Oecologia 135, 67–77. Coplen T. B. (2011) Guidelines and recommended terms for expression of stable-isotope-ratio and gas-ratio measurement results. Rapid Commun. Mass Spectrom. 25, 2538–2560. Dansgaard W. (1964) Stable Isotopes in Precipitation. Tellus 16, 436–468. DeNiro M. J. and Epstein S. (1977) Mechanism of carbon isotope fractionation associated with lipid synthesis. Science 197, 261–263. 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. Ehleringer J. R., Cerling T. E. and Helliker B. R. (1997) C-4 photosynthesis, atmospheric CO2 and climate. Oecologia 112, 285–299. Ellsworth P. Z. and Williams D. G. (2007) Hydrogen isotope fractionation during water uptake by woody xerophytes. Plant Soil 291, 93–107. Farquhar G. D. and Cernusak L. A. (2005) On the isotopic composition of leaf water in the non- steady state. Funct. Plant Biol. 32, 293–303. Farquhar G. D. and Lloyd J. (1993) Carbon and Oxygen Isotope Effects in the Exchange of Carbon Dioxide between Terrestrial Plants and the Atmosphere. Stable isotopes and plant carbon-water relations. Farquhar G., Ehleringer J. and Hubick K. (1989) Carbon Isotope Discrimination and Photosynthesis. Annu. Rev. Plant Physiol. Plant Molec. Biol. 40, 503–537. Farquhar G. and Richards R. (1984) Isotopic Composition of Plant Carbon Correlates with Water-Use Efficiency of Wheat Genotypes. Aust. J. Plant Physiol. 11, 539–552. Farquhar G. and Sharkey T. (1982) Stomatal Conductance and Photosynthesis. Annu. Rev. Plant Physiol. Plant Molec. Biol. 33, 317–345. Farquhar G. D., Cernusak L. A. and Barnes B. (2007) Heavy Water Fractionation during Transpiration. Plant Physiol. 143, 11–18. Feakins S. J., Warny S. and Lee J.-E. (2012) Hydrologic cycling over Antarctica during the middle Miocene warming. Nat. Geosci. 5, 557–560. Franks P. J. and Beerling D. J. (2009) Maximum leaf conductance driven by CO2 effects on stomatal size and density over geologic time. Proc. Natl. Acad. Sci. U. S. A. 106, 10343–10347. Gray J. E., Holroyd G. H., Van der Lee F. M., Bahrami A. R., Sijmons P. C., Woodward F. I., Schuch W. and Heterington A. M. (2000) The HIC signalling pathway links CO2 perception to stomatal development. Nature 408, 713–716. 41 Hetherington A. M. and Woodward F. I. (2003) The role of stomata in sensing and driving environmental change. Nature 424, 901–908. 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. Huang Y. S., Shuman B., Wang Y. and Webb T. (2004) Hydrogen isotope ratios of individual lipids in lake sediments as novel tracers of climatic and environmental change: a surface sediment test. J. Paleolimn. 31, 363–375. Huang Y. S., Shuman B., Wang Y. and Webb T. (2002) Hydrogen isotope ratios of palmitic acid in lacustrine sediments record late Quaternary climate variations. Geology 30, 1103–1106. Kahmen A., Sachse D., Arndt S. K., Tu K. P., Farrington H., Vitousek P. M. and Dawson T. E. (2011) Cellulose (delta)18O is an index of leaf-to-air vapor pressure difference (VPD) in tropical plants. Proc. Natl. Acad. Sci. U.S.A. 108, 1981–1986. Kahmen A., Schefuss E. and Sachse D. (2013) Leaf water deuterium enrichment shapes leaf wax n-alkane delta D values of angiosperm plants I: Experimental evidence and mechanistic insights. Geochim. Cosmochim. Acta 111, 39–49. Kuerschner W. M., Kvacek Z. and Dilcher D. L. (2008) The impact of Miocene atmospheric carbon dioxide fluctuations on climate and the evolution of terrestrial ecosystems. Proc. Natl. Acad. Sci. U. S. A. 105, 449–453. Larcher L., Hara-Nishimura I. and Sternberg L. (2014) Effects of stomatal density and leaf water content on the (18) O enrichment of leaf water. New Phytol. Lin G. and Sternberg L. (1994) Utilization of Surface-Water by Red Mangrove (rhizophora- Mangle L) - an Isotopic Study. Bull. Mar. Sci. 54, 94–102. Meyerowitz E. (1987) Arabidopsis-Thaliana. Annu. Rev. Genet. 21, 93–111. Pagani M., Pedentchouk N., Huber M., Sluijs A., Schouten S., Brinkhuis H., Damste J. S. S. and Dickens G. R. (2006) Arctic hydrology during global warming at the Palaeocene/Eocene thermal maximum. Nature 442, 671–675. Park R. and Epstein S. (1960) Carbon Isotope Fractionation During Photosynthesis. Geochim. Cosmochim. Acta 21, 110–126. Raven J. A. (2002) Selection pressures on stomatal evolution. New Phytol. 153, 371–386. Roden J. S. and Ehleringer James R. (1999) Observations of Hydrogen and Oxygen Isotopes in Leaf Water Confirm the Craig-Gordon Model under Wide-Ranging Environmental Conditions. Plant Physiol. 120, 1165–1174. Royer D. L. (2001) Stomatal density and stomatal index as indicators of paleoatmospheric CO2 concentration. Rev. Palaeobot. Palynology 114, 1–28. 42 Royer D. L., Wing S. L., Beerling D. J., Jolley D. W., Koch P. L., Hickey L. J. and Berner R. A. (2001) Paleobotanical evidence for near present-day levels of atmospheric CO2 during part of the tertiary. Science 292, 2310–2313. Rundgren M. and Bjorck S. (2003) Late-glacial and early Holocene variations in atmospheric CO2 concentration indicated by high-resolution stomatal index data. Earth Planet. Sci. Lett. 213, 191–204. 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 ed. R. Jeanloz. Annu. Rev. Earth Planet. Sci. 40, 221–249. Schefuss E., Schouten S. and Schneider R. R. (2005) Climatic controls on central African hydrology during the past 20,000 years. Nature 437, 1003–1006. Schmidt H. L., Werner R. A. and Eisenreich W. (2003) Systematics of 2 H patterns in natural compounds and its importance for the elucidation of biosynthetic pathways. Phytochemistry Reviews 2, 61–85. Sessions A. L. and Hayes J. M. (2005) Calculation of hydrogen isotopic fractionations in biogeochemical systems. Geochimica et Cosmochimica Acta 69, 593–597. Sternberg L. da S. L. O. (2009) Oxygen stable isotope ratios of tree-ring cellulose: the next phase of understanding. New Phytol. 181, 553–562. Sternberg L., Deniro M. and Ajie H. (1984) Stable Hydrogen Isotope Ratios of Saponifiable Lipids and Cellulose Nitrate from Cam, C-3 and C-4 Plants. Phytochemistry 23, 2475–2477. Sugano S. S., Shimada T., Imai Y., Okawa K., Tamai A., Mori M. and Hara-Nishimura I. (2010) Stomagen positively regulates stomatal density in Arabidopsis. Nature 463, 241–U130. Tanaka Y., Sugano S. S., Shimada T. and Hara-Nishimura I. (2013) Enhancement of leaf photosynthetic capacity through increased stomatal density in Arabidopsis. New Phytol. 198, 757–764. Tierney J. E., Russell J. M., Huang Y., Damste 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. (2010) A 35 Myr North American leaf-wax compound-specific carbon and hydrogen isotope record: Implications for C-4 grasslands and hydrologic cycle dynamics. Earth Planet. Sci. Lett. 299, 250–262. Vandewater P., Leavitt S. and Betancourt J. (1994) Trends in Stomatal Density and C-13/C-12 Ratios of Pinus-Flexilis Needles During Last Glacial-Interglacial Cycle. Science 264, 239–243. Vendramini P. F. and Sternberg L. da S. L. (2007) A faster plant stem-water extraction method. Rapid Commun. Mass Spectrom. 21, 164–168. 43 Wagner F., Below R., Klerk P. D., Dilcher D. L., Joosten H., Kürschner W. M. and Visscher H. (1996) A natural experiment on plant acclimation: lifetime stomatal frequency response of an individual tree to annual atmospheric CO2 increase. PNAS 93, 11705–11708. Woodward F. (1987) Stomatal Numbers Are Sensitive to Increases in Co2 from Preindustrial Levels. Nature 327, 617–618. Zhang Z., Sachs J. P. and Marchetti A. (2009) Hydrogen isotope fractionation in freshwater and marine algae: II. Temperature and nitrogen limited growth rate effects. Org. Geochem. 40, 428– 439. 44 Chapter 3 Comparison of three methods for the methylation of aliphatic and aromatic compounds This chapter was published in 2017 as: Lee H, Feakins SJ, Lu Z, Schimmelmann A, Sessions AL, Tierney JE, Williams TJ (2017) Comparison of three methods for the methylation of aliphatic and aromatic compounds. Rapid Communications in Mass Spectrometry 31:1633–1640. Abstract RATIONALE: Methylation protocols commonly call for acidic, hot conditions that are known to promote organic 1 H/ 2 H exchange in aromatic and aliphatic C—H bonds. Here we tested two such commonly-used methods and compared a third that avoids these acidic conditions, to quantify isotope effects with each method and to directly determine acidic-exchange rates relevant to experimental conditions. METHODS: We compared acidic and non-acidic methylation approaches catalyzed by hydrochloric acid, acetyl chloride and EDCI (1-ethyl-3-(3-dimethylaminopropyl)carbodiimide) / DMAP (4-dimethylaminopyridine) respectively. These were applied to two analytes: phthalic acid (an aromatic) and octacosanoic acid (an aliphatic). We analyzed yield by gas chromatography flame ionization (GC/FID) and hydrogen and carbon isotopic composition by isotope ratio mass spectrometry (GC/IRMS). We quantified the 1 H/ 2 H exchange rate on dimethyl 45 phthalate under acidic conditions with proton nuclear magnetic resonance ( 1 H-NMR) measurements. RESULTS: δ 2 H and δ 13 C values and yield were equivalent among three methods for methyl octacosanoate. The two acidic methods resulted in comparable yield and isotopic composition of dimethyl phthalate, however the non-acidic method resulted in lower δ 2 H and δ 13 C values perhaps due to low yields. Concerns over acid-catalyzed 1 H/ 2 H exchange are unwarranted as the effect was trivial over a 12-hour reaction time. CONCLUSIONS: We find product isolation yield and evaporation to be the main concerns in the accurate determination of isotopic composition. 1 H/ 2 H exchange reactions are too slow to cause measurable isotope fractionation over the typical duration and reaction conditions used in methylation. Thus, we are able to recommend continued use of acidic catalysts in such methylation reactions for both aliphatic and aromatic compounds. 3.1 Introduction Stable hydrogen isotopic compositions (δ 2 H) of organic compounds are widely studied in organic geochemistry, biochemistry, ecology, climate, and forensics. Applications include the study of the hydrologic cycle and climate (Sauer et al., 2001; Tierney et al., 2008; Feakins et al., 2016), ecology (Hobson et al., 1999; Gao et al., 2014; Dawson et al., 2015), biosynthetic pathways of lipids (Sessions et al., 1999; Osburn et al., 2011), paleoelevation (Polissar et al., 2009; Ponton et al., 2014), diagenesis of organic matter (Schimmelmann et al., 2006) and food authenticity and provenance (Camin et al., 2010; Tipple et al., 2012). In many compounds, hydrogen positions include those in C—H bonds that are non-exchangeable after biosynthesis 46 and preserved under low temperature sedimentary conditions. They carry the original biosynthetically fixed H, although even these H may be exchanged under elevated temperatures and over geologic time (Sessions et al., 2004; Sessions, 2016). In contrast, O—H bonds, as found in cellulose or fatty acids, contain H atoms that are readily exchangeable with ambient water in the plant (DeNiro and Epstein, 1981), soil, or sediments after burial (Schimmelmann et al., 2006). In order to recover signals relating to the original plant biosynthesis, analyses typically target only the carbon-bound hydrogens, by masking exchangeable O—H hydrogens. The preparative steps for cellulose involve ‗nitration‘ (i.e. formation of nitric acid esters (Epstein et al., 1976)), whereas derivatization of aliphatic and aromatic acids typically involves esterification with methanol. Here, we examine isotopic exchange associated with reaction conditions used for esterification prior to isotopic analysis. 3.1.1. Acid catalyzed esterification Organic compounds with carboxylic acid groups are commonly esterified with methanol prior to gas chromatography (GC) analysis (Sessions, 2006; de Groot, 2008) using acidic catalysts. Hydrochloric acid (HCl) and acetyl chloride (CH 3 COCl), which forms anhydrous HCl in methanol, are among the most widely used catalysts of aliphatic and aromatic acids. The carboxylic acid is initially protonated and undergoes an exchange reaction with the alcohol to give the reaction intermediate. Then the intermediate can lose a proton to become an ester. The equilibrium point of this reaction is displaced so that esterification proceeds virtually to completion in excess alcohol (Christie, 1989). However, the Brønsted acid catalyst can also promote hydrogens bound to an aromatic ring to undergo 1 H/ 2 H exchange via electrophilic aromatic substitution (Fig. 3.1) (Brown and Okamoto, 1958; Olah, 1971; Werstiuk and Kadai, 47 1974). Thus, during esterification, hydroxyl hydrogens in methanol can exchange with aromatic hydrogens in an analyte molecule and the δ 2 H values can equilibrate. Understanding acid-catalyzed isotope exchange during esterification is relevant to the study of aromatic compounds, but also indirectly for the study of aliphatic compounds such as the alkanoic acids derived from plant leaf waxes and microbial lipids, because aromatic structures are commonly used as part of the isotopic determination (Sauer et al., 2001). One method to determine the δ 2 H and carbon isotopic composition (δ 13 C) values of methyl groups added in alkanoic acid esterification steps involves the use of phthalic acid (Sessions et al., 2002). Phthalic acid of known isotopic composition (measured independently as a sodium salt) can be methylated and measured by gas chromatography isotope ratio mass spectrometry (GC/IRMS) to calculate the isotopic composition of the methyl hydrogens and carbon via mass balance (Sessions, 2006). Because phthalic acid has an aromatic ring, conventional acidic esterification conditions should enable 1H/ 2 H exchange on the aromatic hydrogens (Fig. 3.1) (Brown and Okamoto, 1958; Olah, 1971; Werstiuk and Kadai, 1974). Such exchange may result in an incorrect estimation of the overall dimethyl phthalate isotopic composition and mass balance estimation of the δ 2 H value of methyl hydrogens, needed when calculating the isotopic composition of the original fatty acid (Polissar and D‘Andrea, 2014). However the magnitude of such an effect is currently unknown. 48 Fig. 3.1. Schematic diagram showing the risk of acid-catalyzed 1 H/ 2 H exchange in dimethyl phthalate. 3.1.2. Non-acidic coupling reagents to promote esterification Alternatively, non-acidic approaches to methylation exist that may obviate the above concerns. In this study, we tested a non-acidic method using 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDCI) as a coupling reagent with 4-dimethylaminopyridine (DMAP) as the catalyst (Fig. 3.2) (Lehmann et al., 2006). Carbodiimides, due to their accessibility and versatility, are ranked as one of the most important classes of compounds in organic chemistry (Mikolajczyk and Kielbasiński, 1981). The most widely used carbodiimide reagent, dicyclohexylcarbodiimide (DCC), makes insoluble urea byproducts requiring additional cumbersome purifications (Zhang et al., 2004). Hence we chose EDCI, a water soluble carbodiimide with water soluble urea byproduct (Sheehan et al., 1961), combined with DMAP which has shown to increase the efficiency of esterification by 104 fold (Neises and Steglich, 1978). We compared this approach to conventional acidic methylation reactions. EDCI and DMAP are not expected to lead to 1H/ 2 H exchange because the reaction conditions are buffered, and the merits of these reagents for organic geochemical applications are tested here. 49 Fig. 3.2. The structures of non-acidic reagents EDCI and DMAP. 3.1.3. Method comparison In this study, we tested two acidic and one non-acidic methods of esterification as applied to phthalic acid and octacosanoic acid to evaluate any isotopic fractionations and practical issues arising with each using GC/IRMS and GC/FID approaches. We further directly measured the 1 H/ 2 H exchange rate in phthalic acid methylation with 5 vol% HCl using proton nuclear magnetic resonance ( 1 H-NMR). Based on the exchange rate, we estimate the proportion of aromatic H that are substituted over 12 hours of reaction time from the initial δD values of the reactants. The goal is to quantify the δD and δ 13 C perturbation possible with each esterification method and to make practical recommendations for esterification prior to GC/IRMS analyses. 50 3.2 Experimental 3.2.1 Phthalic acid and methanol of known isotopic composition The hydrogen isotopic composition of phthalate was measured in Na-phthalate to exclude exchangeable, carboxylic acid hydrogens and the δ 13 C value was measured from the free diacid via offline combustion at Indiana University. Offline combustion consisted of phthalic acid combustion in quartz ampoules and cryogenic purification of combustion gases in a vacuum line. The aromatic hydrogen isotopic composition was −95.5 ± 2.2 ‰ (1, n = 4) and the carbon had a δ 13 C value of −27.21 ± 0.02 ‰ (n = 4). The uncertainties reported represent replicate precision. Unquantified uncertainty in the absolute value may exist, and the largest source of concern would be water inclusion in the Na-phthalate crystals. In order to mitigate against the influence of water inclusions, the crystals are warmed before combustion. Methanol (MeOH) was 99.8% anhydrous and the δD value of the methyl group was −141 ± 3 ‰ (n = 3), calculated via mass balance between bulk methanol and hydroxyl hydrogen at Indiana University. Excess sodium metal and small amount of methanol were reacted in an evacuated, sealed ampoule to liberate H 2 gas from the hydroxyl hydrogen for isotopic determination. The uncertainties reported represent replicate precision. Unquantified uncertainty in the absolute value may exist, the largest source of such error would be associated with residual methanol vapor, although this is mitigated against by allowing the reaction to proceed for several days. Hydroxyl hydrogen has a δD value of −27 ‰ and thus can provide the relatively 2 H-enriched hydrogen source for exchange into the phthalate aromatic ring during acid-catalyzed equilibration. Other H + in the 5 vol% catalyst (from HCl and water) can exchange freely with methanol‘s hydroxyl hydrogen and their isotopic compositions are unknown. Nonetheless, the 51 dominant source of exchangeable hydrogens is from methanol which is 95 % of the reaction solvent. The δ 13 C value of the methanol was −46.77 ± 0.04 ‰, from bulk measurement. The isotopic composition of phthalic acid and methanol were used to calculate the predicted isotope values of dimethyl phthalate via Eq. 1 for hydrogen and Eq. 2 for carbon. (1) (2) 3.2.2 The acidic methods: HCl and CH 3 COCl For the HCl catalyzed methylation, 3 mg of phthalic acid or 3 mg of C 28 n-alkanoic acid were methylated with 1 mL of a mixture of 5 vol% HCl and 95 % MeOH in a tightly sealed 15 mL culture tube at 70 °C overnight. For the acetyl chloride catalyzed methylation, the same procedure was followed using CH 3 COCl (instead of HCl) and reacting at 50 °C overnight. Practical guidance on methylation protocols are provided in the supplement, and are briefly summarized here. Following overnight methylation reactions, the samples were removed from the heat, then, 1 mL of Milli-Q water (Millipore Milli-Q Plus QPAK 2) was added to produce an aqueous phase. The analyte was extracted using liquid-liquid extraction with 1 mL of dichloromethane (DCM) for dimethyl phthalate and hexane for n-alkanoic acid methyl esters. For each extraction (repeated 3 or more times), the mixture was vigorously shaken, left to partition into two phases and the organic phase, containing dimethyl phthalate or fatty acid methyl ester (FAME), was extracted by pipette and dried by passage through a column of anhydrous Na2SO4. The organic fraction was further purified by passing through a silica column (5 cm ⅹ 40 mm Pasteur pipette, 5 wt% 52 water-deactivated silica gel, 100-200 mesh) eluting with three column volumes of DCM for dimethyl phthalate and for FAME. 3.2.3 The non-acidic method: EDCI & DMAP A general procedure for synthesis of ester derivatives in Lehmann et al.,(Lehmann et al., 2006) was followed with adjustments for scale. 22.5 mg of EDCI (8 mol ratio to phthalic acid or octacosanoic acid) and 1.1 mg of DMAP (0.5 mol ratio) and 3 mg of phthalic acid or octacosanoic acid (C 28 n-alkanoic acid) were added to 1 mL of isotopically known methanol and the solution was kept at room temperature overnight. The solution was diluted with 1 mL of Milli-Q water. The methylated product was extracted through the addition of 1 mL of ethyl acetate. The mixture was shaken vigorously to allow liquid:liquid extraction to proceed, and then allowed to settle into two phases, and this process was repeated at least three times. The organic fraction, containing dimethyl phthalate or methyl octacosanoate, was dried by passage through a column of anhydrous Na 2 SO 4 . It was further purified by eluting the ester fraction through a silica column with three column volumes of ethyl acetate. 3.2.4 GC/MS/FID analysis The methylation products generated in this study, including dimethyl phthalate and C 28 FAME, were identified and quantified using gas chromatography (Agilent 6890 series) coupled with mass-selective detector (Agilent 5973; equipped with quadrupole mass analyzer and electron impact with ionization energy of 69.9 eV) and flame ionization detection (GC/MS/FID supplied by Agilent, Palo Alto, USA). The instrument was equipped with a Rxi-5 ms column (30 m ⅹ 0.25 mm, film thickness 0.25 m supplied by Restek Corporation, Bellefonte, PA, USA) with a 4 mL min−1 constant column flow, split passively 4:1 (based on theoretical calculations following 53 Poiseulle‘s law) between the MS (via a 80 cm ⅹ 0.25 mm silica capillary) and FID (via a 120 cm ⅹ 0.1 mm silica capillary). Helium was used as the carrier gas. The initial temperature of 50 °C was held for 3.5 min, followed by a temperature ramp of 20 °C min−1 to 300 °C maintained for 10 min. 1 L of sample was injected in splitless mode. Absolute abundance was calculated using a calibration curve of an in-house standard (mixture of n-alkanes and n-alkanoic acids) and their peak area response on the FID. Yields were calculated based on the chemical equation using masses of reagents, quantified product abundances and are reported as % of predicted product. All uncertainties are reported as standard deviations () unless otherwise stated. 3.2.5 GC/IRMS analysis Compound-specific hydrogen and carbon isotopic compositions were measured using GC/IRMS. Samples were introduced via a programmable temperature vaporizing (PTV) injector in splitless mode at 50 º C to a Thermo Scientific® Trace gas chromatograph. The gas chromatograph was equipped with a Rxi-5 ms column (30 m ⅹ 0.25 mm, film thickness 0.25 m) and connected via an Isolink with pyrolysis furnace at 1400 °C for hydrogen isotopic analysis, and via a combustion furnace at 1000 °C for carbon isotopic analysis, through a Conflo IV interface to a DeltaVPlus isotope ratio mass spectrometer (all supplied by Thermo Scientific, Bremen, Germany). The gas chromatograph was operated with a He flow of 1 mL min−1 from an initial temperature of 50 °C, followed by a ramp of 20 °C min−1 to 300 °C, and held for 5 min. For hydrogen, the H 3+ factor was measured daily to correct for the formation of H 3+ in the ion source (Sessions et al., 2001), and to check for linearity, with a mean value over the course of measurements reported here of 3.6 ppm mV−1  0.35 (n = 6) across a range of 1−8 V. For carbon, the standard deviation of CO2 pulses spanning 1–8 V was determined daily to check for 54 linearity and the standard deviation across that range averaged 0.02 ‰. Gas peaks of H2 or CO2 as appropriate were co-injected bracketing the analyte peaks during the GC/IRMS run. Two of these peaks were used for standardization between sample and standard, while the rest were treated as unknowns to assess precision. Data were then normalized to the Vienna Standard Mean Ocean Water (VSMOW)/Standard Light Antarctic Precipitation (SLAP) hydrogen isotopic scale via comparison with an external standard with 15 n-alkanes (C16-C30 n-alkane mixture of type A, isotopic standard from Indiana University), with δ 2 H values ranging from −46 to −227 ‰. For carbon, isotopic results were normalized to Vienna PeeDee Belemnite (VPDB) / lithium carbonate prepared by H. Svec (LSVEC) by comparing with A mix standard δ13C values ranging from −28.6 to −33.3 ‰. The standard was run daily throughout the sequence. Measured δ 2 H values of the samples are normalized using a regression between the known and measured values of the standards. The root-mean-square error of replicate analysis of the external standard (A mix) averaged 4.6 ‰ for hydrogen and 0.2 ‰ for carbon. The results are reported using conventional delta (δ) notation in permil (‰). 55 Table 3.1. Hydrogen and carbon isotope values calculated for C 28 and C 16 FAMEs and propagated uncertainties based on the precision and accuracy of dimethyl phthalate analyses. 3.2.6 1 H/ 2 H exchange rate using 1 H-NMR We conducted a proton nuclear magnetic resonance ( 1 H-NMR) kinetic experiment to measure the rate of 1 H/ 2 H exchange on dimethyl phthalate in 5 vol% HCl (12 M) using a Varian VNMRS-600 (supplied by Varian, Inc. now Agilent Technologies, Palo Alto, USA) at the University of Southern California. We reacted 30 mg of phthalic acid in a solution of 0.60 mL of methanol-d4 (C 2 H 3 O 2 H) and 5 vol% HCl in a Wilmad J Young NMR tube (type of NMR tube with a Teflon cap that can be attached to a vacuum line, supplied by Wilmad-labglass, Vineland, NJ, USA) sealed with plastic cap and parafilm, in an oil bath at 70 º C. We measured the phthalate H peak areas daily for ten days after the overnight esterification reaction. The tube was δ 2 H (‰) δ 13 C (‰) HCl CH 3 COCl EDCI/DMAP HCl CH 3 COCl EDCI/DMAP Mean dimethyl phthalate -107 -105 -137 -31.3 -31.5 -32.7 Uncertainties i. Precision † 1 2 1 0.1 0.0 0.4 ii. Accuracy° 16 18 14 0.2 0.4 1.6 Estimated methyl group -115 -112 -165 -47.9 -48.9 -54.5 Assigned FA value -100 -20 C 28 FAME calculated -101 -101 -103 -21.0 -21.0 -21.2 Propagated s * i. Precision 6 6 6 0.2 0.2 0.2 ii. Accuracy 6 6 6 0.2 0.2 0.3 C 16 FAME calculated -101 -101 -106 -21.6 -21.7 -22.0 Propagated s * i. Precision 6 6 6 0.2 0.2 0.2 ii. Accuracy 7 7 7 0.2 0.2 0.5 † Reproducibility of phthalic acid methylation and preparative chemistry. ° Difference between known and measured dimethyl phthalate value from Fig. 3. * Propagated standard deviation calculated following the approach of Polissar and D’Andrea [26] . 56 placed back in the oil bath after measurement to maintain the reaction conditions over the ten days. The peaks were analyzed with a relaxation delay of 10 sec. A decrease in peak area detected by 1 H-NMR indicated exchange of 1 H on the phthalate with 2 H from the methanol-d 4 solvent. 3.3 Results and discussion 3.3.1 Product yield We found product yields varied between methods and analytes. For dimethyl phthalate, the acidic methods resulted in > 70 % product yield, whereas the carbodiimide method resulted in yields < 30 % (Fig. 3.3, Supplemental Table 3.1). For methyl octacosanoate, all methods yielded > 80 % product yield, but again yields were highest for the acidic methods > 94 % (Supplemental Table 3.2). Low product yield (observed here for the dimethyl phthalate using the non-acidic method) is a concern given the potential for isotopic fractionation associated with incomplete reaction or recovery. Low yields are also problematic for many applications because sample size requirements increase. We investigate here the potential reasons for low yields and discuss our attempts to improve them. 3.3.1.1 Yield for acidic methods High yields were obtained using acidic methods for methylation, with acetyl chloride (mean yield 80 %,  = 5, n = 10) or HCl (mean yield 73 %,  = 5, n = 13) as the starting reagent, consistent with theoretical product yield for Fischer esterification (Christie, 1989). The higher yield for the acetyl chloride method is due to the lower water content of the reagent, compared to 57 HCl (12 M solution). These high yields were achieved after experimentation to optimize handling to minimize loss (Supplemental Table 3.3). For example, the boiling point of dimethyl phthalate is 283 °C and we observed significant removal of the product by GC/FID under normal solvent blow-down on a hotplate, thus we avoided using the hotplate. We also improved our yield by up to 10 % by switching from hexane (used for aliphatic esters) to DCM during liquid- liquid extraction, given the greater polarity of dimethyl phthalate. Still, we found dimethyl phthalate has a lower isolated yield than aliphatic esters (Supplemental Table 3.1), suggesting that extraction and purification of dimethyl phthalate could be further improved. Less than 100 % isolated yield could be due to either incomplete liquid-liquid extraction or evaporative loss prior to measurement. 3.3.1.2 Yield for carbodiimide method The isolated yield of dimethyl phthalate was low (mean 22 %) using the carbodiimide method which could be due to incomplete reaction or loss during purification. We monitored the reaction with thin layer chromatography (TLC) at the beginning and end of the reaction (after 12 hrs) and we confirmed that the product had formed while phthalic acid had been completely consumed. Additional evidence for reaction completion was obtained by testing for the absence of phthalic anhydride, a reaction intermediate for the carbodiimide method, at the end of the experiment by GC/MS. Thus, we can rule out incomplete reaction. Low product yield using carbodiimide reagents for esterification has been previously noted, due to the carboxylate being consumed during the formation of an N-acylurea byproduct (Tsakos et al., 2015). Addition of DMAP as a catalyst inhibits the byproduct formation and facilitates the reaction to form the ester, which has been shown to increase yield (Neises and Steglich, 1978). 58 Following this protocol, we obtained < 30 % expected yield of dimethyl phthalate, while obtaining > 80 % yield of methyl octacosanoate. The low yield of dimethyl phthalate may be explained by its greater volatility compared to methyl octacosanoate. We tested yield response to various reaction conditions such as reagent concentration, temperature and duration of reaction for the carbodiimide approach, although tests of intermediates showed that reactions went to completion. In case isolation and purification steps were the issues affecting low isolation yields we tested different solvents for extraction, as well as citric acid washing and brine washing in efforts to improve the isolated product yield. Each of these efforts failed to generate yield > 30 % (Supplemental Table 3.4). Having experimented considerably with this approach, we find it is not currently a pragmatic protocol for low concentration (3 mg) dimethyl phthalate derivatization applications. In contrast, the method does have promise as a viable method for derivatization of aliphatic acids. Nevertheless, we note additional practical challenges for the use of the carbodiimide method prior to GC analyses. We found it necessary to clean the syringe, inlet and trim the column after running the carbodiimide samples by GC/IRMS likely due to urea persisting in the mixture even after liquid-liquid extraction. If these compounds were injected in measurable quantity, due to their high polarizability and protic nature, they could interact with the GC column. We do not however observe any humps or fronting and tailing of the analyte peak in the chromatogram, our analytical results for both quantification and isotopic composition therefore appear to be unaffected by any such polar contaminants. Regardless of analyte, the carbodiimide method presented challenges for subsequent GC analyses that should be addressed if this approach is to be adopted for aliphatic GC applications. 59 3.3.2 Isotopic analyses The expected isotopic values of dimethyl phthalate calculated from mass balance (Eq. 1 and Eq. 2) were δ 2 H values of –123 ‰ and δ13C values of –31.1 ‰. For carbon isotopes, acidic methods successfully resulted in values that are close to expected (Fig. 3.3). Acetyl chloride and HCl methods produced δ13C values of –31.5 ± 0.0 ‰ (n = 3) and –31.3 ± 0.1 ‰ (n = 3). The non- acidic method resulted in a more negative value of –32.7 ± 0.4 ‰ (n = 4). The hydrogen isotopic composition of dimethyl phthalate differed between the acidic and non- acidic esterification methods (Fig. 3.3). Dimethyl phthalate produced with acetyl chloride and HCl resulted in δ 2 H values of −105 ± 2 ‰ (n = 3) and −107 ± 1 ‰ (n = 3) respectively, thus being equal within uncertainty, however +16 to +18 ‰ more enriched than expected. The carbodiimide method resulted in a more negative δ 2 H value of −137 ± 1 ‰ (n = 4). We calculated the initial methyl hydrogen and carbon isotopic composition of methanol (Sessions et al., 2002) by mass balance and found none of the methods accurately assessed the (known) initial methyl δ 2 H and δ 13 C values. Acidic methods produced 25 – 29 ‰ more positive estimates for methyl δ 2 H values whereas the carbodiimide method produced 25 ‰ more negative estimates compared to the known value of the methanol −141 ± 3 ‰, n = 3 (Table 3.1). This inaccuracy is greater than analytical errors, the largest of which is the calibration to the international VSMOW-SLAP isotopic scale (see methods). This suggests an isotopic fractionation during the methylation reaction and/or purification and one that differs between the acidic and non-acidic methods. 60 3.3.2.1 Is acid-catalyzed exchange a cause of 2 H-enrichment? We directly measured 1 H/ 2 H exchange rates for dimethyl phthalate by 1 H-NMR under conditions relevant to the acidic methylation method. We found the same exchange rates (slopes) within uncertainties for the hydrogens on alpha and beta carbon positions of dimethyl phthalate (Fig. 3.4). Using the mean slope as the exchange rate constant (4.94  10-4 hr-1), this corresponds to an exchange half-life (the time to exchange 50 % of aromatic hydrogens, ) of 1403 hours (58.5 days) at 70 °C. For a 100 ‰ disequilibrium in δ 2 H values between aromatic H and methanol hydroxyl H, this would induce a change in the δ 2 H value of the aromatic H of just 0.035 ‰ over a 1-hour reaction, which is not detectable by GC/IRMS. For reagents described here, with 70 ‰ disequilibrium in δ 2 H of dimethyl phthalate and methanol hydroxyl H, and a 12- hour reaction time (0.4 % exchange), we would expect 0.3 ‰ 2 H-enrichment. Even with a 12- hour reaction (0.4 % exchange) and a 1000 ‰ disequilibrium, such as might be encountered working on extraterrestrial samples, a shift of only 4 ‰ is predicted, which is still smaller than analytical uncertainties for GC/IRMS. Although 1 H/ 2 H exchange is taking place as expected, we directly establish, for the first time, that the rates are too slow to cause any significant changes in δ -values over the reaction conditions used in the HCl method. Still, when using strong acids as catalysts in methylation, it is good practice to not extend reaction times. 3.3.2.2 Is evaporation a cause of 2 H-fractionation? Guarding against evaporation is always a concern for volatile analytes intended for isotopic analyses, requiring careful laboratory handling. Vapor pressure isotope effects vary between molecules, depending on intermolecular vibrations (bonding) and frequency shifts on condensation, both of which vary with temperature (Höpfner, 1969). We directly tested 61 fractionation for dimethyl phthalate under worst-case scenario laboratory handling with evaporative losses to demonstrate the scale of possible errors. To mimic extreme sample preparation error, we evaporatively removed all solvent and ~45 % of the dimethyl phthalate using N2 blowdown, redissolved the compound in solvent and observed +14 ‰ 2 H-enrichment upon reanalysis. We also tested GC inlet conditions: we compared δ 2 H values of dimethyl phthalate injected via a PTV inlet in splitless mode, compared to split-mode at 150 °C. With the inlet-evaporation test, peak area decreased by half and we measured a 2 H-enrichment of +8 ‰. These tests indicate that it is possible to fractionate dimethyl phthalate during handling, such as complete solvent blow-down, or during analysis, via the use of heated, split injection modes. To avoid such effects, we dried the solvents under a gentle N 2 (g) stream without the use of heat and ran samples in splitless injection mode. Yields were still below 100 % for acidic methods, leaving open the possibility that some product evaporated 3.3.2.3 2 H- and 13 C-depletion using the carbodiimide method The carbodiimide method resulted in more negative δ 2 H and δ 13 C values than expected for dimethyl phthalate compared to the known isotopic composition. We infer a kinetic isotope effect, with the preferential reaction of the lighter isotope, by the carbodiimide method.(Marlier, 2001) Although we have not fully resolved the cause of the isotopic offset using this method (Figs. 3.3, 3.4) we suggest that securing higher yields and determining the potential for kinetic isotope effects with the carbodiimide reaction may be relevant questions for future research if this method is to be used as preparative method prior to GC/IRMS analyses. 62 Fig. 3.3. Dimethyl phthalate isolated yield and isotopic results. Carbon and hydrogen isotopic measurements are compared to known dimethyl phthalate values (i.e., target values) calculated by isotopic mass balance. Diamonds represent carbon isotope results whereas circles represent hydrogen isotopes. 3.3.3 Implications for aromatic and aliphatic acid esterification For methyl octacosanoate, prepared by each of three methods, the δ 2 H and δ 13 C values were within 3 ‰ and 0.5 ‰ uncertainty envelopes, respectively (Fig. 3.5). Thus, eah of the three methods are suitable preparative methods for GC/FID and GC/IRMS applications for long chain fatty acids, although the acidic methods remain the recommended methods. 63 For phthalic acid, we found that the two acidic methods are equivalent, and δ 13 C values were as predicted (within 0.4‰, similar within uncertainties), although we found δ 2 H values were elevated by 15‰ (Fig. 3.3). Although acidic exchange can lead to 2 H-enrichment, rates established here are far too slow to explain observations. Small 2 H-enrichment can be reproduced through evaporation, however evaporation cannot fully explain the observed enrichment given the high yields obtained, given that ~50% of the analyte would have to be lost to explain such an enrichment. Thus, the remaining offset may reflect a fractionation associated with the methylation reaction. Nevertheless, from a practical standpoint, we emphasize that volatility and evaporation is a greater concern than acidic exchange on aromatic compounds, when pursuing phthalic acid methylation or methylation of reagents with similar or lower boiling points. We recommend that analytical conditions be carefully designed and monitored to minimize evaporative losses and further isotopic enrichment, by testing quantitative yields by GC/FID in parallel with isotopic determinations by GC/IRMS. In contrast, the non-acidic carbodiimide method performed poorly in phthalic acid methylation experiments in terms of product yield and depletion in both 2 H (by up to 15 ‰). and 13 C (by up to 2 ‰; Fig. 3.3). We hypothesize the low isolated yield is related to the isotopic depletion but the yield could not be improved in our attempts shown here and thus the consequent isotope effects could not be directly tested. Since acidic-exchange is not quantitatively important for these applications, we have no reason to extend testing of the non-acidic method here. Phthalic acid methylation is commonly used to determine the isotopic composition of methyl groups in methanol used for fatty acid esterification (Sessions, 2006). While the different methods induce isotopic offsets from predicted values in this study (Table 1), the error in the isotopic composition of the methanol becomes trivial for long chain n-alkanoic acids due to the 64 small proportion of methyl hydrogens and carbon. For example, a 54 ‰ offset in methyl δ 2 H becomes 2 ‰ for δ 2 H of C 28 FAME and 5 ‰ for C 16 FAME and a 7 ‰ offset in methyl δ 13 C becomes 0.2 ‰ and 0.3 ‰ for δ 13 C of C 28 and C 16 FAME. These accuracy and precision errors contribute to methyl group uncertainty, and to the propagated uncertainties in the calculation of analyte δ 2 H values (Polissar and D‘Andrea, 2014). Overall, the magnitude of methyl hydrogen uncertainty on derivatized compounds remains small relative to other sources of uncertainties in the estimation of high molecular weight aliphatic acids e.g., C 28 n-alkanoic acid, a constituent of plant leaf waxes (Table 1). For short chain lipids such as the C 16 n-alkanoic acid (common in microbial, plant and animal cells) the mean isotope values differ more between methods (up to 5 ‰) and uncertainties increase. For practical guidance, we provide detailed protocol and error propagations in the supplement. Fig. 3.4. Pseudo-first order rate (k nmr ) of HCl-mediated 1 H/ 2 H exchange rate in dimethyl phthalate. Hydrogens attached to the beta carbon position (red symbols, red line) and to the alpha carbon (blue triangles, blue line). 65 Fig. 3.5. Methyl octacosanoate isolated yield, carbon and hydrogen isotopic composition. Error bars are replicate instrument precision, results are the same within analytical uncertainty. 3.4 Conclusions Methyl esterification is a commonly used derivatization procedure for both aromatic and aliphatic compounds with carboxylic acid groups prior to GC analysis. We have studied hydrogen and carbon isotopic fractionations during methylation using two widely used acidic catalysts and one non-acidic method using a carbodiimide. We found practical advantages to the acidic methods over the carbodiimide method, notably higher isolated yields. We determined that rates of 1 H/ 2 H exchange (4.94  10 -4 hr -1 ) are too slow to be of quantitative importance and thus find that the acidic conditions do not present problems for most applications, even with aromatic compounds. We thus recommend following one of the acidic protocols. 66 A phthalic acid standard is commonly used to determine the isotopic composition of methanol used to derivatize both the standard and the aliphatic or aromatic acid analyte for GC/IRMS applications. Using phthalic acid and methanol standards of known values here, we establish that for both carbon and hydrogen isotopic analyses the carbodiimide method produces systematically more depleted values compared to the acidic methods. As we have shown that acidic-exchange is not quantitatively significant, the greatest issues appear to be low yield (carbodiimide method) and evaporation (under poor laboratory handling scenarios). Although we observe a small systematic 13 C- and 2 H-enrichment in dimethyl phthalate from known reagents using the acidic methods, the propagated errors for long chain n-alkanoic acid applications are smaller than analytical uncertainty of δ 2 H by GC/IRMS at least for high molecular weight analytes. Acknowledgements This work was supported by funding from the US American Chemical Society Award #PRF- 53747-ND2 to SF, from the US National Science Foundation Award CHE-1566167 to TW, and from the USC Provost fellowship to HL. Federal support of NMR spectrometers (NSF DBI- 0821671, CHE-0840366; NIH S10 RR25432) is gratefully acknowledged. We thank Elias Karkabi for discussions. This manuscript was improved with the comments from three reviewers and Editor John Monaghan. 67 References Brown H. C. and Okamoto Y. (1958) Electrophilic substituent constants. Journal of the American Chemical Society 420, 4979–4987. Camin F., Bontempo L., Ziller L., Piangiolino C. and Morchio G. (2010) Stable isotope ratios of carbon and hydrogen to distinguish olive oil from shark squalene-squalane. Rapid Communications in Mass Spectrometry 24, 1810–1816. Christie W. W. (1989) Gas Chromatography and Lipids : a practical guide., Oily Press, Ayr. Dawson K. S., Osburn M. R., Sessions A. L. and Orphan V. J. (2015) Metabolic associations with archaea drive shifts in hydrogen isotope fractionation in sulfate-reducing bacterial lipids in cocultures and methane seeps. Geobiology 13, 462–477. DeNiro M. J. and Epstein S. (1981) Isotopic composition of cellulose from aquatic organisms. Geochimica et Cosmochimica Acta 45, 1885–1894. Epstein S., Yapp C. J. and Hall J. H. (1976) The determination of the D/H ratio of non- exchangeable hydrogen in cellulose extracted from aquatic and land plants. Earth and Planetary Science Letters 30, 241–251. Feakins S. J., Bentley L. P., Salinas N., Shenkin A., Blonder B., Goldsmith G. R., Ponton C., Arvin L. J., Wu M. S., Peters T., West A. J., Martin R. E., Enquist B. J., Asner G. P. and Malhi Y. (2016) Plant leaf wax biomarkers capture gradients in hydrogen isotopes of precipitation from the Andes and Amazon. Geochimica et Cosmochimica Acta 182, 155–172. Gao L., Edwards E. J., Zeng Y. and Huang Y. (2014) Major evolutionary trends in hydrogen isotope fractionation of vascular plant leaf waxes ed. R. Aroca. PLoS ONE 9, 1–10. de Groot P. A. (2008) Handbook of Stable Isotope Analytical Techniques., Elsevier, Amsterdam. Hobson K. A., Atwell L. and Wassenaar L. I. (1999) Influence of drinking water and diet on the stable-hydrogen isotope ratios of animal tissues. Proceedings of the National Academy of Sciences of the United States of America 96, 8003–6. Höpfner A. (1969) Vapor Pressure Isotope Effects. Angewandte Chemie International Edition in English 8, 689–699. Lehmann F., Pettersen A., Currier E. A., Sherbukhin V., Olsson R., Hacksell U. and Luthman K. (2006) Novel potent and efficacious nonpeptidic urotensin II receptor agonists. Journal of Medicinal Chemistry 49, 2232–2240. Marlier J. F. (2001) Multiple isotope effects on the acyl group transfer reactions of amides and esters. Accounts of Chemical Research 34, 283–290. 68 Mikolajczyk M. and Kielbasiński P. (1981) Recent developments in the carbodiimide chemistry. Tetrahedron 37, 233–284. Neises B. and Steglich W. (1978) Simple Method for the Esterification of Carboxylic Acids. Angewandte Chemie International Edition in English 17, 522–524. Olah G. A. (1971) Mechanism of electrophilic aromatic substitutions. Accounts of Chemical Research 4, 240–248. Osburn M. R., Sessions A. L., Pepe-Ranney C. and Spear J. R. (2011) Hydrogen-isotopic variability in fatty acids from Yellowstone National Park hot spring microbial communities. Geochimica et Cosmochimica Acta 75, 4830–4845. Polissar P. J. and D‘Andrea W. J. (2014) Uncertainty in paleohydrologic reconstructions from molecular δD values. Geochimica et Cosmochimica Acta 129, 146–156. Polissar P. J., Freeman K. H., Rowley D. B., McInerney F. A. and Currie B. S. (2009) Paleoaltimetry of the Tibetan Plateau from D/H ratios of lipid biomarkers. Earth and Planetary Science Letters 287, 64–76. Ponton C., West A. J., Feakins S. J. and Galy V. (2014) Leaf wax biomarkers in transit record river catchment composition. Geophysical Research Letters 41, 6420–6427. Sauer P. E., Eglinton T. I., Hayes J. M., Schimmelmann A. and Sessions A. L. (2001) Compound-specific D/H ratios of lipid biomarkers from sediments as a proxy for environmental and climatic conditions. Geochimica et Cosmochimica Acta 65, 213–222. Sauer P. E., Eglinton T. I., Hayes J. M., Schimmelmann A. and Sessions A. L. (2001) Compound-specific D/H ratios of lipid biomarkers from sediments as a proxy for environmental and climatic conditions. Geochimica et Cosmochimica Acta 65, 213–222. Schimmelmann A., Sessions A. L. and Mastalerz M. (2006) Hydrogen isotopic (D/H) composition of organic matter during diagenesis and thermal maturation. Annual Review of Earth and Planetary Sciences 34, 501–533. Sessions A. L. (2016) Factors controlling the deuterium contents of sedimentary hydrocarbons. Organic Geochemistry 96, 43–64. Sessions A. L. (2006) Isotope-ratio detection for gas chromatography. Journal of Separation Science 29, 1946–1961. Sessions A. L., Burgoyne T. W. and Hayes J. M. (2001) Determination of the H3 Factor in hydrogen isotope ratio monitoring mass spectrometry. Analytical Chemistry 73, 200–207. Sessions A. L., Burgoyne T. W., Schimmelmann A. and Hayes J. M. (1999) Fractionation of hydrogen isotopes in lipid biosynthesis. Organic Geochemistry 30, 1193–1200. 69 Sessions A. L., Jahnke L. L., Schimmelmann A. and Hayes J. M. (2002) Hydrogen isotope fractionation in lipids of the methane-oxidizing bacterium Methylococcus capsulatus. Geochimica et Cosmochimica Acta 66, 3955–3969. Sessions A. L., Sylva S. P., Summons R. E. and Hayes J. M. (2004) Isotopic exchange of carbon- bound hydrogen over geologic timescales. Geochimica et Cosmochimica Acta 68, 1545–1559. Sheehan J., Cruickshank P. and Boshart G. (1961) Convenient Synthesis of Water-Soluble Carbodiimides. The Journal of Organic Chemistry 26, 2525–2528. Tierney J. E., Russell J. M., Huang Y., Damste 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., Chesson L. A., Erkkila B. R., Cerling T. E. and Ehleringer J. R. (2012) B-HIVE: Beeswax hydrogen isotopes as validation of environment, part II. Compound-specific hydrogen isotope analysis. Food Chemistry 134, 494–501. Tsakos M., Schaffert E. S., Clement L. L., Villadsen N. L., Poulsen T. B., Sasaki M., Mynott R., Lehmann C. W., Ye T., Littlefield B. A., Giannakakou P., Hamel E., Watashi K., Kobayashi S., Sugawara F., Murthi K. K., Gentile L. N. and Liu J. H. (2015) Ester coupling reactions – an enduring challenge in the chemical synthesis of bioactive natural products. Nat. Prod. Rep. 32, 605–632. Werstiuk N. H. and Kadai T. (1974) The high temperature and dilute acid (HTDA) procedure as a general method of replacing aromatic hydrogen by deuterium. II. Canadian Journal of Chemistry 52, 2169–2171. Zhang M., Vedantham P., Daniel L. Flynn A. and Hanson P. R. (2004) High-Load, Soluble Oligomeric Carbodiimide: Synthesis and Application in Coupling Reactions. J. Org. Chem. 69, 8340–8344. 70 Chapter 4 Quantitative determination of lignin methoxyl concentration and isotopic composition Hyejung Lee a , Sarah Feakins a , Xiaojuan Feng b a Department of Earth Sciences, University of Southern California, Los Angeles, California, USA; b State Key Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China; Abstract Lignin is the second most abundant biochemical made by plants after cellulose. Its constituents include methoxyl groups that are targets for analysis both for structural characterization and stable hydrogen and carbon isotopic analyses. Here we test quantitative recovery and analysis associated with two methods of preparation and sample introduction for gas chromatography of the analyte iodomethane, contrasting headspace and liquid methods (dissolved in isooctane). We also analyze carbon and hydrogen isotopic composition and sensitivity to analytical conditions. The goal is to develop applications for the quantification and isotopic determination of lignin methoxyl components in a range of substrates, including isolated lignin and components, tree wood, as well as sedimentary lignin from peat, lignite and as modified in coals. We establish quantitative recovery of methoxyl from solid standards of phenolic compounds of known stoichiometry using the liquid method. Solid standards are convenient for quantification and 71 isotope biogeochemical applications. For hydrogen isotopic analyses quantitative recovery is important, with D-depletion associated with incomplete conversion. 4.1 Introduction Lignin, a large heterogeneous aromatic polymer, is the second most abundant plant biochemical after cellulose. In plants, lignin serves primarily as structural support by chemically connecting to cellulose and hemicellulose in the fiber walls (Hwanc, 1992) and it represents a significant part of plant litter input to soils (Thevenot et al., 2010). Lignin is thus an important component in organic-rich soils including peats and their buried counterparts (immature) lignites (Philp, 1985). Fundamentally, the lignin polymer comprises of three types of monomers: p-coumaryl alcohol (H unit), coniferyl alcohol (G unit), and sinapyl alcohol (S unit) linked predominantly with –O– 4 ( –aryl ether) bonds via oxidative coupling reactions in varying proportions depending on the morphological parts and positions within the plant (Adler, 1977; Boerjan et al., 2003; Vanholme et al., 2010). One common feature of lignin polymer and its phenol oxidation derivatives is methoxyl groups (–OCH 3 ) which has been used to characterize the heterogeneous macromolecule. Vanillyl phenols, derived from the G unit monomer, contain one methoxyl group, whereas syringyl phenols (S unit) contain two, and H unit derivatives have no methoxyl groups. The hydrogens in the methoxyl groups are of interest for isotope biogeochemistry due to the stability of the C–H bonds, making them suitable for hydrogen isotopic analyses to determine primary isotopic signatures during plant biosynthesis (Schimmelmann et al., 2006). In contrast, the bulk wood and lignin includes –OH groups where the hydrogen is exchangeable, leading to incorporation of secondary signals overprinting the primary, such as with water in the plant and in sedimentary burial (DeNiro and Epstein, 1981). 72 The hydrogen isotopic composition of the methoxyl groups (δD methoxyl ) on lignin from tree wood has been shown to record plant source water (Keppler et al., 2007; Feakins et al., 2013a; Anhäuser et al., 2017b) offering a proxy for past precipitation isotopic composition for paleoclimate reconstructions (Anhäuser et al., 2014, 2017b, 2018; Mischel et al., 2015; Riechelmann et al., 2016). δD methoxyl values have also been used for forensics, revealing the sourcing of plant materials (Keppler and Hamilton, 2008; Gori et al., 2015; Konopleva et al., 2017). The carbon isotopic composition of methoxyl groups has also been investigated (Greule et al., 2009) for climate (Riechelmann et al., 2016) and food authentication (Greule et al., 2010; Hansen et al., 2014). In the paper pulping and biochemical industries, methoxyl concentrations have been of interest for lignin structural characterization (Li et al., 2012; Sumerskii et al., 2017). No prior publications to our knowledge have reported both the quantity and isotopic composition of methoxyl groups, however, although such a combination may have merit both to monitor for quantitative recovery during sample preparation and may provide new insights for geological applications. Here we both quantify methoxyl yields and analyze hydrogen and carbon isotopic composition of methoxyl groups in a suite of standards and a range of substrates in order to calibrate and test the method for its potential applications to geological and biogeochemical research questions. 4.1.1 Analytical considerations The first step in lignin methoxyl analysis involves isolation of the methoxyl groups. The classic procedure (Zeisel, 1885), uses hydroiodic acid (HI) to cleave the ether bonds and yield iodomethane (CH 3 I). Iodomethane concentrations can be determined via gravimetry (Hadžija 73 and Tonković, 1975), titration (Vieböck and Schwappach, 1930) or gas chromatography (GC) (Keppler et al., 2007; Li et al., 2012). Several studies have used the GC approach to quantify (Li et al., 2012; Sumerskii et al., 2017) and determine the hydrogen (Keppler et al., 2007; Greule et al., 2008; Feakins et al., 2013a; Anhäuser et al., 2015, 2018) and carbon (Greule et al., 2010; Hansen et al., 2014; Anhäuser et al., 2015) isotopic composition of iodomethane. Analytical methods for the quantification and isotopic composition of iodomethane typically use gas chromatography, and the main difference in analytical conditions has been the mode of sample introduction which has included liquid injection (Baker, 1996), headspace injection with a dedicated headspace autosampler (Li et al., 2012) or without a headspace autosampler (Greule et al., 2008; Feakins et al., 2013b). Here we measure iodomethane from standards with known stoichiometry in order to constrain yields and sources of uncertainty, testing the headspace and liquid approaches. The goal is to optimize methods of analyzing iodomethane, i.e. to improve analytical precision in quantification and isotopic determination towards a range of potential geologic, biologic and forensic applications. 4.2 Materials and methods 4.2.1 Materials We selected a series of analytes for iodomethane quantification and isotopic determinations, including iodomethane (CAS # 74-88-4, Sigma-Aldrich), as well as in-house standards: lignin monomers, lignin, wood, and peat, each expected to evolve iodomethane upon reaction with HI acid. The iodomethane and lignin monomers of known stoichiometry serve as standards for quantification of yields. Lignin, wood and peat were homogenized in large batches to achieve 74 like-with-like standards for assessment of uncertainties associated with reaction chemistry and analytical conditions in real world samples of unknown composition. The four in-house standards established here are USC Lignin, Bamboo, Poplar, and Irish Peat. Samples of lignite and coal were also measured to test a range of geologically relevant substrates. 4.2.2 Iodomethane isolation and injection Iodomethane (CAS # 74-88-4, Sigma-Aldrich) was analyzed via gas chromatography for both quantification or isotopic composition either directly by headspace or dissolved in isooctane (hereafter, the ―liquid method‖). 1 – 10 L of iodomethane was injected into 2 mL GC vials with Al crimp caps with PTFE/rubber TF2 septum gas tight septa for the headspace approach, whereas for the liquid approach iodomethane was dissolved in 500 mL of isooctane. For phenol compounds, USC standards, and samples, 1 – 100 mg of material was reacted with 0.1 – 0.15 mL of hydroiodic acid (55 % HI) in 2 mL GC vials with crimp caps. Samples were heated at 120 °C for 30 min while shielded from light. Vials were cooled, held at ambient conditions (~22 °C) for at least 30 min to allow the iodomethane product to equilibrate into headspace. Samples were neutralized via the injection of 0.15 mL 5M KOH through the septum with additional 0.05 – 0.1 mL until neutralized (as detected by a color change). For headspace analyses, samples were ready to be measured at this point (―headspace step‖). For the liquid method, we next performed liquid – liquid extraction to partition the iodomethane into the organic, taking advantage of the high partition coefficient of iodomethane into isooctane from gas or aqueous phase at room temperature (25 °C) (Stephens et al., 2012). 250 L of isooctane was injected through the septum into the vial at the ―headspace step‖ above, and the mixture was vigorously shaken for 30 sec, centrifuged to facilitate separation into organic and 75 aqueous layers. The organic phase, containing iodomethane was extracted by syringe 3 times and the fractional (vol/vol) recovery was recorded. 4.2.3 Identification and quantification by GC – MS/FID Iodomethane was identified and quantified using gas chromatography (Agilent 6890 series) coupled with a single quadrupole mass – selective detector (Agilent 5973, using electron ionization with an ionization energy of 69.9 eV) for identification and flame ionization detection (GC-MS/FID) for quantification. The instrument was equipped with a Rxi-5 ms column (30 m × 0.25 mm, film thickness 0.25 m supplied; Restek Corp., Bellefonte, PA, USA) with a 2 mL min-1 constant flow rate. Helium was used as the carrier gas. For the headspace method, the GC method had an initial temperature of 31 °C, held for 3 min, followed by a temperature ramp of 20 °C min-1 up to 70 °C and held for 1 min. For the liquid method, the GC method was further ramped to a final temperature of 120 °C and the mass spectrometer detector was turned off at 3.70 min to avoid detection of the high boiling point solvent (isooctane) that elutes after the volatile analyte peak (iodomethane). Samples were injected in splitless mode for both headspace and liquid methods. Calibration was achieved with injection of iodomethane of known concentrations across the 1 – 10 L to define peak area response on the flame ionization detector. Calibration that accounts for the larger uncertainty associated with reaction chemistry and recovery of iodomethane from samples was achieved by defining response curves using phenols with one and two methoxyl groups, vanillyl and syringyl phenols respectively, in the range 0.5 – 10 mg. 76 4.2.4 Hydrogen isotopic analyses The hydrogen isotopic values (δD) of iodomethane were determined at USC using a Trace GC connected via a GC-Isolink pyrolysis furnace (at 1400 °C), passing through a cold trap (liquid nitrogen) for HI entrapment (Feakins et al., 2013b), and a Conflo IV interface to Delta VPlus IRMS (all supplied by Thermo Scientific, Bremen, Germany). The cold trap consisted of a fused silica capillary (1 m × 0.32 mm ID) with 30 cm of a u-shaped bend in the capillary immersed in a dewar of liquid nitrogen (Feakins et al., 2013b). The trapped HI was periodically purged by removing the capillary from the cold trap and venting to an exhaust vent. For the headspace method, 10 – 60 L of gas from sample vials was manually injected using a gas tight syringe into a Programmable Temperature Vaporizing (PTV) inlet operated at constant temperature (200 °C) and with a split ratio of 25 to generate a narrow peak chromatography of the gas phase analyte. The GC was fitted with a ZB-5 ms column (30 m × 0.25 mm × 1 m) and conditions included a 2 mL min-1 constant column flow. The GC oven temperature started at 33 °C, which was held for 2.5 min, and was followed by a temperature ramp of 20 °C min-1 up to 46 °C and was held for 1 min. For the liquid method, 1 – 10 mL of isooctane containing iodomethane was injected using a gas tight syringe into a split/splitless (SSL) inlet operated at constant temperature (200 °C) in splitless mode. The flow rate was at 3 mL min-1 with the oven temperature starting at 33 °C, held for 2.5 min, followed by a 20 °C min-1 ramp up to 130 °C which was held for 0.6 min to remove the isooctane. The GC Isolink backflush multi-functional valve controller (MFVC) was heated to 65 º C to aid backflush of isooctane. 77 Four peaks of hydrogen reference gas bracket the iodomethane analyte peak during the GC- IRMS run. One of the initial peaks was used for standardization of the isotopic analyses, while the other three bracketing peaks were treated as unknowns (precision averaged 0.4 ‰, 1, n = 208). The known reference peak was set by comparing with an external standard obtained from A. Schimmelmann, Indiana University, Bloomington, containing 15 n-alkane compounds (C 16 to C 30 ), with δD values spanning –9 to –254 ‰. The RMS error determined by replicate measurements of the standard across the course of analyses was 4.2‰. Data were then normalized to the Vienna Standard Mean Ocean Water / Standard Light Antarctic Precipitation (VSMOW/SLAP) isotopic scale by comparison with both the Amix and an external standard of 99.7% purity iodomethane analyzed by offline combustion and analysis by dual inlet IRMS with δD values –95.6 ± 1.6 ‰ n = 6 (the δD value was analyzed and supplied by A. Schimmelmann, Indiana University, Bloomington). The results are reported using conventional delta notation. 4.2.5 Carbon isotopic analyses The δ 13 C values of iodomethane were determined at USC using a Trace GC connected via a GC- Isolink combustion furnace (at 1000 °C), and a Conflo IV interface to Delta VPlus IRMS (all supplied by Thermo Scientific, Bremen, Germany). 30−50 L of headspace gas from sample vials was injected using a gas tight syringe into a Programmable Temperature Vaporizing (PTV) inlet operated at constant temperature (200 °C) and with a split ratio of 25 needed to generate narrow peak chromatography of the gas phase analyte. The GC was fitted with a ZB-5ms column (30 m x 0.25 mm x 1 m) and conditions included a 2 mL/min constant column flow, and a temperature program at 33 °C for 2.5 min followed by a ramp of 20 °C/min to 46 °C, held for 1 minute. 78 Peaks of carbon dioxide reference gas bracket the iodomethane analyte peak during the course of the GC-C-IRMS run. One of the initial peaks was used for standardization of the isotopic analyses, while the other three bracketing peaks were treated as unknowns (precision averaged 0.08 ‰, 1, n = 353). δ 13 C value of the initial peak was set by comparing with an external standard, A6 mix, containing 15 n-alkane compounds (C 16 to C 30 ), with δ 13 C values ranging from –33.3 to –28.6 ‰, obtained from A. Schimmelmann, Indiana University, Bloomington. Data were then normalized to the Vienna Pee Dee Belemnite (VPDB) / lithium carbonate prepared by H. Svec (LSVEC) isotopic scale by comparison with both the A6 mix and an external standard of 99.7 % purity CH 3 I analyzed by offline combustion and analysis by dual inlet IRMS with δ 13 C values −54.09 ± 0.02 ‰, n = 6 (the δ 13 C value was analyzed and supplied by A. Schimmelmann, Indiana University, Bloomington). The results are reported using conventional delta notation. 4.2.6 Lignin phenol analyses The four USC standards (Lignin, Bamboo, Poplar, and Irish Peat) were analyzed for phenol ratios at the Chinese Academy of Sciences. Ground wood was transferred into Teflon lined bombs for CuO oxidation to isolate lignin derived phenols together with 0.5 g CuO, 100 mg ammonium iron (II) sulfate hexahydrate and 3 mL of 12 M NaOH under N 2 and heated at at 170 º C for 2.5 h. After the reaction, the water phase was acidified to pH 1 with 6 M HCl and kept for 1 h at room temperature in the dark to prevent reactions of cinnamic acids. After centrifugation (2500 rpm for 30 min), the supernatants were liquid-liquid extracted with ethyl acetate. The extracts were concentrated by rotary evaporation, spiked with a known amount of internal standard (ethyl vanillin), and derivatized with N,O-bis-(trimethylsilyl) trifluoroacetamide 79 (BSTFA) and pyridine (70°C, 1 h) to yield trimethylsilyl (TMS) derivatives before quantification on GC–MS. Lignin phenols were quantified using internal standards on a Trace 1310 gas chromatograph coupled to an ISQ mass spectrometer (Thermo Fisher Scientific, USA) using a DB-5MS column (30 m × 0.25 mm i.d., film thickness, 0.25 μm). The GC oven temperature was held at 65°C for 2 min, increased from 65 to 300°C at a rate of 6°C min -1 with final isothermal hold at 300°C for 20 min. Helium was used as carrier gas (0.8 mL min −1 ). The mass spectrometer was operated in the electron impact mode (EI) at 70 eV and scanned from 50 to 650 daltons. Vanillyl (vanillin, acetovanillone, vanillic acid), syringyl (syringaldehyde, acetosyringone, syringic acid), and cinnamyl (p-coumaric acid, ferulic acid) phenols were summarized to represent lignin phenols. 4.3 Results 4.3.1 Calibration for iodomethane quantification The iodomethane standard was analyzed at different concentrations to establish a standard calibration for quantification of unknowns. To achieve a calibration, we placed 1 – 10 L of CH 3 I(l) into a 2 mL GC crimp-capped vial and injected 1 – 5 L of headspace gas into the gas chromatograph. Concentration of CH 3 I is reported in g converted from L using density (2.28 g mL -1 ) and molecular weight (141.94 g mol -1 ). The headspace method resulted in variable peak areas depending on the injection concentration (Figure 4.1a-c). We found a (3x) greater response with large injections (5 L, Figure 4.1c) of low concentration relative to smaller injection volumes (1 L, Figure 4.1a) at higher 80 concentrations. Above 5 – 8 g of iodomethane, the peak area response plateaued, likely reflecting incomplete transfer due to condensation on the surface of the vial (Figure 4.1a-c). We observed condensation of iodomethane when >4 L was placed in the GC vial (> 9.12 mg of iodomethane in 2 mL headspace volume). This implies headspace inside the vial is not homogenous at room temperature especially at higher concentrations. Fig. 4.1. Comparison of the (a-c) headspace methods (open symbols) and (d) liquid (solid symbols) of iodomethane quantification by GC–FID. Samples were injected in headspace mode, with a range of injection volumes (a) 1 L, (b) 2 L and (c) 5 L and liquid mode in (d) 1 l isooctane. Response is linear for liquid mode (d), but for headspace injections non-linear, with saturation response curves indicating incomplete transfer from vial or syringe. In order to avoid condensation issues with headspace analysis, we adopted and modified the liquid method described in Baker (1996) that analyzed iodomethane dissolved in pentane. We used isooctane, to dissolve the iodomethane, selecting a high boiling point organic solvent to achieve better separation between solvent and analyte during analysis via GC. When dissolving 1 – 10 L of iodomethane into isooctane and injecting liquid isooctane into the GC we find a linear 81 peak area response with increasing iodomethane concentration (Figure 4.1d). This indicates that dissolution into isooctane is effective and avoids the problems observed with incomplete transfer using the headspace method (Figure 4.1a-c), that may result from the analyte remaining in the aqueous phase, condensing onto the glass vial or syringe. 4.3.2 Calibration for methoxy group quantification Lignin phenols of known stoichiometry, reacted with HI, should also yield predictable amounts of iodomethane. However, some of the analyte may be lost in the sample preparation step. We therefore compare the iodomethane-calibration to the lignin phenol response curve, both using the liquid method (Figure 4.2). We achieved 85.2 % (1 = 5.89%) recovery yields of iodomethane from the four phenolic compounds tested in the range of 0.7 – 12.7 g iodomethane equivalent to 0.8 – 15.0 mg of V phenol or 0.5 – 8.9 mg of S phenols (Figure 4.2, Table 4.1). We use this phenol-calibration to quantify lignin, wood, lignite, and coal samples containing unknown stoichiometry, assuming consistent yield. Fig. 4.2. Comparison of the iodomethane standard quantification calibration (black filled circles) with the quantification calibration of iodomethane produced via the Ziesel method and liquid-liquid extraction of phenolic 82 substrates of known stoichiometry (open symbols). The phenol-calibration quantifies the iodomethane yields (open symbols) from four different lignin phenols (Vanillin, Vanillic acid, Syringaldehyde, Syringic acid) with one or two methoxyl groups (V and S groups respectively) using the liquid method. The chemistry yield, including reaction and calibration is linear across the measured range. Table 4.1. List of materials with known and unknown methyl mass analyzed using the liquid method. For lignin oxidative phenols, reaction yields from the liquid method are reported relative to known methyl mass from stoichiometry. For wood and lignin standards, and the lignite and coal samples, quantification was achieved relative to the lignin monomeric calibration. 4.3.3 Iodomethane quantification in heterogeneous wood samples Based upon GC-FID iodomethane measurements and the phenol calibration, we determined the abundance of methoxyl groups in wood products reporting as methyl content (wt %) as it is the methyl portion of the methoxy group that is cleaved during the Ziesel reaction. We quantified the methyl abundance in a suite of standards and samples (Table 4.1) including the USC Lignin Standard, a kraft lignin product (i.e. the lignin product from the manufacturing process that removes lignin from paper pulp, purchased from Sigma Aldrich), wood standards (USC Bamboo and Poplar Standards, supplier: Amazon.com, ground in a Wiley Mill). We found the USC Poplar and Bamboo Standards contained fewer methyl groups (1.9 and 2.2 wt% methyl Methyl content (wt %) Phenol ratio Phenol concentration (g/g) Materials Molecular formula Stoichiometry Liq method estimate (wt %) Liq method yield (%) S/V C/V P/V V S C P DHA Dimers Lignin phenols Vanillin C 8 H 8 O 3 9.9 8.5 86.1 Vanillic acid C 8 H 8 O 4 8.9 8.0 89.8 Syringealdehyde C 9 H 10 O 4 16.5 13.7 83.4 Syringic acid C 9 H 10 O 5 15.1 13.9 91.6 In-house standards and samples with unknown methyl mass USC Lignin 3.0 0.04 0.01 0.04 20.60 0.74 0.28 0.87 0.06 0.75 USC Bamboo 1.9 2.98 1.45 0.36 9.55 28.44 13.82 3.43 0.05 0.58 USC Poplar 2.2 3.02 0.01 0.02 8.76 26.51 0.11 0.13 0.03 1.08 USC Irish Peat 0.6 0.90 0.76 0.88 3.92 3.54 2.98 3.46 1.65 0.12 Polish lignite 1.6 - 3.7 KT coal 0.01 - 0.05 83 respectively) than the USC Lignin Standard (3.0 wt% methyl) as expected given that cellulose and hemicellulose major components diluting lignin in wood (Hedges et al., 1985). 4.3.4 Methoxyl quantification in lignin-bearing sedimentary deposits We test a variety of geological substrates in order to assess changes in methyl group concentration during early and late lignin maturation stages in peat, lignite and coal. We analyzed peat from Ireland (USC Irish Peat Standard, supplier: Amazon.com), and lignite samples of Miocene age from Poland (Drobniak and Mastalerz, 2006) and coal samples from the Cretaceous – Paleogene (K–Pg) boundary from Montana, USA. The USC Irish Peat Standard methyl content was found to be 0.6% (Table 4.1) which indicates either the methoxyl composition of bog plants were lower than that of the two wood types measured here and/or dilution with other sedimentary components. We analyzed a series of lignites, from a low species diversity, Miocene age low-maturity deposit in Poland (Drobniak and Mastalerz, 2006). The Polish lignite samples had been categorized into four groups (Supplementary Table 4.1) based on their morphological, petrographic and bulk chemical properties, with increasing maturity from group 1 to 4. Semi-quantitative analysis by Fourier Transform Infra-Red (FTIR) spectroscopy found no clear change in methoxyl group content for samples in groups 1 to 4. Lignin to cellulose ratios, another early diagenesis metric, do however vary and increases with increasing transformation of lignites (Drobniak and Mastalerz, 2006) (Figure 4.3). Using the liquid method, methyl (within methoxyl) content in 36 Polish lignites were quantified with at least 4 samples in each group. We find methyl content increases with increasing lignin to cellulose ratio, i.e. increased degradation loss of cellulose 84 leads to a concentration of lignin and thus methyl group yields (Figure 4.3). The median methyl content increased from 2.5 to 3.5% in groups 1 to 4 (Figure 4.3, Supplementary Table 4.1). Fig. 4.3. The methyl concentrations of Miocene age Polish lignites from Drobniak and Mastalerz, (2006), are analyzed here using the liquid method. Four groups (annotated as grp 1, 2, 3, and 4) are displayed in box plots according to their lignin to cellulose ratio. Coal samples from the Cretaceous – Paleogene (K – Pg) boundary at the Hell Creek Formation in Montana were analyzed for methyl (within methoxyl) content. Three samples collected at 50, 60 and 70 cm below the boundary were black in color, showing extensive coalification. The methyl concentration ranged from 0.01 to 0.05 wt% (Table 4.1). Due to the very low methyl content, large amounts of coal (up to 100 mg) were reacted for measurement and analytical error including replicate vials was 0.01 wt% (1). 85 4.3.5 Lignin phenol concentrations and ratios of the USC Standards Lignin phenol characteristics were analyzed for the four USC standards (Table 4.1) which include kraft lignin, two angiosperm plants (grass and wood) and boreal peat that contain varying abundance of methoxyl groups. Concentrations of oxidative phenols (V for vanillyl, S for syringyl, C for coniferyl and P for p-hydroxyl) and ratios between them display plant characteristics of the USC standards. USC Lignin standard is a kraft lignin and is mainly composed of vanillyl phenols and has a very low S/V ratios suggesting a gymnosperm origin (Table 4.1) (Hedges and Mann, 1979). In contrast, USC Bamboo and Poplar standards have high S/V ratios because they are both angiosperms. USC Bamboo standards has a high C/V due to structural differences in grass and wood type plants. USC Irish Peat standard also has a relatively high C/V ratio and the highest P/V ratio among the USC standards because boreal peats are composed of mosses with high relative concentrations of P and C phenols to V and S phenols (Weng and Chapple, 2010). Phenol monomers in each group were quantified (Supplemental Table 4.2) and methoxyl content of the four USC standards were estimated based on the stoichiometry of the phenols. USC Bamboo, Poplar, Lignin and Irish Peat standards contained 0.6, 0.5, 0.2 and 0.1% of methyl content (within methoxyl groups) (Supplemental Table 4.2). In comparison, the liquid method quantification yielded 1.0, 1.9, 2.2, 0.6% respectively which is more than 30% of the phenol estimates. The estimate of methyl content based on phenol concentrations is lower because the CuO oxidation method does not completely depolymerize lignin, leading to fractional yields of phenols from bulk lignin (~30%) (Ma et al., 2018). 86 4.3.6 Hydrogen isotopic composition of the USC Standards Here we compare the headspace and liquid methods of isolation and introduction of iodomethane for gas chromatography. The two methods have been shown previously to differ in yields, with headspace yields being lower associated with condensation onto vial or syringe surfaces. This comparison of methods allows us to further test the effect of iodomethane condensation on the hydrogen isotopic composition. We find that the headspace method yield δD values that are 1 – 15 ‰ more D-enriched than values from the liquid method (Table 4.2). Split flow setting for headspace injection could make the results more positive but the magnitude of that enrichment was variable while the split ratio stayed the same. Uncertainties vary between hydrogen isotopic measurements using the headspace and liquid approaches. We categorize two levels of uncertainties, one being solely instrumental between triplicate analyses of one sample vial, and the other representing uncertainties in the sample preparation process by analyzing three replicate vials of one sample. The uncertainties in the instrument triplicate were 2.5‰ and 2.7‰ for headspace and liquid method and in vial replicates, 3.3‰ and 2.7‰, respectively. When combined, the overall uncertainty was 4.3‰ for headspace and 3.1‰ for liquid method (Table 4.2). Table 4.2. Hydrogen isotope results of the standard suite in both headspace and liquid method and carbon isotope results in headspace method via GC – IRMS. Instrument  represents an averaged uncertainty of triplicate measurements of one sample vial whereas chemistry  represents the standard deviation between several replicate vials. The combined  was determined by the square root of the sum of the squares of the instrument and chemistry errors. Hydrogen isotopic composition Carbon isotopic composition Headspace method Liquid method Headspace method Standards D instrument  n chemistry  n combined  D instrument  n chemistry  n combined   13 C instrument  n chemistry  n combined  USC Lignin standard -254 3.2 12 3.0 4 4.4 -255 1.5 21 2.9 7 3.3 -21.0 0.26 33 0.63 9 0.68 USC Bamboo standard -173 2.4 9 3.1 3 3.9 -187 1.0 9 1.1 3 1.5 -26.8 0.09 11 0.47 3 0.48 USC Poplar standard -315 1.0 8 1.4 3 1.7 -330 1.7 9 1.6 3 2.3 -26.8 0.10 9 0.11 3 0.14 USC Irish Peat standard -265 1.9 7 2.2 3 2.9 -27.0 0.39 9 0.65 3 0.76 Vanillin -21 3.7 9 4.5 3 5.8 -35 1.3 9 3.4 3 3.6 -37.4 0.38 9 0.51 3 0.64 Vanillic acid 202 1.8 9 2.0 3 2.7 180 1.4 9 3.8 3 4.0 -24.8 0.46 8 0.52 3 0.70 Syringealdehyde -171 2.6 9 3.4 3 4.3 -180 1.3 9 3.9 3 4.1 -40.9 0.11 10 0.41 3 0.43 Syringic acid -167 2.2 9 2.4 3 3.3 -172 1.0 9 2.0 3 2.2 -27.5 0.51 8 0.62 3 0.80 Iodomethane -82 3.4 25 6.8 7 7.6 -96 1.6 47 3.2 15 3.6 -54.6 0.17 31 0.49 9 0.52 mean 2.5 3.3 4.2 1.4 2.7 3.1 -31.9 0.49 0.57 87 The hydrogen isotopic composition of the suite of substrates spanned –255 to +180 ‰ (Figure 4.4). The D-enriched values for vanillic acid and vanillin deviate from values common in nature, but could be explained by their synthetic origin (Fischer, 2007) and indicate a potential for food testing and forensic applications. In natural products, we observe a 150 ‰ difference between the USC Bamboo and USC Poplar wood standards. Although the geographic origins are unknown, there may be both source water δD value and biosynthetic fractionation differences between these two groups. The global mean fractionation ( 2  wax/precip ) for tree wood δD methoxyl relative to precipitation δD has been reported to be -216 ± 19‰ (Keppler et al., 2007) but there may be biosynthetic differences between species or plant types. In particular, we note that the hydrogen isotopic fractionation between source water and methoxyl groups within the graminoids remains untested. Fig. 4.4. Cross plot of hydrogen and carbon isotopic composition of iodomethane (black square), and the methoxyl groups released from phenolic compounds (purple triangles), wood and peat (orange circles). These results demonstrate the isotopic range of natural and synthetic products tested here. All hydrogen isotopic compositions were measured using the liquid method, all carbon isotopic compositions were measured using the headspace method. 88 We were unable to measure the δD values of methoxyl groups within the USC Irish Peat by the headspace method due to low concentrations. However, we were able to analyze this substrate via the liquid method, by reacting 100 mg of the USC Irish Peat in order to collect the iodomethane analyte at sufficient concentration for analyses. However, replicate uncertainties are large, which may reflect analytical challenges such as associated with reaction processes at low concentrations, i.e. this may be a matrix-interaction or heterogeneity effect. Given these observations we used the liquid method for hydrogen isotopic analyses of methoxyl groups in geological (potentially low methoxyl concentration) substrates e.g. lignite and coal. 4.3.7 Carbon isotopic composition of the USC Standards Carbon isotopic compositions were only analyzed using the headspace method because condensation and evaporation of hydrocarbons are known to not have a large carbon isotope effect (Whittaker et al., 1995; Harrington et al., 1999). For example, in monoaromatic hydrocarbons, phase change led to an isotopic fractionation of about 0.2 ‰ which is smaller than the analytical error for online GC–IRMS measurements (Harrington et al., 1999). Using the headspace method we found the carbon isotopic composition of the USC standards range from – 21.0 to –54.6 ‰ (Figure 4.4) establishing a broad range for comparison to natural samples. The chemical replicate errors were bigger than instrumental uncertainty for all standards, with combined uncertainty averaging 0.57 ‰. δ 13 C methoxyl of all of the wood standards were close to that expected for C 3 plants (Cerling et al., 1997). Variability in δ 13 C methoxyl -21.4 to -54.6 was observed among the phenol standards as these are likely synthetic rather than plant-derived. The iodomethane standard had the most 13 C-depleted value (-54.6 ‰; Table 4.2). 89 4.4 Discussion 4.4.1 Recommended practices for quantification of iodomethane We recommend using the liquid method for accurate and precise quantification of iodomethane using GC–MS/FID. The liquid method linearly corresponds concentration to peak area and the results are consistent with varying injection volumes (Figure 4.1d), whereas headspace injections suffered losses assumed to be related to condensation on the vial or syringe surfaces. Incomplete transfer is a concern both for quantification and isotopic analyses. Furthermore, when quantifying methoxyl groups in natural samples prepared via the Ziesel process, water is present within the vial, into which iodomethane will partition (Ogg, 1938). One solution to this problem, established here, is to partition the iodomethane product into isooctane for analyses. Another reported solution is to use a headspace autosampler with appropriate vial conditions and sufficient equilibration time (Li et al., 2012). Loss of iodomethane to the aqueous phase or to condensation onto the vial or syringe, is shown here to affect hydrogen isotopic analyses. The lighter isotope preferentially condenses at ambient room temperature conditions due to the inverse vapor pressure isotope effect in most hydrocarbons and halogenated hydrocarbons (Höpfner, 1969), thus the headspace iodomethane would be D-enriched. This is opposite to the familiar effect in water, whereby condensation favors the heavier isotope (Craig et al., 1963). Another potential source of D-enrichment is associated with the use of split flow at the GC inlet as commonly used with headspace approaches. In order to achieve a narrow peak shape with gas injections, a split ratio of 25 was set for all headspace injections here, similar to that reported in prior work (Greule et al., 2008; Feakins et al., 2013b; Anhäuser et al., 2018). D-diffusion is 90 slower and thus fractionation may occur if using split flow with small analytes (e.g., Lee et al., 2017). For the liquid method, the samples are injected in splitless mode, without any compromise to peak shape, which is already narrow, thus there are no concerns about fractionation within the instrument. 4.4.2 Lignin methoxyl δD and δ 13 C analyses Our tests show that condensation can affect the hydrogen isotopic composition of iodomethane measured by headspace GC approaches. The extent of condensation can vary depending on the concentration inside the vial, injection volume, amount of water in the vial, and we observed up to 15‰ positive shift in the hydrogen isotopic composition with the headspace method compared to the liquid method. Moreover, the liquid method showed higher precision for both instrumental and replicate sample vials compared to the headspace approach and allowed lower methoxyl concentration samples to be analyzed. For δ 13 C however, condensation during headspace method should only have a small effect on the isotopic composition; smaller than the analytical uncertainties. We have characterized wood, lignin and phenolic compounds yielding hydrogen and carbon isotopic compositions with a range of over 500‰ and 40‰ respectively. Phenols and wood offer advantages as isotope standards are advantageous over the iodomethane standard for the following reasons. First, the analysis of solid standards requires the same preparative Ziesel step that samples require, meaning like treatment for sample and standard. Second, storage and transport are easier with solid compounds than the volatile iodomethane standard which may evaporate with time and needs to be handled carefully for inter-laboratory comparisons. Compounds measured here have δD methoxyl values that range over 500‰ and encompass 91 published lignin methoxyl values (Keppler et al., 2007; Feakins et al., 2013a; Anhäuser et al., 2015, 2017b, 2017a, 2018; Riechelmann et al., 2017) except a few samples that come from the Arctic circle, Yemen and tropical saltwater mangroves. A single iodomethane standard is available, analyzed with offline gas chromatography isotope ratio mass spectrometry against the VSMOW/SLAP and LSVEC/VPDB scale, otherwise broader isotopic calibration must be achieved by different compounds such as an alkane mixture (all supplied by A. Schimmelmann Indiana University). Here we report a collection of USC Standards that can define a range of lignin methoxyl hydrogen and carbon isotopic values that constrain the range of values likely in a variety of natural product and synthetic products from the food and paper industries as well as constraining a range of living and sedimentary plant derived lignin materials. 4.4.3 Fate of methoxyl during early diagenesis and maturation Since carbohydrate is more degradable than lignin (Hedges et al., 1985), the lignin percentage increases during early stages of decomposition and thus we might expect an increase in methoxyl concentration from living or recently fallen wood to decomposing wood and soil or peat accumulations under woody forests. This effect is indirectly shown in the higher methoxyl content in the USC Lignin standard compared to the USC Bamboo and USC Poplar standards. USC Lignin standard does not contain any carbohydrates such as cellulose and hemicellulose, representing wood composition during early diagenesis. Likewise in the Polish lignites, although these samples are of Miocene age, we see a slight increase in the methyl content due to the increase in lignin to cellulose ratio (Drobniak and Mastalerz, 2006). Even much older woody samples given low transformation, can be analyzed 92 for lignin methoxyl quantification and isotopic composition (Anhäuser et al., 2018) due to the high resistance of lignin to degradation. As diagenesis proceeds, wood transforms and the lignin structure changes. Characteristics include a reduction in methoxyl groups and phenolic carbon (Hatcher et al., 1981) and an increase in the acid to aldehyde ratio within cupric oxide oxidation products (Goñi et al., 2000). Among the USC standards and the Polish lignite and Montana K – Pg coal samples, we see a decrease in methyl content across the maturity gradient as expected. Given the different starting point of woods with different phenolic content, however, the variation in initial methoxyl abundance should be considered in studies of sediments and interpretations in terms of source ecosystem and maturity. One solution could be the combination of classic lignin analysis of lignin phenols and their ratios (Goñi et al., 2000), in tandem with methyl concentration measurements. 4.4.4 Methoxyl concentrations in peats under different ecosystems Boreal peats represent a rich archive of organic matter and lignin of global importance (carbon pool of 455 Pg), with distributions that include high latitudes (> 35°N and S) and altitude, water logged locations (Gorham, 1991). High latitude and altitude, boreal peats are dominantly formed from the accumulation of sphagnum moss (Tsutsuki and Kondo, 1995). While peats should have high lignin to cellulose ratios (Hedges et al., 1985), moss lignin is mostly composed of p- hydroxyphenyl monomers with no methoxyl groups (Wang et al., 2010), with a more minor component (> 30 %) of the lignin polymer comprised of vanillin with each monomer yielding 1 methoxyl group. The USC Irish Peat standard yielded low methyl concentrations (0.6 wt%), presumably reflecting the monomeric composition of sphagnum, which is the likely source of the 93 peat, as in much of Ireland (Hammond, 1981). Following our protocols, such low concentrations of methoxyl present analytical challenges for sufficient yield for isotopic analyses, without further methodological adaptations, and the p-hydroxyphenyl dominance means this approach may be inherently limited in boreal peats. While we find lignin methoxyl approaches have limited potential in sphagnum – moss peats, these approaches have better potential in peats formed under woody vegetation, such as the swamp forests of tropical climates today (Lähteenoja et al., 2013) and in the geological past (Naafs et al., 2018). Angiosperm forests are expected to be particularly promising given that the syringyl phenols with 2 methoxyl groups, are abundant in angiosperms whereas the vanillyl phenols dominating in gymnosperms contain just one methoxyl group (Weng and Chapple, 2010). Thus we expect more methoxyl in angiosperm lignites than in those from the older gymnosperm forests of the pre-angiosperm world. 4.5 Conclusions We combine quantification and isotopic analyses of lignin methoxyl groups in various substrates (iodomethane, lignin monomers, pure lignin, wood, peat, lignite and coal) to optimize analytical methods towards biogeochemical and geological applications. Quantification of methoxyl concentration is of interest for composition and preservation analyses and for assuring quantitative recovery sufficient for isotopic determinations. We compare two methods for iodomethane quantification that reveal significant losses can occur due to condensation, and that these can be resolved by analysis of iodomethane in isooctane. Condensation losses result in underestimates of iodomethane concentrations and are associated with D-depletion of the 94 residual, consistent with theory. Techniques that avoid condensation losses are thus important for both quantification and δD analyses, whereas this issue is not a concern for δ 13 C analyses. We emphasize that the volatility of iodomethane presents challenges for storage as an isotopic standard for calibration purposes. Solid standards of lignin phenols have known stoichiometry, and thus provide advantages both for secure quantification and for isotopic calibration. Based on the presence of methoxyl on different monomeric lignin units, quantification of methoxyl has potential to questions of different lignin composition within plant physiology (e.g. compression versus extension wood) and plant biogeochemistry, wood and food forensics, and other applications to studying lignin in plants and immature sediments containing lignin in the environment. We find a decrease in methoxyl concentration associated with maturation from lignite to coal, such that this approach may provide a maturation indicator to complement classic indicators (e.g. lignin/cellulose, vitrinite reflectance). The isotopic determinations of methoxyl groups have scope for application to tree wood, and their sedimentary archives e.g. buried logs, peats and lignites, especially those derived from woody species (gymnosperms and angiosperms). However boreal peats derived from sphagnum and sapropelic lignites are not promising targets given the low methoxyl concentrations in source organisms. 95 Acknowledgements We thank Maria Mastalerz and Nathan Carroll for Polish lignite and Montana coal samples. Camilo Ponton, Travis Williams, Taylor Graham, Max Lloyd, Arndt Schimmelmann, and Marcus Greule are thanked for helpful discussions. This study was supported by the American Chemical Society Petroleum Research Fund (US) granted to SF (53747-ND2), USC Provost fellowship and USC Earth Science department graduate student research fund to HL. 96 References Adler, E., 1977. Lignin chemistry-past, present and future. Wood Science and Technology 11, 169–218. Anhäuser, T., Greule, M., Keppler, F., 2017a. Stable hydrogen isotope values of lignin methoxyl groups of four tree species across Germany and their implication for temperature reconstruction. Science of the Total Environment 579, 263–271. Anhäuser, T., Greule, M., Polag, D., Bowen, G.J., Keppler, F., 2017b. Mean annual temperatures of mid-latitude regions derived from δ 2 H values of wood lignin methoxyl groups and its implications for paleoclimate studies. Science of the Total Environment 574, 1276–1282. Anhäuser, T., Greule, M., Zech, M., Kalbitz, K., McRoberts, C., Keppler, F., 2015. Stable hydrogen and carbon isotope ratios of methoxyl groups during plant litter degradation. Isotopes in Environmental and Health Studies 51, 143–154. Anhäuser, T., Hook, B.A., Halfar, J., Greule, M., Keppler, F., 2018. Earliest Eocene cold period and polar amplification - Insights from δ 2 H values of lignin methoxyl groups of mummified wood. Palaeogeography, Palaeoclimatology, Palaeoecology 505, 326–336. Anhäuser, T., Sirocko, F., Greule, M., Esper, J., Keppler, F., 2014. D/H ratios of methoxyl groups of the sedimentary organic matter of Lake Holzmaar (Eifel, Germany): A potential palaeoclimate/-hydrology proxy. Geochimica et Cosmochimica Acta 142, 39–52. Baker, S.M., 1996. Rapid methoxyl analysis of lignins using gas chromatography. Holzforschung 50, 573–574. Boerjan, W., Ralph, J., Baucher, M., 2003. LIGNIN BIOSYNTHESIS. Annual Review of Plant Biology 54, 519–546. Cerling, T.E., Harris, J.M., MacFadden, B.J., Leakey, M.G., Quade, J., Eisenmann, V., Ehieringer, J.R., 1997. Global vegetation change through the Miocene/Pliocene boundary. Nature 389, 153–158. Craig, H., Gordon, L.I., Horibe, Y., 1963. Isotopic exchange effects in the evaporation of water: 1. Low-temperature experimental results. Journal of Geophysical Research 68, 5079–5087. DeNiro, M.J., Epstein, S., 1981. Isotopic composition of cellulose from aquatic organisms. Geochimica et Cosmochimica Acta 45, 1885–1894. Drobniak, A., Mastalerz, M., 2006. Chemical evolution of Miocene wood: Example from the Belchatow brown coal deposit, central Poland. International Journal of Coal Geology 66, 157– 178. Feakins, S.J., Ellsworth, P. V., Sternberg, L. da S.L., 2013a. Lignin methoxyl hydrogen isotope ratios in a coastal ecosystem. Geochimica et Cosmochimica Acta 121, 54–66. 97 Feakins, S.J., Rincon, M., Pinedo, P., 2013b. Analytical challenges in the quantitative determination of 2 H/ 1 H ratios of methyl iodide. Rapid Communications in Mass Spectrometry 27, 430–436. Fischer, U., 2007. Flavours and Fragrances: Chemistry, Bioprocessing and Sustainability, Flavours and Fragrances: Chemistry, Bioprocessing and Sustainability. Springer. doi:10.1007/978-3-540-49339-6_11 Goñi, M.A., Thomas, K.A., Goni, M.A., 2000. Sources and Transformations of Organic Matter in Surface Soils and Sediments from a Tidal Estuary (North Inlet, South Carolina, USA). Estuaries 23, 548. Gorham, E., 1991. Northern peatlands: role in the carbon cycle and probable responses to climatic warming. Ecological Applications 1, 182–195. Gori, Y., Wehrens, R., La Porta, N., Camin, F., 2015. Oxygen and hydrogen stable isotope ratios of bulk needles reveal the geographic origin of Norway spruce in the European Alps. PLoS ONE 10, 1–14. Greule, M., Mosandl, A., Hamilton, J.T.G., Keppler, F., 2008. A rapid and precise method for determination of D/H ratios of plant methoxyl groups. Rapid Communications in Mass Spectrometry 22, 3983–3988. Greule, M., Mosandl, A., Hamilton, J.T.G., Keppler, F., 2009. A simple rapid method to precisely determine 13C/12C ratios of plant methoxyl groups. Rapid Communications in Mass Spectrometry 23, 1710–1714. Greule, M., Tumino, L.D., Kronewald, T., Hener, U., Schleucher, J., Mosandl, A., Keppler, F., 2010. Improved rapid authentication of vanillin using δ13C and δ2H values. European Food Research and Technology 231, 933–941. Hadžija, O., Tonković, M., 1975. Gravimetric microdetermination of alkoxy groups in sulphur containing compounds. Mikrochimica Acta 64, 213–217. Hammond, R.F., 1981. The peatlands of Ireland. An Foras Taluntais 1–63. Hansen, A.M.S., Fromberg, A., Frandsen, H.L., 2014. Authenticity and traceability of vanilla flavors by analysis of stable isotopes of carbon and hydrogen. Journal of Agricultural and Food Chemistry 62, 10326–10331. Harrington, R.R., Poulson, S.R., Drever, J.I., Colberg, P.J.S., Kelly, E.F., 1999. Carbon isotope systematics of monoaromatic hydrocarbons: Vaporization and adsorption experiments. Organic Geochemistry 30, 765–775. Hatcher, P.G., Breger, I.A., Earl, W.L., 1981. Nuclear magnetic resonance studies of ancient buried wood-I. Observations on the origin of coal to the brown coal stage. Organic Geochemistry 3, 49–55. Hedges, J.I., Cowie, G.L., Ertel, J.R., James Barbour, R., Hatcher, P.G., 1985. Degradation of carbohydrates and lignins in buried woods. Geochimica et Cosmochimica Acta 49, 701–711. 98 Hedges, J.I., Mann, D.C., 1979. The lignin geochemistry of marine sediments from the southern Washington coast. Geochimica et Cosmochimica Acta 43, 1809–1818. Höpfner, A., 1969. Vapor Pressure Isotope Effects. Angewandte Chemie International Edition in English 8, 689–699. Hwanc, S., 1992. The Chemistry of, Chemistry is Life. Academic Press Inc., New York. doi:10.1002/0470864028 Keppler, F., Hamilton, J., 2008. Tracing the geographical origin of early potato tubers using stable hydrogen isotope ratios of methoxyl groups. Isotopes in Environmental and Health Studies 44, 337–347. Keppler, F., Harper, D.B., Kalin, R.M., Meier-Augenstein, W., Farmer, N., Davis, S., Schmidt, H.L., Brown, D.M., Hamilton, J.T.G., 2007. Stable hydrogen isotope ratios of lignin methoxyl groups as a paleoclimate proxy and constraint of the geographical origin of wood. New Phytologist 176, 600–609. Konopleva, N.G., Ivanyuk, G.Y., Pakhomovsky, Y.A., Yakovenchuk, V.N., Mikhailova, Y.A., 2017. Loparite-(Ce) from the Khibiny Alkaline Pluton, Kola Peninsula, Russia. Geology of Ore Deposits 59, 729–737. Lähteenoja, O., Flores, B., Nelson, B., 2013. Tropical peat accumulation in Central Amazonia. Wetlands 33, 495–503. Lee, H., Feakins, S.J., Lu, Z., Schimmelmann, A., Sessions, A.L., Tierney, J.E., Williams, T.J., 2017. Comparison of three methods for the methylation of aliphatic and aromatic compounds. Rapid Communications in Mass Spectrometry 31, 1633–1640. Li, H., Chai, X.S., Zhan, H., Liu, M., Fu, S., 2012. A Novel Method for Determination of Ethoxyl Content in Ethyl Cellulose by Headspace Gas Chromatography. Analytical Letters 45, 1028–1035. Ma, T., Zhu, S., Wang, Z., Chen, D., Dai, G., Feng, B., Su, X., Hu, H., Li, K., Han, W., Liang, C., Bai, Y., Feng, X., 2018. Divergent accumulation of microbial necromass and plant lignin components in grassland soils. Nature Communications 9, 3480. Mischel, M., Esper, J., Keppler, F., Greule, M., Werner, W., 2015. δ 2 H, δ 13 C and δ 18 O from whole wood, α-cellulose and lignin methoxyl groups in Pinus sylvestris: a multi-parameter approach. Isotopes in Environmental and Health Studies 51, 553–568. Ogg, R.A., 1938. The Hydrolysis of Methyl Iodide. Journal of the American Chemical Society 60, 2000–2001. Philp, R.P., 1985. Petroleum Formation and Occurrence, Eos, Transactions American Geophysical Union. Springer-Verlag, New York. doi:10.1029/EO066i037p00643 Riechelmann, D.F.C., Greule, M., Siegwolf, R.T.W., Anhäuser, T., Esper, J., Keppler, F., 2017. Warm season precipitation signal in δ 2 H values of wood lignin methoxyl groups from high 99 elevation larch trees in Switzerland. Rapid Communications in Mass Spectrometry 31, 1589– 1598. Riechelmann, D.F.C., Greule, M., Treydte, K., Esper, J., Keppler, F., 2016. Climate signals in δ13C of wood lignin methoxyl groups from high-elevation larch trees. Palaeogeography, Palaeoclimatology, Palaeoecology 445, 60–71. Schimmelmann, A., Sessions, A.L., Mastalerz, M., 2006. Hydrogen Isotopic (D/H) Composition of Organic Matter During Diagenesis and Thermal Maturation. Annual Review of Earth and Planetary Sciences 34, 501–533. Stephens, T.W., Wilson, A., Dabadge, N., Tian, A., Zimmerman, M., Hensley, H.J., Acree, W.E.J., Abraham, M.H., 2012. Correlation of solute partitioning into isooctane from water and from the gas phase based on updated Abraham equations. Global Journal of Physical Chemistry 3, 1–16. Sumerskii, I., Zweckmair, T., Hettegger, H., Zinovyev, G., Bacher, M., Rosenau, T., Potthast, A., 2017. A fast track for the accurate determination of methoxyl and ethoxyl groups in lignin. RSC Advances 7, 22974–22982. Thevenot, M., Dignac, M.F., Rumpel, C., 2010. Fate of lignins in soils: A review. Soil Biology and Biochemistry 42, 1200–1211. Tsutsuki, K., Kondo, R., 1995. Lignin-derived phenolic compounds in different types of peat profiles in Hokkaido, Japan. Soil Science and Plant Nutrition 41, 515–527. Vanholme, R., Demedts, B., Morreel, K., Ralph, J., Boerjan, W., 2010. Lignin Biosynthesis and Structure. Plant Physiology 153, 895–905. Vieböck, F., Schwappach, A., 1930. Eine neue Methode zur maßanalytischen Bestimmung der Methoxyl-und Ä thoxylgruppe. Berichte der deutschen chemischen Gesellschaft (A and B Series) 63, 2818–2823. Weng, J.K., Chapple, C., 2010. The origin and evolution of lignin biosynthesis. New Phytologist 187, 273–285. Whittaker, M., Pollard, S.J.T., Fallick, T.O.E., 1995. Characterisation of refractory wastes at heavy oil-contaminated sites: A review of conventional and novel analytical methods. Environmental Technology (United Kingdom) 16, 1009–1033. Zeisel, S., 1885. Ü ber ein Verfahren zum quantitativen Nachweise von Methoxyl. Monatshefte für Chemie 6, 989–997. 100 Chapter 5 Sustained wood burial in the Bengal Fan over the last 19 million years Hyejung Lee a , Valier Galy b , Xiaojuan Feng c , Camilo Ponton a,d , Albert Galy e , Christian France- Lanord e , Sarah J. Feakins a a Department of Earth Sciences, University of Southern California, Los Angeles, California, USA; b Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA; c State Key Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China; d Department of Geology, Western Washington University, Bellingham, Washington, USA; e Centre de Recherches Pétrographiques et Géochimiques, Université de Lorraine – CNRS, Vandoeuvre-lès-Nancy, France; Abstract The Ganges – Brahmaputra River system transports over a billion tons of sediment every year from the Himalayan Mountains to the Bay of Bengal and has built the world‘s largest active sedimentary deposit, the Bengal Fan. High sedimentation rates drive exceptional organic matter preservation that represents a long-term sink for atmospheric CO2. While much attention has been paid to organic-rich fine sediments, coarse sediments have generally been overlooked as a locus of organic carbon burial. However, IODP Expedition 354 recently discovered abundant woody debris (mm to cm size fragments) preserved within the coarse sediment layers of 101 turbiditic sequences recovered from 6 marine drill sites along a transect through the Bengal fan (~8°N, ~3700 m water depth) with recovery spanning 19 million years (Myr). Analysis of bulk wood and lignin reveals sustained delivery and preservation of woody debris dominantly sourced from the lowland portion of the catchment, with only episodic – albeit spectacular - occurrence of wood source from high elevations. Woody debris include C 4 plants after their expansion in the late Miocene and wood composition and sourcing becomes more variable in the Pleistocene. Estimates of wood burial flux would require further accounting throughout turbidic units and across the fan, but this targeted study of woody debris shows sustained wood transport and preservation throughout the last 19 Myr - a delivery and burial mechanism generally neglected in carbon budgets. 5.1 Introduction The burial of terrestrial biospheric organic carbon (OC) in marine sediments represents a net transfer of carbon from the atmosphere to the sedimentary reservoir where it may be stored over geologic timescales, thereby decreasing CO 2 and increasing O 2 levels in the atmosphere (Derry and France-Lanord, 1997; Hayes and Waldbauer, 2006). Fine-grained sediments enriched in phyllosilicates with high surface area are characterized by high OC loading and can shield OC from degradation through mineral interactions (Mayer, 1994). As such, the supply and accumulation of clay minerals has been considered to be key to the preservation of both marine and terrestrial OC in marine sediments (Hedges and Keil, 1995; Kennedy and Wagner, 2011) and coarse sediments and OC particles have received little attention (Sparkes et al., 2015). 102 Coarse plant debris has recently been reported as a significant component of the OC load in small mountainous rivers characterized by high erosion rate and short sediment transit time (West et al., 2011; Turowski et al., 2016; Kramer et al., 2017). If coarse wood is efficiently supplied to and preserved in marine sediments, as anecdotal evidence suggests (Saller et al., 2006; Sparkes et al., 2015), the burial of coarse OC particles may thus play an important – yet overlooked – role in the sequestration of atmospheric CO 2 . Here, we investigate the sourcing, transport and burial of coarse woody debris in the largest active sedimentary system on Earth, the Bengal Fan. International Ocean Discovery Program Expedition 354 drilled seven sites on a W – E transect at 8º N in the middle Bengal Fan, ca. 2000 km away from the mouth of the rivers that currently supply sediments to the Bengal Fan (Fig. A.1). Three deep cores recovered Neogene and Quaternary sediments and together provide a record of sedimentation history over the last 19 Myr (France-Lanord et al., 2016) (Fig. A.2). Wood was visually identified and removed from split cores (Fig. 5.1) onboard the R/V Joides Resolution and further recovery was achieved in the lab by sieving mm to cm-scale wood from the sediment matrix (as detailed in the Appendix). 103 Fig. 5.1. Five examples of wood deposits in IODP Exp 354 cores from Sites U1451 and U1455. Buried wood sizes vary from mm to cm scale and are denoted by black fragments on core images (IODP LIMS Reports). Samples with readily visible wood pieces such as in (A) U1451B-23R-1-47-49 cm, (B) U1455C-44R-3-85-87 cm, and (D) U1451A-56F-2-4-7 cm were picked directly from the face of the split cores. Other layers with visible specks thought to be wood, were sampled and sieved (1mm) to isolate fragments, including samples of (C) U1451B-19R-3-36-40 cm and (E) U1451B-58R-1-50-56 cm. All rulers show cm. We analyzed the molecular and isotopic composition of the woody debris (as detailed in the Appendix) to reveal information about sourcing within the catchment based on environmental gradients (e.g., elevation, C 3 /C 4 vegetation and precipitation). We further explore the 104 implications of sourcing and composition to reveal transport and preservation/degradation mechanisms of this overlooked component of fluvial export of terrestrial carbon. Given the size of the G – B catchment (1.6  10 6 km 2 ) and the Bengal Fan (3  10 6 km 2 ), and the long distance of sediment transport (>2000 km offshore), the frequent presence of wood fragments is remarkable. Although clay minerals have been considered to be the predominant control on the preservation of terrestrial OC in marine sediments, evidence for wood burial in coarse turbidic sediments, means similar units should not be discounted in OC accounting in similar marginal settings. 5.2 Results 5.2.1 Abundance of wood in the Bengal Fan Turbiditic units dominate the sedimentary accumulation of the Bengal Fan and mm to cm size wood debris was picked from the coarse-grained layers of turbiditic sequences (especially sub- units Tb and Td of the Bouma sequence (Bouma, 1962) during visual description of the split cores onboard the R/V Joides Resolution. We found wood at each of the drill sites that we studied (U1450, U1451, U1452, U1453, U1454, U1455) (Figs. 5.1 and 5.2A; Figs. A.2 and A.3). Wood was mainly found as i) isolated large fragments within thick sandy deposits, ii) small fragments dispersed throughout the coarse basal units of turbidite layers and, iii) (sub)horizontal layers of fragments concentrated at the transition between coarse and fine sediments, i.e. the subunit Td of the Bouma sequence (Fig 5.1). The wood fragments varied in color and size, but were most commonly grey or black in color, in thin sheets (likely as a result of compaction and dewatering) or small amorphous fragments. Bigger pieces (up to 3 cm  2 cm) were mostly brown to black, 105 but some layers yielded abundant and apparently unaltered (pale colored) wood fragments. Five coarse woody horizons notably yielded >1 – 5 g of wood (out of ~50 g of sediment or picked from the cut face of the core), distributed across the Miocene and Pliocene (Fig. 5.2B). Fifty- eight visible woody horizons yielded lower amounts of wood. In addition, we sieved (>1 mm) and recovered finer wood debris from 37 out of 73 turbiditic basal units studied, recovering 0.1 – 2.5 mg g -1 (Fig A.2, Table A.6). In total, we recovered 95 wood samples spanning the last 19 Myr (Fig. 5.2C). The age-depth distribution (Fig. 5.2D) reveals the episodic deposition of woody horizons and overlapping sediment recovery provides corroboration (Fig. 5.2D). Fig. 5.2. Occurrence of wood in IODP Expedition 354 Sites U1450, U1451, U1452, U1453, U1454, U1455 in the Bengal Fan. (A) Total wood mass per sample from two datasets on log scale: those wood fragments picked from shipboard visual inspection of split core, and fragments sieved and picked from 10 – 80 g of sediments post-expedition (symbol color denotes Site – see legend). (B) Sum of wood mass in 1-Myr bins (blue bars); a wood-rich sample from a core catcher within the 0-1 Ma interval estimated > 5g. (C) The number of samples studied (white bars), of which containing wood (cyan bars) displayed in 1- Myr bins. (D) The age–depth distribution of sampled wood fragments (colors, as in A) and summary illustration of drill site temporal coverage, showing approximate age ranges of the cores sampled in colored bars to the right. Note core recovery is discontinuous and sedimentation rates vary (Fig A.2). Age estimates for sample depths are provided in Table A.6 with explanation in the Appendix. 106 5.2.2 Lignin evidence for plant type and degradation Lignin, a large heterogeneous aromatic polymer and the second most abundant plant biochemical (Hedges et al., 1985), carries information about wood type and degradation (Hedges and Mann, 1979). A subset of wood horizons (n = 15) were analyzed for lignin phenol composition with multiple samples possible from one horizon (n = 9). Results for syringyl (S) to vanillyl (V) phenols indicated that all Neogene wood fragments were derived from angiosperms (S/V > 0.02) (Winterfeld et al., 2015) (Fig. 5.3C). As angiosperms dominate below 2500 m in the G – B catchment, and conifers dominate above 3000 m (Carpenter, 2005), the phenol results constrain the elevation source to the lowlands. Only in the Pleistocene did we find a dominance of gymnosperms (S/V = 0.0) in 6 of 9 aliquots of a single core catcher sample at ~0.05 Ma (Fig. 5.3C, Table A.7). This massive accumulation of wood debris derived from high altitude (>2500 m) represents an unusual deposit in the 19 Myr record. In the Neogene samples (n = 14), we observe progressive degradation from 6 – 18 Ma based on declining concentrations of S, V and p-hydroxyl (P) phenols and increasing acid-to-aldehyde ratios within the S and V groups (Sd/Sl and Vd/Vl, respectively) (Hedges and Mann, 1979) (Fig. A.4). As S phenols are preferentially lost compared to V phenols during wood degradation (Hedges et al., 1985), the directional trend in S/V over the Neogene is ascribed to progressive degradation, instead of a continuous change in relative angiosperms and gymnosperms inputs (Fig. 5.3). However, degradation is not extensive as both Sd/Sl (mean = 0.32, 1 = 0.15, n = 23) and Vd/Vl (mean = 0.37, 1 = 0.15, n = 23) fall largely within the range (<0.4) expected for fresh wood (Winterfeld et al., 2015). 107 Fig. 5.3. Isotopic composition and phenol ratio of wood fragments in the Bengal Fan from the Expedition 354. (A) Syringyl to vanillyl phenol ratios of wood fragments: S/V < 0.2 indicates gymnosperms in 6 of 9 aliquots from one horizon at 0.05 Ma (dark green dot); all remaining samples (n = 14) have S/V > 0.2 interpreted as angiosperms (red shading) and arrow indicates direction of S/V change with progressive degradation. (B) Bulk wood  13 C values (triplicates in each horizon reflect wood heterogeneity, with each sample representing 50-100 g of fragments).  13 C values show a consistent signal from 19 to 6 Ma, with 13 C-enriched values appearing 5 Ma and especially after 2 Ma. (C) Lignin methoxyl D values from 20- 80 mg of wood fragments with error bars representing instrumental error (triplicate analyses). In panels (B) and (C) C 4 samples (n = 18) are highlighted in orange; samples with corresponding phenol measurements from the same sediment horizon (not aliquots of a homogenized sample) are denoted in grey. δD values show no systematic trend over 19 Myr. 5.2.3 Bulk wood carbon concentration and isotopic composition All 97 wood-bearing turbiditic units were sampled in triplicate for organic carbon concentration and carbon isotopic composition. We found wood OC ranging from 0.8 to 81.4% (mean = 33.5%, 1 =14.0%, n = 257, Table A.8). This measurement provides constrains on alteration from fresh wood that is typically 40 – 55% OC (Thomas and Martin, 2012). Low OC (< 40%) implies matrix dilution, presumably due to silicate mineral matrix embedded within wood fragments, or degradation. Elevated OC% (~60%) may relate to the presence of some elemental C, such as derived from charring. Three samples with 80% OC are far outside of the range of living plant 108 tissues and thus likely represent charcoal. We find no systematic trend in wood OC content over the 19 Myr record, further suggesting limited degradation of wood debris following burial. Carbon stable isotopic composition of triplicates were measured to identify variations in source wood altitude, humidity and C 3 vs C 4 photosynthetic pathway. We find no correlation between wood OC content and the carbon isotopic composition (Fig. A.5) which implies that matrix dilution and/or degradation does not systematically affect carbon isotopic signatures. Bulk wood  13 C ( 13 C wood ) values range from -30.4 to -11.4‰ in 97 turbiditic units (each horizon sampled in triplicate, yielding 257 individual measurements spanning the last 19 Myr (Fig. 5.3). Aliquots within a horizon have an average variability of 0.7‰ (1) that represents heterogeneity within a single coarse-grained horizon of the turbiditic deposit, but heterogeneity is not directly comparable between samples that include a variable number of wood fragments. Nevertheless, heterogeneity implies mixing in the processes of erosion, fluvial and sub-marine transport to the edge of the shelf, and later resuspension in turbiditic transport to the mid-fan core location. We find  13 C wood values in discrete turbiditic horizons range between -29.9 and -23.8‰ (mean = - 26.6‰, 1 = 1.2‰, n = 187) from 18.4 to 5.8 Ma. This 6.1‰ range indicates considerable heterogeneity in sourcing associated with differences between species, microhabitat (canopy closure), source region (relative humidity and altitude), but with apparent stationarity from 18.4 to 5.8 Ma (Fig. 5.3). As bulk tree wood is infrequently reported in modern ecosystems, we report  13 C wood from a survey of the woody tissues of trees, bamboo and ferns collected along two elevation transects in Arunachal Pradesh (India) and Nepal within the G – B catchment in 2004 – 2007 (Tables A.11 and A.12) and find modern wood  13 C values range from -23.1 to -33.2‰. We account for the 109 year of sampling, to remove the Suess effect and determine pre-Industrial equivalent values (δ 13 C pre-I ) relative to an atmospheric δ 13 C atm value of -6.4‰ (Friedli et al., 1986). δ 13 C pre-I range from -21.4 to -31.4‰ with an average of -26.8‰, similar to the range we find in the Bengal Fan buried wood samples from 18.4 to 5.8Ma (Table A.12). We find no systematic difference between plant type (gymnosperm, angiosperm tree or shrub, bamboo or fern) but we do observe an overall tendency to 13 C-enrichment with elevation of +0.8‰/km -1 (r 2 = 0.22, p < 0.001, n = 53, Fig. A.7). One plant at 1500 m yielded a δ 13 C value of -12.7‰, indicating use of the C 4 pathway, and was therefore not included in the above C 3 summary. Another plant (-12.4‰) was collected in the lowlands, from the bank of the Ganges, where C 4 are dominant (Table A.11). We also compile modern surveys of bulk tree wood  13 C data, from the literature, similarly corrected to pre-Industrial equivalents, for context (Table A.12). These modern surveys confirm that the buried wood samples fall in the expected range of tree wood sourced from the catchment during most of the Neogene (Fig. A.8). More enriched values are first seen in the Pliocene, where the maximum  13 C wood value is -18.9‰ (at 4.4 Ma), which is 4.9‰ more positive than the most 13 C- enriched sample during the previous 9 Myr. Such a value is unlikely to be pure C 4 and may indicate a mixture of C 3 and C 4 fragments in a sample (Figs. 5.3, 5.4). More 13 C-enriched values during the Pleistocene (orange dots, Fig. 5.3B) fall within the range of C 4 plants (mean = -13.1‰,  = 1.0‰, n = 17). Although there is some variability in δ 13 C atm within the Neogene, we do not attempt correction of the buried wood samples given age model uncertainties and retain the measured values as they carry the accumulated signal of all sources of environmental variability. 110 5.2.4 Lignin methoxyl hydrogen isotopic composition A subset of the wood horizons contained enough material to be analyzed for the hydrogen isotopic composition of methoxyl groups on lignin. In 59 samples spanning 19 Myr in the Bengal Fan, D methoxyl values range from -300 to -150‰ (mean = -253‰, 1 = 30.5‰, n = 59; Fig. 5.3, Table A.9). From 19 – 4 Ma, values cluster between -280 to -230‰ (mean = -258‰, 1 = 22.4‰, n = 39). From 2 to 0 Ma, excluding two outliers, D methoxyl values are similar to the Neogene (mean = -253‰, 1 = 34‰, n = 18). A modestly expanded variance may reflect the contribution of D-enriched precipitation in dry times in the lowlands and D-depleted precipitation from high elevation sources especially in the anomalous gymnosperm-rich deposit. About half of the Pleistocene samples are from one core catcher U1455-5F-CC, which have more negative δD methoxyl (mean = -265‰, 1 = 40‰, n = 9) due to inputs from high elevation conifers, as evidenced by lignin phenol composition. Another caveat associated with the interpretation of the isotopic range in D methoxyl analyses is that each reported value represents averaging of a variable number of heterogeneous fragments (1 to >100 pieces of wood, sufficient to achieve 20 – 80 mg). Samples derived from a small number of fragments, from coarser woody debris, may be biased to the discrete conditions of a few trees (with potential differences in D precipitation, habitat and species-specific fractionation). This is the case for the coarse fragments in U1455-5F-CC (0.05 Ma), where all data derive from individual trees. 111 Fig. 5.4. Comparison of D methoxyl and  13 C wood for paired analyses within discrete same sedimentary horizons.  13 C wood show triplicate variability (linked by a line) with one corresponding D methoxyl measurement, from the same sediment horizon. Only a subset of all  13 C wood analyses are shown here, where corresponding D methoxyl values were available. 5.3 Discussion 5.3.1 Wood burial in the Bengal Fan Despite the length of the Ganges and Brahmaputra rivers and the >1500 km distance from the river mouth to the deposition sites, woody fragments have been transported, deposited and preserved in deep sea sediments of the mid Bengal Fan throughout the last 19 Myr. Given changes in climate, ecosystems and the migration of submarine channels, this constitutes remarkable consistency. Most studies have focused on mineral-bound OC in fine grained layers of the upper turbidites (subunit Te of the Bouma sequence); however, we find wood fragments in the coarse-grained layers. Our sampling of the transect of IODP Exp 354 cores, focused on 112 locating visible wood fragments within the coarse, turbiditic units to establish the abundance and characteristics of such deposits. This discovery indicates the future need for accounting of buried wood in the carbon budget (Table A.8) by surveying C throughout turbiditic deposits in efforts to quantify the contribution of wood to the sequestration of atmospheric CO 2 for long-term (10 7 -yr) burial. The drill sites sit along a 300 km transect orthogonal to the main direction of sediment delivery in efforts to capture the migrating channels supplying turbiditic sediments to the fan (France- Lanord et al., 2016). However, we do not yet have a view of transport of wood from the subaerial delta to the fan‘s terminal lobes (across 3  10 6 km 2 ), as would be needed to obtain a comprehensive budget of the overall burial of woody debris. Within the coarse turbiditic horizons, wood OC is generally 0.15 ± 0.43% (1 n=34) of sediment by mass (Table A.8). While low, this abundance is still significant compared to the OC content (typically ca. 0.05%) of wood-free coarse sediments from the modern Ganges-Brahmaputra river system (Galy et al., 2008). Furthermore, OC in coarse river sediments has been shown to be almost entirely composed of petrogenic C (Galy et al., 2008), in stark contrast to the observation of frequent wood debris (i.e. biospheric C) accumulation in Bengal Fan sediments. Coarse sediment horizons are not normally considered in the OC burial estimates which have focused on the fine-grained turbiditic sediments (Derry and France-Lanord, 1997). Nonetheless, the results of IODP expedition 354 show that coarse sediments represent a large proportion of the overall sediment deposition in the distal Bengal Fan (France-Lanord et al., 2016). Combined with our observations of frequent wood debris preservation in coarse sediments, this indicates the contribution of wood debris to the overall burial of biospheric OC may be significant. 113 Although large wood debris has been analyzed in rivers (Seo et al., 2008; Hilton et al., 2010; Turowski et al., 2016; Wohl, 2017), there are only a few studies of wood export to the ocean (Kramer et al., 2017) and even fewer investigating wood preservation and burial in marine sediments (Saller et al., 2006; Sparkes et al., 2015). Intense physical erosion events such as landslides triggered by storms and earthquakes may export wood to the oceans in small mountainous river. For example, Typhoon Morakot triggered landslides in Taiwan in August 2009 delivering approximately 3.8 – 8.4 Tg of coarse woody debris to the oceans, representing 30 – 60% of the exported biospheric carbon (West et al., 2011) and partly transported and buried by turbiditic current (Sparkes et al., 2015). In the G – B catchment, wood export has been observed and/or documented during earthquake, landslide and glacial-dam outburst floods (23– 26) (Cenderelli and Wohl, 2001; Huang et al., 2014; Cook et al., 2018)(Montgomery et al., 2004), but we are aware of no studies that have tried to quantify wood export fluxes to the oceans from large rivers. Strong monsoonal precipitation is a seasonal driver of physical erosion in South and East Asia. High sedimentation rates in the South China Sea (Clift, 2006) and the incision of the Mekong River (Nie et al., 2018) have been interpreted as a strengthening of the monsoon during the Mid Miocene Climatic Optimum (MMCO), between 17 – 14 Ma. Our sampling does not reveal evidence for increased wood export at the MMCO, nor a change in the character of wood exported, however we find coarse woody material has been mobilized throughout the Miocene with frequent woody layers within the 18 – 17 and 12 – 11 Ma bins. Wood is present at a relatively consistent frequency in the three deep sites, 270 km apart over the last 19 Myr (Fig. 5.2D). This observation suggests that it may be an important component of OC burial. However, accurate accounting for such fluxes is challenging. Both mobilization and 114 export of woody debris is episodic in nature, associated with high flow events. In a very different fluvial system, the upper Rhone catchment, observed annual wood output during normal flows averaged 0.8 Mg km -2 whereas in a single flood, the wood output was 6 Mg km -2 (Gurnell et al., 2002). This makes it challenging to capture both in modern river systems and in ancient fan deposits where the locus of deposition also shifts over time. In the Bengal Fan, we expect the locus of deposition will change as the turbiditic canyon-levee systems migrate across the fan (Weber et al., 2003). In addition, drilling recovery of coarse-grained sediments is typically lower than that of fine-grained sediments (Fig. A.2), which may lead to underestimates of wood burial. 5.3.2 Lowland dominated, fresh wood export Lignin phenol S/V ratios indicate that woody debris exported and buried in the Bengal Fan is almost entirely from angiosperms (Fig. 5.3), which dominate the G – B catchment below 2500 m (Carpenter, 2005). One instance of high elevation gymnosperm export may reflect a rare mass wasting events e.g., during to glacial-dam failures (Huang et al., 2014; Cook et al., 2018), providing an efficient means of transporting coarse OC from the Himalayas to the Bay of Bengal over a short timescale. Angiosperm composition clearly implies dominantly lowland sourcing of woody debris throughout the record. A predominant lowland provenance has previously been documented in the modern Ganges-Brahmaputra River system for bulk organic carbon (Galy et al., 2008) and vascular plant biomarkers (Galy et al., 2011). Similarly, a persistent lowland provenance was inferred for organic carbon associated with fine sediments deposited in the Bengal Fan over the last 20 kyrs (Hein et al., 2017). The results presented here thus extend this observation to woody material exported to the Bengal Fan over the last 19 Myrs. 115 5.3.3. Vegetation change in the G-B catchment There is considerable carbon isotopic variability in the wood fragments recovered within closely spaced samples the Bengal Fan throughout the record. Based on the expected drivers of 13 C variability (canopy closure, relative humidity and elevation) (Diefendorf et al., 2010; Wu et al., 2017), and the range of modern sampled vegetation from the Himalaya and around the world (Tables A.11 and A.12), the Neogene variability is within the expected range for this catchment today (Fig. A.8). This inference is reasonable as pollen from the Siwalik Formation in the Himalayan foothills show that Miocene floodplain was covered with tropical forest with subtropical to temperate broadleaf forests in the foothills similar to the present (Hoorn et al., 2000). Pollen reconstructions from the Bengal Fan, from Deep Sea Drilling Project (DSDP) Site 218 find a mixture of tropical forest taxa and temperate taxa, with the gymnosperms (Pinaceae) dominating the pollen reaching the southern tip of the fan as pines are abundant pollen producers, in comparison to many tropical angiosperms (Chandra and Kumar, 1997). Overall, we find a stable (but heterogenous) buried wood signal across a 13-Myr period from 19 to 5.9 Ma that averages -26.6‰ (1 = 1.2‰, n = 180). This stability likely indicates similar vegetation sustained by a stable climate regime, with similar modes of erosion associated with periodic high-flow erosion and transport events. Heterogeneity within the samples likely reflects broad integration of wood from across the expansive and ecologically diverse lowland part of the drainage and the mixing of those sources during fluvial erosion, shelf storage and turbiditic resuspension. After the stability of the  13 C wood signal from 19 to 5.9 Ma, we have a gap in wood recovery (5.9 – 4.3 Ma) and the next sampled woody horizon at 4.3 Ma has a  13 C wood value of -18.9‰, which 116 is 4.9‰ more positive than the most 13 C-enriched sample during the previous 9 Myr (Fig. 5.3). As this is beyond the range of C 3 plants, even those in dry and high locations based on our survey (Table A.11 and A.12), we infer the presence of a mixture of C 3 and C 4 fragments within the wood sample. In total, three samples in the early Pliocene yield heterogeneous  13 C wood values between aliquots within a horizon, indicating a mixture C 3 and C 4 species in those sediment layers (Fig. 5.4). Although the presence of C 4 wood may be surprising, the presence of C 4 plants is not. In the late Miocene, there was an ecosystem transformation associated with a C 4 grassland expansion (monocotyledonous plants or monocots) that spread across the lowlands (France-Lanord and Derry, 1994; Quade et al., 1995). This region was the center of radiation in the C 4 Andropogoneae subfamily (Osborne, 2008). Today the catchment includes many tall C 4 grasses (Miscanthus, Saccharum, Hyparrhenia and Heteropogon), with stems that include woody tissues that may be strong enough to survive transport. For example, some of these species include stems that are bamboo-like in appearance, including one specimen sampled for the modern plant survey at 1.5 km in elevation (where otherwise C 3 bamboos are abundant; Table A.11). The abundance of tall grasses allows for the possibility that small fragments of woody-grass stems (including those of C 4 ) may be represented in the Bengal Fan. The C 4 pathway is also known within woody, dicotyledonous plants (dicots). Rarely present in trees, one major (angiosperm) tree group expressing the C 4 pathway is the Euphorbiaceae (Sage, 2001). This group is regionally significant, at least in the Pleistocene, as evidenced by pollen records from Ocean Drilling Program (ODP) Site 717 from the distal Bengal Fan, south of Sri Lanka, (Yasuda et al., 1990). Although pollen cannot discern whether shrubs or trees, it is 117 possible that the 13 C-enriched wood fragments come from the Euphorbiaceae, known to be present within the catchment. A third possibility could be contributions from xeric C 4 woody shrubs (dicots), although not currently growing in this very wet climatic zone. Yet there is pollen evidence from the Siwaliks that in the late Miocene, xeric C 4 woody shrubs including Amaranthaceae and Chenopodaceae appeared in the catchment (Hoorn et al., 2000). Whether monocot or dicot, C 4 plants grow mostly <1.5 km, and thus add to the evidence for the importance of lowland sourcing of woody debris found in the Bengal Fan. One might expect that the late Miocene expansion of C 4 grasslands would reduce the available woody biomass for erosion, however we do not find clear evidence for a step shift. After the gap in wood recovery from 5.9 to 4.3 Ma, we find continued (possibly reduced) delivery and burial of woody material in the Bengal Fan (Fig. 5.2). As the grasslands transformed the floodplain landscape (across expansive regions < 300 m) this suggests the foothills (ca. 300 – 1500 m) may be the dominant source region of the wood. The 13 C-enrichment trend continues into the Pleistocene. Some of the variability in Pleistocene  13 C wood could derive from large, cyclical variations associated with Pleistocene glacial cycles can drive up to 3‰ variations in  13 C atm (Tipple et al., 2010) (A.3), with additional variations resulting from the inclusion of wet, closed lowland forest or dry woody taxa at both extremes (Van Campo, 1986; Yasuda et al., 1990). In addition, between 0.4 and 0.1 Ma, some aliquots yielded  13 C values characteristic of pure C 4 plants (mean -13.0‰,  = 1.0‰, n = 17). We hypothesize that these fragments were mainly deposited during glacial periods when aridity favored C 4 plants in lowland part of the Ganges-Brahmaputra catchment (Hein et al., 2017). 118 Pleistocene pollen in DSDP Site 717 were dominated by Gramineae (grass) pollen (presumably C 4 ), with enhanced transport of Pinus pollen from Himalayan sources in cold intervals (Yasuda et al., 1990). Although expansion of the C 4 pathway is most commonly associated with grasslands (monocots), and many of these are woody enough to produce the buried C 4 woods observed, it is also possible that C 4 dicots were present. The C 4 pathway appeared in dicots during the cool, low atmospheric CO 2 times of the Pleistocene (Parapsychological Association, 1989; Ehleringer et al., 1997). C 4 dicots occupy ephemeral streams, disturbed or saline habitats. Only in glacial periods with lowered pCO 2 and drier conditions in the region, may C 4 dicots have a competitive advantage over other plants (Ehleringer et al., 1997). The buried wood record from the Bengal Fan indicates that C 4 woody debris is at times very abundant in the Pleistocene and provides catchment-scale evidence for the timing and magnitude of their expansion. Very low  13 C wood values are also observed during the Pleistocene, suggesting transport and burial processes are also effective during interglacial periods. Overall, we find a consistency in both the mean and the variance of the  13 C values of woody debris, which points to a large area sourcing and averaging process to maintain such stability. We find no evidence for single point source, such as catastrophic events with a homogenous signature but see in all wood horizons a mixing of heterogenous sources during the transport to the Bengal Fan deposition. This is consistent with evidence for lowland sourcing, as catastrophic erosion and transport processes such as landslides and glacial dam rupture are absent in the lowlands. Overall, the mode of wood mobilization from the landscape and delivery to the deep Bengal Fan differs from observations in steep erosive systems where mass wasting events and hyperpycnal flows account for most wood transport (Sparkes et al., 2015; Turowski et al., 2016). The Pleistocene stands out as an interval of extreme variability that reflects cyclical climate 119 variability and the resulting changes in vegetation and erosion as conditions change, including some high elevation sourcing in contrast to the dominant lowland mode of the Neogene. 5.3.4. Constraints on the isotopic composition of precipitation D methoxyl values are clustered between -280 and -230‰ across the entire 19 Myr record. The 50‰ range may represent spatial variations in D precip across the catchment, differences in microhabitat and fractionations ( 2  methoxyl/precip ) between species. The range of δD methoxyl values is consistent with mixed sourcing from a large area over the last 19 Myrs. Modern trees have been studied to constrain the biological fractionation ( 2  methoxyl/precip ) between D precip and D methoxyl and 2  methoxyl/precip is thought to be quite uniform (-216  17‰, 1, n = 222) based on sampling in Scandinavia and limited sampling around the world (Keppler et al., 2007; Anhäuser et al., 2017). A subset of oak, birch and pine, (representing genera present in the G – B catchment), yielded an indistinguishable mean value (-216  16 ‰, 1, n = 30) thus we use the overall mean 2  methoxyl/precip . Estimated δD precip from the wood samples recovered from the last 19 Ma range from +68 to -108‰ (mean = -48‰,  = 39‰, n = 59; Table A.9) with a cluster between -10 and -80‰. Modern δD precip weighted by precipitation amount range from -28 to -81‰ in the catchment below 2000 m (Fig. A.6, Table A.10). The overlap suggests broadly comparable monsoonal precipitation patterns during the last 19 Myr. However species-specific fractionations are entirely uncalibrated in this region and thus constitute a major uncertainty and possible source of the spread in D methoxyl values. A few samples in the Miocene with D methoxyl values below -280‰ and  13 C wood >-26.0‰ may derive from high altitude. Those samples at 11.8, 7.5 and 9.8 Ma yield D precip estimates of -87, - 120 90, and -107‰ respectively, values that suggest a high elevation source or a period of intensified monsoonal rainfall (Garzione et al., 2000; Gajurel et al., 2006). In a single Pleistocene horizon, lignin analyses reveal 6 of 9 aliquots were gymnosperm, thus derived from high elevation >2500 m (based on modern catchment survey; Dataset S6), mixed with 3 aliquots that were angiosperms likely from lower in the catchment. The gymnosperm fragments from this horizon yield D precip estimates as low as -108‰ (mean = -95‰,  = 14‰, n = 6; Table A.9), corroborating a high elevation provenance. In addition, the angiosperm debris from this same horizon yield higher D precip estimates (mean = +3‰,  = 6‰, n = 3; Table A.9), indicative of lowland sourcing. There are several mechanisms possible to explain high elevation sourcing including earthquake-triggered landsliding, or a glacial lake outburst flood associated with Pleistocene climate variability. Such an event could mobilize a large amount of woody debris from a high elevation ―point source‖ that can be transported rapidly to lower elevations, entraining wood from the lowlands, and perhaps triggering hyperpycnal flow facilitating rapid delivery to the deep sea. 5.4 Conclusion Geochemical analyses of wood preserved in the Bengal Fan show remarkably well-preserved, lowland angiosperms wood dominates the largest size fraction of the wood recovered. Although most woody horizons carry the hallmarks of lowland sources, we find one unusual horizon at 0.05Ma that indicates catastrophic erosion processes carrying gymnosperms from high altitude to the mid-Bengal Fan. Throughout the record, the range of  13 C wood and D methoxyl within sedimentary horizons indicates mixing of heterogenous sources within the lowland part of the 121 catchment. Although the climate, vegetation and wood-erosive regime of the Neogene appears relatively stable from this wood-derived record, we see a major transition with  13 C evidence for an increase in woody C 4 vegetation in the catchment in the late Miocene and further expansion in the glacial phases of the Pleistocene. The wood horizons of the Pleistocene stand out against the 19-Myr record as a time of heightened variability in source characteristics. Organic carbon burial in the Bengal Fan has been shown to be 2 – 3 times more efficient in sequestering atmospheric CO 2 than Himalayan silicate weathering, and has therefore been proposed as one of the forcing factors of Cenozoic cooling (Derry and France-Lanord, 1997; Galy et al., 2007). The discovery of woody fragments in the coarse grain layers of turbiditic sequences across 19 Myr, indicates that woody input may be a neglected contributor to the organic carbon cycle in the Bengal Fan and in other continental margins over geologic timescales. This has potential implications not only for organic carbon burial budgets (Burdige, 2007) but also for the fossil fuel generation potential of coarse siliciclastic sediments (Saller et al., 2006). The rapid delivery and burial of undegraded wood material represents a highly efficient pathway of atmospheric CO 2 sequestration. Wood burial indeed bypasses extensive organic carbon degradation that occurs during soil formation and that has been suggested to limit the efficiency of organic carbon burial at very high erosion rates (Galy et al., 2015). If triggered by high discharge events, wood erosion may follow the ―pulse-shunt‖ concept proposed for dissolved organic carbon (Raymond et al., 2016) and thus help sustain high organic carbon burial efficiency under enhanced physical erosion rates. 122 5.5 Materials and Methods Study Site. IODP Expedition 354 drilled Neogene and Quaternary sediments in the Bengal Fan and we studied 6 sites across a 320 km transect at 8º N in the mid Fan (mean water depth 3650 m, deepest sediment recovery 1200 mbsf). Middle to Late Pleistocene ages were modelled using an approach that factors litho-, magneto-, bio-, cyclo-, and seismic stratigraphic constraints, based on results from the International Ocean Discovery Program Expedition 354 Bengal Fan and analysis of the GeoB97-020/027 seismic line (54). Isolation of wood from the sediment core. Wood fragments that were visually identified were manually picked from the split core face using stainless-steel tweezers (58 samples) and coarse grain turbiditic layers were additionally sampled for wood. We freeze dried, disassociated and sieved (> 1 mm) to retrieve wood fragments from the matrix typically of 25 g of sediment, yielding an additional 37 wood samples from 73 samples examined. In this manner, we retrieved a semi-quantitative recovery of wood fragments from the marine sediments spanning 19 to 0.1 Ma. Isotopic analyses. For bulk wood OC and  13 C analyses, 50 to 150 g aliquots were isolated from wood samples (which consisted of one large piece to over a hundred small fragments). Samples were ground and treated with 4% HCl at 85º C for 1 h to remove carbonate, before measurement by EA-IRMS. TOC + indicates correction for loss during acidification (Galy et al., 2007).  13 C values are reported on the Vienna Pee Dee Belemnite / Lithium carbonate standard by L. Svec (VPDB/LSVEC) isotopic scale, with precision and accuracy better than 0.3 ‰. Variability between replicates represents sample heterogeneity as samples were not homogenized prior to analyses. For lignin methoxyl hydrogen isotopic analyses, 20 to 80 mg of powdered 123 wood was reacted with HI to release CH 3 I, then dissolved into isooctane for analysis by GC- IRMS. Data are reported on the Vienna Standard Mean Ocean Water / Standard Light Antarctic Precipitation (VSMOW/SLAP) isotopic scale, with precision and accuracy better than 6 ‰. Lignin analyses. Ground wood was transferred into Teflon lined bombs for CuO oxidation to isolate lignin derived phenols. The 0.5 g CuO, 100 mg ammonium iron (II) sulfate hexahydrate and 3 mL of 12 M NaOH under N 2 at 170 º C for 2.5 h. After the reaction, the water phase was acidified to pH 1 with 6 M HCl and kept for 1 h at room temperature in the dark to prevent reactions of cinnamic acids. After centrifugation (2500 rpm for 30 min), the supernatants were liquid-liquid extracted with diethyl ether. The ether extracts were concentrated by rotary evaporation, spiked with a known amount of internal standard (ethyl vanillin), transferred to 2 mL glass vials and dried under N 2 for GC–MS analysis. Acknowledgements This work was funded by US National Science Foundation OCE-1401217 and COL-T354A55 to SF and OCE-1400805 to VG. Graduate student participation in the project received support from University of Southern California Provost‘s Fellowship to HL. Samples were provided by the International Ocean Discovery Program. We are grateful for the efforts of the Expedition 354 Science Party, Carl Johnson and Zongguang Liu. CFL and AG were supported by IODP-France. 124 References Anhäuser, T., Greule, M., Keppler, F., 2017. Stable hydrogen isotope values of lignin methoxyl groups of four tree species across Germany and their implication for temperature reconstruction. Science of the Total Environment 579, 263–271. Bouma, A.H., 1962. Sedimentology of Some Flysch Deposits, Elsevier, Amsterdam. Elsevier Pub. Co., Amsterdam ;New York. Burdige, D.J., 2007. Preservation of organic matter in marine sediments: Controls, mechanisms, and an imbalance in sediment organic carbon budgets? Chemical Reviews 107, 467–485. Carpenter, C., 2005. The environmental control of plant species density on a Himalayan elevation gradient. Journal of Biogeography 32, 999–1018. Cenderelli, D.A., Wohl, E.E., 2001. Peak discharge estimates of glacial-lake outburst floods and ―normal‖ climatic floods in the Mount Everest region, Nepal. Geomorphology 40, 57–90. Chandra, A., Kumar, M., 1997. Palynology of the Late Tertiary sediments (DSDP Site 218) in the Bengal Fan, Indian Ocean. Palaeobotanist 46, 51–69. Clift, P.D., 2006. Controls on the erosion of Cenozoic Asia and the flux of clastic sediment to the ocean. Earth and Planetary Science Letters 241, 571–580. Cook, K.L., Andermann, C., Gimbert, F., Adhikari, B.R., Hovius, N., 2018. Glacial lake outburst floods as drivers of fluvial erosion in the Himalaya. Science 362, 53–57. Derry, L., France-Lanord, C., 1997. Organic carbon burial forcing of the carbon cycle from Himalayan erosion. Nature 390, 65–67. Diefendorf, A.F., Mueller, K.E., Wing, S.L., Koch, P.L., Freeman, K.H., 2010. Global patterns in leaf 13C discrimination and implications for studies of past and future climate. Proceedings of the National Academy of Sciences 107, 5738–5743. Ehleringer, J.R., Cerling, T.E., Helliker, B.R., 1997. C4 photosynthesis, atmospheric CO2, and climate. Oecologia 112, 285–299. France-Lanord, C., Derry, L.A., 1994. δ13C of organic carbon in the Bengal Fan: Source evolution and transport of C3 and C4 plant carbon to marine sediments. Geochimica et Cosmochimica Acta 58, 4809–4814. France-Lanord, C., Spiess, V., Klaus, A., Adhikari, R.R., Adhikari, S.K., Bahk, J.-J., Baxter, A.T., Cruz, J.W., Das, S.K., Dekens, P., Duleba, W., Fox, L.R., Galy, A., Galy, V., Ge, J., Gleason, J.D., Gyawali, B.R., Huyghe, P., Jia, G., Lantzsch, H., Manoj, M.C., Martos Martin, Y., Meynadier, L., Najman, Y.M.R., Nakajima, A., Ponton, C., Reilly, B.T., Rogers, K.G., Savian, J.F., Schwenk, T., Selkin, P.A., Weber, M.E., Williams, T., Yoshida, K., 2016. 125 Proceedings of the International Ocean Discovery Program. International Ocean Discovery Program 354. doi:10.14379/iodp/proc.354.101.2016 Friedli, H., Lötscher, H., Oeschger, H., Siegenthaler, U., Stauffer, B., 1986. Ice core record of the 13C/12C ratio of atmospheric CO2 in the past two centuries. Nature 324, 237–238. Gajurel, A.P., France-Lanord, C., Huyghe, P., Guilmette, C., Gurung, D., 2006. C and O isotope compositions of modern fresh-water mollusc shells and river waters from the Himalaya and Ganga plain. Chemical Geology 233, 156–183. Galy, V., Eglinton, T., France-Lanord, C., Sylva, S., 2011. The provenance of vegetation and environmental signatures encoded in vascular plant biomarkers carried by the Ganges- Brahmaputra rivers. Earth and Planetary Science Letters 304, 1–12. Galy, V., France-Lanord, C., Beyssac, O., Faure, P., Kudrass, H., Palhol, F., 2007. Efficient organic carbon burial in the Bengal fan sustained by the Himalayan erosional system. Nature 450, 407–410. Galy, V., France-Lanord, C., Lartiges, B., 2008. Loading and fate of particulate organic carbon from the Himalaya to the Ganga-Brahmaputra delta. Geochimica et Cosmochimica Acta 72, 1767–1787. Galy, V., Peucker-Ehrenbrink, B., Eglinton, T., 2015. Global carbon export from the terrestrial biosphere controlled by erosion. Nature 521, 204–207. Garzione, C.N., Quade, J., DeCelles, P.G., English, N.B., 2000. Predicting paleoelevation of Tibet and the Himalaya from δ18O vs. altitude gradients in meteoric water across the Nepal Himalaya. Earth and Planetary Science Letters 183, 215–229. Gurnell, A.M., Piégay, H., Swanson, F.J., Gregory, S. V., 2002. Large wood and fluvial processes. Freshwater Biology 47, 601–619. Hayes, J.M., Waldbauer, J.R., 2006. The carbon cycle and associated redox processes through time. Philosophical Transactions of the Royal Society B: Biological Sciences 361, 931–950. Hedges, J.I., Cowie, G.L., Ertel, J.R., James Barbour, R., Hatcher, P.G., 1985. Degradation of carbohydrates and lignins in buried woods. Geochimica et Cosmochimica Acta 49, 701–711. Hedges, J.I., Keil, R.G., 1995. Sedimentary organic matter preservation: an assessment and speculative synthesis. Marine Chemistry 49, 81–115. Hedges, J.I., Mann, D.C., 1979. The characterization of plant tissues by their lignin oxidation products. Geochimica et Cosmochimica Acta 43, 1803–1807. Hein, C.J., Galy, V., Galy, A., France-Lanord, C., Kudrass, H., Schwenk, T., 2017. Post-glacial climate forcing of surface processes in the Ganges–Brahmaputra river basin and 126 implications for carbon sequestration. Earth and Planetary Science Letters 478, 89–101. Hilton, R.G., Galy, A., Hovius, N., Horng, M.-J., Chen, H., 2010. The isotopic composition of particulate organic carbon in mountain rivers of Taiwan. Geochimica et Cosmochimica Acta 74, 3164–3181. Hoorn, C., Ohja, T., Quade, J., 2000. Palynological evidence for vegetation development and climatic change in the Sub-Himalayan Zone (Neogene, Central Nepal). Palaeogeography, Palaeoclimatology, Palaeoecology 163, 133–161. Huang, S.Y., Chen, Y.G., Burr, G.S., Jaiswal, M.K., Lin, Y.N., Yin, G., Liu, J., Zhao, S., Cao, Z., 2014. Late Pleistocene sedimentary history of multiple glacially dammed lake episodes along the Yarlung-Tsangpo river, southeast Tibet. Quaternary Research (United States) 82, 430–440. Kennedy, M.J., Wagner, T., 2011. Clay mineral continental amplifier for marine carbon sequestration in a greenhouse ocean. Proceedings of the National Academy of Sciences 108, 9776–9781. Keppler, F., Harper, D.B., Kalin, R.M., Meier-Augenstein, W., Farmer, N., Davis, S., Schmidt, H.L., Brown, D.M., Hamilton, J.T.G., 2007. Stable hydrogen isotope ratios of lignin methoxyl groups as a paleoclimate proxy and constraint of the geographical origin of wood. New Phytologist 176, 600–609. Kramer, N., Wohl, E., Hess-Homeier, B., Leisz, S., 2017. The pulse of driftwood export from a very large forested river basin over multiple time scales, Slave River, Canada. Water Resources Research 53, 1928–1947. Mayer, L.M., 1994. Surface area control of organic carbon accumulation in continental shelf sediments. Geochimica et Cosmochimica Acta 58, 1271–1284. Montgomery, D.R., Hallet, B., Yuping, L., Finnegan, N., Anders, A., Gillespie, A., Greenberg, H.M., 2004. Evidence for Holocene megafloods down the Tsangpo River gorge, southeastern Tibet. Quaternary Research 62, 201–207. Nie, J., Ruetenik, G., Gallagher, K., Hoke, G., Garzione, C.N., Wang, W., Stockli, D., Hu, X., Wang, Z., Wang, Y., Stevens, T., Danišík, M., Liu, S., 2018. Rapid incision of the Mekong River in the middle Miocene linked to monsoonal precipitation. Nature Geoscience 11. doi:10.1038/s41561-018-0244-z Osborne, C.P., 2008. Atmosphere, ecology and evolution: What drove the Miocene expansion of C4grasslands? Journal of Ecology. Parapsychological Association, 1989. Terms and Methods in Parapsychological Research. Journal of Humanistic Psychology 29, 394–399. 127 Quade, J., Cater, J.M.L., Ojha, T.P., Adam, J., Harrison, T.M., 1995. Late Miocene environmental change in Nepal and the northern Indian subcontinent: stable isotopic evidence from paleosols. Geological Society of America Bulletin 107, 1381–1397. Raymond, P.A., Saiers, J.E., Sobczak, W. V., 2016. Hydrological and biogeochemical controls on watershed dissolved organic matter transport: Pulse- shunt concept. Ecology 97, 5–16. Reilly, B. (2018) Deciphering Quaternary Geomagnetic, Glacial, and Depositional Histories Using Paleomagnetism in Tandem with Other Chronostratigraphic and Sedimentological Approaches. PhD thesis, Oregon State University, Corvallis. Sage, R.F., 2001. Environmental and evolutionary preconditions for the origin and diversification of the C4 photosynthetic syndrome. Plant Biology 3, 202–213. Saller, A., Lin, R., Dunham, J., 2006. Leaves in turbidite sands: The main source of oil and gas in the deep-water Kutei Basin, Indonesia. AAPG Bulletin 90, 1585–1608. Seo, J. Il, Nakamura, F., Nakano, D., Ichiyanagi, H., Chun, K.W., 2008. Factors controlling the fluvial export of large woody debris, and its contribution to organic carbon budgets at watershed scales. Water Resources Research 44, 1–13. Sparkes, R.B., Lin, I.T., Hovius, N., Galy, A., Liu, J.T., Xu, X., Yang, R., 2015. Redistribution of multi-phase particulate organic carbon in a marine shelf and canyon system during an exceptional river flood: Effects of Typhoon Morakot on the Gaoping River-Canyon system. Marine Geology 363, 191–201. Thomas, S.C., Martin, A.R., 2012. Carbon content of tree tissues: A synthesis. Forests 3, 332– 352. Tipple, B.J., Meyers, S.R., Pagani, M., 2010. Carbon isotope ratio of Cenozoic CO 2 : A comparative evaluation of available geochemical proxies. Paleoceanography 25, 1–11. Turowski, J.M., Hilton, R.G., Sparkes, R., 2016. Decadal carbon discharge by a mountain stream is dominated by coarse organic matter. Geology 44, 27–30. Van Campo, E., 1986. Monsoon fluctuations in two 20,000-Yr B.P. Oxygen-isotope/pollen records off southwest India. Quaternary Research 26, 376–388. Weber, M.E., Wiedicke-Hombach, M., Kudrass, H.R., Erlenkeuser, H., 2003. Bengal Fan sediment transport activity and response to climate forcing inferred from sediment physical properties. Sedimentary Geology 155, 361–381. West, A.J., Lin, C.W., Lin, T.C., Hilton, R.G., Liu, S.H., Chang, C.T., Lin, K.C., Galy, A., Sparkes, R.B., Hovius, N., 2011. Mobilization and transport of coarse woody debris to the oceans triggered by an extreme tropical storm. Limnology and Oceanography 56, 77–85. Winterfeld, M., Laepple, T., Mollenhauer, G., 2015. Characterization of particulate organic 128 matter in the Lena River delta and adjacent nearshore zone, NE Siberia - Part I: Radiocarbon inventories. Biogeosciences 12, 3769–3788. Wohl, E., 2017. Bridging the gaps: An overview of wood across time and space in diverse rivers. Geomorphology 279, 3–26. Wu, M.S., Feakins, S.J., Martin, R.E., Shenkin, A., Bentley, L.P., Blonder, B., Salinas, N., Asner, G.P., Malhi, Y., 2017. Altitude effect on leaf wax carbon isotopic composition in humid tropical forests. Geochimica et Cosmochimica Acta 206, 1–17. Yasuda, Y., Amano, K., Yamanoi, T., 1990. Pleistocene Climatic Changes as Deduced from a Pollen Analysis of Site 717 Cores. Proceedings of the Ocean Drilling Program, 116 Scientific Results 116, 249–257. 129 Chapter 6 Dissertation Conclusions Hydrogen and carbon isotopic compositions of plant waxes and more recently methoxyl groups within the lignin macromolecule, have been used to provide information about the changes in vegetation and hydroclimate of the past. The paleoclimate applications of both biomarkers are established upon numerous modern calibration studies that demonstrate a robust relationship between their isotopic signatures and environmental factors. However, secondary isotopic fractionations from molecular synthesis in plants to sample analysis can potentially obscure the climatic signal of interest. Chapters 2 and 3 focus on the plant wax biomarker and I use controlled experiments to determine how stomatal density and acid-catalyzed methylation affect isotopic compositions of n-alkanes and n-alkanoic acids. Chapters 4 and 5 branch out to study climatic information embedded in another plant molecule, lignin. Hydrogen isotopic composition of methoxyl groups in lignin have been suggested as a potential paleoclimate indicator but analytical methods and standardization for isotope work are in its early stages. In chapter 4, I use a liquid injection method to analyze the concentration and hydrogen isotopic composition of methoxyl groups in lignin-bearing samples that have potential to serve as isotopic standards. I then apply the proxy towards a set of buried wood samples found in the Bengal Fan over the last 19 million years to reconstruct the source of wood in the Ganges – Brahmaputra (G–B) catchment. 130 Chapter reviews and future work Chapter 2 investigates the effects of varying stomatal density on plant wax isotopic composition. Stomata are pores on leaves that regulate carbon uptake and water loss during photosynthesis. Changes in the stomatal density are known to be due to atmospheric CO 2 concentration and large swings in pCO 2 over the geologic timescale raised a concern that isotope effects associated with stomatal density may need to be taken into account for paleoclimate studies using plant waxes. I used genetically modified Arabidopsis thaliana, to provide the first controlled tests of stomatal density on carbon (δ 13 C) and hydrogen (δD) isotopic compositions of n-alkanes and found that the overexpressed A.thaliana mutant was shifted by -3‰ and +15‰ for δ 13 C and δD compared to the suppressed mutant. The shifts in δD and δ 13 C of n-alkanes with stomatal density changes could have implications for paleoclimate studies especially for high pCO 2 , hyperthermal events and glacial periods when stomatal densities varied the most. However, climatic factors associated with pCO 2 , such as temperature and relative humidity, would also influence stomatal conductance and transpiration, which may offset the observed D-enrichment and 13 C-depletion in the plant wax n-alkanes. We suggest experiments with these genetic lineages under varying environmental conditions would help quantify those additional effects further. In terms of future applications, we recommend taking advantage of genetic mutations of A. thaliana to constrain isotopic fractionations caused by other plant physiological and biochemical changes relevant for paleoclimate applications. For example, genes that affect not only stomatal density but size and closure of stomata are available, along with mutations on aquaporin, leaf venation patterns and cell wall degradability. In chapter 3, hydrogen and carbon isotopic fractionations associated with methylation protocols of aromatic and aliphatic molecules were examined. Methylation is a widely-used procedure for 131 samples with carboxylic acid groups prior to gas chromatography and commonly calls for acidic, hot reaction conditions that are known to promote H/D exchange in aromatic and aliphatic C–H bonds. We tested two commonly-used methods and compared to a third method that avoids these acidic conditions to quantify isotopic fractionations associated with using an acidic catalyst. δD, δ 13 C values and reaction yields were equivalent among the three methods for methyl octacosanoate. For dimethyl phthalate, a small aromatic ester, the two acidic methods resulted in comparable yield and isotopic compositions with D-enrichment shifts but the non-acidic method resulted in more negative δD and δ 13 C values perhaps due to low yields. Using nuclear magnetic resonance, we found that the acid-catalyzed H/D exchange rates (4.94  10 -4 hr -1 ) are too slow to cause measurable isotopic fractionation over the typical duration and reaction conditions used in phthalic acid methylation. Product isolation yield and evaporation were more important factors for caution to accurately determine δD of dimethyl phthalate and fatty acid methyl esters. Fortunately, even with isotopic fractionations observed in δD in dimethyl phthalate, propagated errors for n-alkanoic acid applications were small, within typical instrumental uncertainty. Determining isotope exchange rates using nuclear magnetic resonance can be applied in other reaction conditions relevant for plant biomarkers. Base can also catalyze methylation and their effect on hydrogen and carbon isotopic compositions remain untested. Chapter 4 investigates quantification and isotopic analyses of lignin methoxyl groups to optimize methods for geological applications. Lignin-bearing substrates are reacted with hydroiodic acid to evolve iodomethane gas which is then analyzed in gas chromatography systems. I compare headspace and liquid injection methods for iodomethane quantification and hydrogen isotopic composition and show limitations of the former approach due to the low boiling point of iodomethane. Using the liquid method by dissolving iodomethane in isooctane is recommended 132 for both quantification and δD analyses. Quantification results show an increase in lignin methoxyl content during early diagenesis of wood when cellulose degrades and a decrease in methoxyl content associated with coal maturation. Our approach in lignin methoxyl quantification also complements classic maturation indicators such as lignin to cellulose ratios. The isotopic determinations of methoxyl groups have potential for application to tree wood, and their sedimentary archives such as buried logs, peats and lignites, especially those derived from gymnosperms and angiosperms. However, boreal peats derived from sphagnum and sapropelic lignites are not promising targets given the low methoxyl concentrations in moss lignin. Hydrogen isotopic composition of lignin methoxyl groups is an exciting new proxy for precipitation δD and in conjunction with abundance of methoxyl groups, various interesting questions related to lignin biogeochemistry, wood and food forensics, and paleoclimate/paleoecology can be explored. Remaining work includes establishing international isotopic standards to accurately determine of lignin methoxyl δD and δ 13 C samples. In addition, more plant types and species especially in tropical/subtropical regions need to be analyzed to further constrain the net fractionation term between lignin methoxyl δD and precipitation δD. For example, plants using the C 4 photosynthetic pathway are missing from the global survey for lignin methoxyl δD. Chapter 5 applies the analytical techniques described in chapter 4 along with other geochemical proxies to study fossil wood preserved in the Bengal Fan over the last 19 million years. Wood export has been observed in small mountainous rivers during catastrophic flood events or earthquake-triggered landslides, however wood was not widely thought to survive transport to and burial in the oceans. This study shows that woody debris can be transported thousands of km in rivers and in under-sea turbiditic currents to be deposited and preserved in the middle of a 133 marine fan. We found that the buried wood fragments were minimally degraded despite their age and were dominantly derived from lowland angiosperms. Throughout the record, the large range of carbon isotopic composition of wood ( 13 C wood ) and hydrogen isotopic composition of lignin methoxyl groups (D methoxyl ) indicated mixing of heterogenous sources. We also see a transition in  13 C wood for an increase in woody C 4 vegetation in the catchment in the late Miocene and further expansion in the glacial periods in the Pleistocene. The discovery of woody fragments in the coarse grain layers of turbiditic sequences across 19 million years, indicates that woody input may be an overlooked contributor to the organic carbon cycle in the Bengal Fan and in other continental margins over geologic timescales. Future work could include systematic sampling of buried wood found in the Expedition 354 turbidites and quantifying the wood organic carbon concentration in the Bengal Fan sediments. A similar approach shown in this chapter can also be applied to other fans and continental margins to study global wood export and burial patterns. 134 Appendix A.1 Supplementary Information for Chapter 3 A.1.1 Protocol for phthalic acid methylation A.1.1.1 Purpose This is a modification of a standard esterification procedure (Wakeham and Pease, 1992) to minimize isotopic fractionation and optimize yield for phthalic acid methylation. Fatty acids are derivatized prior to gas chromatography typically by the addition of a methyl group. The isotopic composition of that added group must be known to calculate the isotopic composition of the fatty acid. It has previously been suggested that phthalic acid presents a useful reagent for determination of the isotopic composition of the added methyl group as phthalic acid is stable for long term storage and shipment, and its isotopic composition can be independently measured offline relative to the international reference standards (Sessions et al., 2002). In this way, the isotopic composition of the methyl group in methanol can be readily determined in-house without need for shipments and storage of a volatile compound. Here we report a recommended method for the preparation of dimethyl phthalate yields for both carbon and hydrogen isotopic analysis by gas chromatography flame ionization detection/isotope ratio mass spectrometry (GC/FID/IRMS) using hydrochloric acid. Other common catalysts include acetyl chloride and boron trifluoride (see text). A.1.1.2 Necessary materials 1. Isotopically-determined phthalic acid 2. 5 wt% H 2 O deactivated silica gel 135 3. Solvents: ultra-pure water (Milli-Q or similar); dichloromethane (DCM) and hexane both gas chromatography (GC) grade 4. Hydrochloric acid 5. Combusted borosilicate glass Pasteur pipettes and glass wool for plugging 6. Sodium sulfate (Na 2 SO 4 ) anhydrous 7. Hot plate 8. 15 mL borosilicate glass culture tube and Teflon lined cap 9. Teflon tape 10. 40 mL borosilicate glass vial 11. 2 mL vial and Teflon-silicone-Teflon septum cap 12. N 2 gas stream for evaporation. A.1.1.3 Procedure 1. Weigh out 3 mg of phthalic acid and record the weight (for yield calculations). 2. Prepare 5 vol% HCl in methanol. 3. Put phthalic acid (1) and 1 mL of the mixture (2) to into a 15 mL culture tube. 4. Cap vial with Teflon lined cap and wrap Teflon tape around the cap to ensure a good seal. 5. Place in a hot block at 70 °C overnight. 6. At completion, check no loss of reagent (solvent level unchanged). 7. Turn off the hot plate and allow to cool. (Dimethyl phthalate will undergo ~ 0.3 ‰ 2 H- enrichment for every 12 hours, so should be promptly removed from the heat and acidic solution, and processed after the overnight reaction.) 8. Perform liquid-liquid extraction and pass desired fraction through sodium sulfate column. a. Add 1 mL of MilliQ to dilute the reagent to enable separation. 136 b. Add 1 mL of DCM, cap tightly. c. Shake vigorously for 30 s. d. Leave to settle, until two distinct density-separated layers form: the (upper) aqueous layer and the (lower) DCM layer. (Note if using hexanes for fatty acid extractions (b) hexanes will be the upper layer). e. Extract the bottom layer (DCM) with a short pipette. f. Load the extracted fraction onto a column containing Na 2 SO 4 anhydrous. Collect the eluent in a 40 mL vial. g. Repeat steps b – f 3 to 5 times. 9. Silica gel. a. Prepare a silica gel column and flush with DCM. b. Load the fraction collected in 8f onto the silica gel column and elute with DCM. 10. Dry and transfer. a. Evaporate the solvent from 9b under gentle N 2(g) stream (without heating). Do not dry completely to avoid evaporating dimethyl phthalate. b. Transfer to a 2 mL GC vial with DCM. c. Repeat a. d. Add known volume of hexanes and analyze the dimethyl phthalate yield by GC/FID and isotopic composition by GC/IRMS. Recap between measurements and store Teflon taped in solvent. A.1.1.4 Isotope mass balance 1. Dimethyl phthalate ((C 2 H 3 O 2 ) 2 C 6 H 4 ) 137 a. b. 2. n-Alkanoic acid (C n H 2n-1 O 2 ) a. b. A.1.1.5 Error propagation 1. GC/IRMS analytical uncertainties on an ester The combined uncertainty on GC/IRMS δ 2 H determinations of an analyte, derivatized as a methyl ester, is often ~5 ‰. As described in Polissar and D‘Andrea (2014), propagated uncertainties include: a. instrument precision for replicate measurements of the analyte, and b. the reference gas precision during sample and standard runs, and c. the uncertainty when comparing measured and known standard values, as well as d. The uncertainty on the methyl group added during methylation (2). 2. Phthalic acid methyl ester chemical preparative uncertainties 138 The uncertainty on the methyl group added during methylation is often determined by derivatizing a phthalic acid (known isotopic composition), a molecule presenting two locations for methyl group additions, with an unknown methanol. a. The uncertainty reported on Na-phthalate δ 2 H is 2.2 ‰ (1, n = 3). b. The reproducibility (reaction chemistry uncertainty) when methylating phthalic acid under normal, careful laboratory procedures is ~1 ‰ (1, n = 32). We recommend at least 3 reactions are performed to check preparative reproducibility. c. When reacting isotopically-known methanol and phthalic acid, we found the calculated isotopic compositions of methyl hydrogen to be offset from that of the known methanol value by 16 ‰ (n = 10; Fig 3.3). This inaccuracy points to an isotope effect during the methylation reaction or an isotopic fractionation during preparation. i. We find negligible hydrogen isotope exchange under reaction conditions. ii. We find yield to be only a problem for the non-acidic method, both acidic methods have quantitative yields. iii. Evaporation during the preparation or measurement may lead to 2 H- enrichment of ~10 ‰ in extreme user-error scenarios. This is avoidable and can be monitored by quantification. 3. This study shows that the uncertainty on the methyl group added in methylation reactions is underestimated when based on GC/IRMS instrument precision (1), Na-phthalate (2a) and chemistry-reproducibility (2b). The largest source of uncertainty arises from inaccuracy in determinations (2c) and we find this is not due to acidic exchange. We find phthalic acid is a particularly susceptible analyte due to its volatility. The key implication 139 is that evaporation should be minimized and analyte yields should be monitored, especially for volatiles. 4. For the unknown analyte, the uncertainty in δ 2 H determination, introduced by the methyl group uncertainty will vary by mass balance (D). For long-chain compounds, e.g. C 28 n- alkanoic acid (from plant leaf wax), the weight of methyl group is small such that the combined, propagated errors are within 1 ‰ even if the methyl group using dimethyl phthalate are as large as 18 ‰, as identified here. For shorter-chain compounds such as C 16 n-alkanoic acid (a lipid present in many organisms) propagated errors are commensurately larger by 1 ‰ (Table 1). Table A.1 Comparison of dimethyl phthalate yield and carbon and hydrogen isotopic composition by three methylation approaches. Table A.2 Comparison of methyl octacosanoate yield and carbon and hydrogen isotopic composition by three methylation approaches. Method Yield (%) n  13 C (‰) known n  2  ‰ known n  13 C (‰) in-house n  2  ‰ in-house n Acidic methods: CH 3COCl 80 ± 5 10 -31.5 ± 0.1 3 -105 ± 2 3 -26.4 ± 0.1 7 -145 ± 1 7 HCl 73 ± 5 13 -31.3 ± 0.1 3 -107 ± 2 3 -26.7 ± 0.1 10 -152 ± 3 10 Mean 76 ± 7 23 -31.4 ± 0.1 6 -106 ± 3 6 -26.6 ± 0.1 17 -149 ± 3 17 Non-acidic method: EDCI/DMAP 22 ± 6 34 -32.7 ± 0.4 4 -137 ± 2 4 -27.6 ± 0.2 9 -178 ± 3 9 ―Known‖ – using the phthalic acid standard and methanol of known isotopic composition provided by University of Indiana, see text for details. ―in-house‖ – this is a standard purchased for in-house use isotopic composition. Method Yield (%)  13 C (‰)  2  ‰ n CH 3COCl 95 ± 2 -27.0 ± 0.1 -222 ± 0 3 HCl 94 ± 3 -26.6 ± 0.0 -223 ± 0 3 EDCI/DMAP 82 ± 4 -26.7 ± 0.1 -224 ± 0 3 140 Table A.3 Individual replicate tests of methylation of phthalic acid, data summarized in Table A.1. Yield (%)  2 H (‰)  ‰  13 C (‰)  ‰ CH3COCl 80 -143 1 -26.4 0.1 83 -143 1 -26.4 0.0 84 -143 3 -26.4 0.0 74 -146 2 -26.6 0.0 91 -145 1 -26.6 0.0 77 -145 1 -26.2 0.0 80 -146 1 -26.2 0.0 HCl 75 -155 1 -26.6 0.0 75 -155 1 -26.6 0.0 70 -156 0 -26.5 0.0 81 -154 0 -26.8 0.1 67 -150 2 -26.5 0.0 65 -151 3 -26.6 0.0 69 -149 3 -26.7 0.1 77 -151 1 -26.8 0.0 78 -151 2 -26.9 0.0 70 -150 1 -26.7 0.0 EDCI/DMAP 21 -174 2 -27.9 0.1 15 -177 0 -27.8 0.0 22 -180 1 -27.6 0.1 30 -176 2 -27.6 0.1 21 -175 3 -27.4 0.1 20 -177 1 -27.5 0.1 22 -185 3 -27.5 0.0 32 -178 2 -27.5 0.0 141 Table A.4 Yields of dimethyl phthalate with variations to the stated experimental protocol for EDCI/DMAP method. Yield (%) Experimental variation 20 DCM solvent for liquid-liquid extraction 19 DCM solvent for liquid-liquid extraction 31 DCM solvent for liquid-liquid extraction 26 1:1 DCM:MeOH solvent for liquid-liquid extraction 24 1:1 DCM:MeOH solvent for liquid-liquid extraction 31 Ethyl acetate solvent for liquid-liquid extraction 28 Ethyl acetate solvent for liquid-liquid extraction 21 Evaporation of solvent before liquid-liquid extraction 24 Evaporation of solvent before liquid-liquid extraction 18 Evaporation of solvent before liquid-liquid extraction 19 Reaction temperature, 70 C 22 Reaction temperature, 70 C 17 Reaction temperature, 70 C, 36 hr reaction 18 48 hr reaction 12 96 hr reaction 31 8 mol EDCI 12 16 mol EDCI 10 24 mol EDCI 22 brine wash 18 citric acid wash 16 half silica gel mass for the separation column 142 A.1.2 References S. G. Wakeham and T. K. Pease, Lipid Analysis in Marine Particle and Sediment Samples. A Laboratory handbook. Unpublished manuscript 1992. A. L. Sessions, L. L. Jahnke, A. Schimmelmann, J. Hayes. Hydrogen isotope fractionation in lipids of the methane-oxidizing bacterium Methylococcus capsulatus. Geochim. Cosmochim. Acta 2002, 66, 3955. P. J. Polissar, W. J. D‘Andrea. Uncertainty in paleohydrologic reconstructions from molecular δ 2 H values. Geochim. Cosmochim. Acta 2014, 129, 146. 143 A.2 Supplementary Information for Chapter 4 Table A.5 Miocene age lignite samples from Belchatow deposit in Poland (Drobniak and Mastalerz, 2006). Samples were classified into four groups with regard to degree of gelification from 1 (non-gelified) to 4 (high gelification). Sample name g methyl / g sample (wt %) Group 1 10/1402 3.14 31/98 3.22 18/98 2.48 10/98 2.57 19/98 2.23 18/96 2.49 14/96 2.33 Group 2 10/384 2.83 6/98 3.17 16/96 2.64 4/384 2.88 12/98 2.42 6/1402 2.88 8/384 2.83 9/1402 1.82 34/98 3.37 39/98 3.33 7/96 3.28 Group 3 13/384 1.70 37/98 3.06 7/384 3.25 20/98 3.20 3/98 3.71 11/384 3.03 42/98 2.03 8/1402 2.94 8/98 2.95 24/98 2.79 2/98 2.86 26/98 3.37 27/98 3.07 28/98 2.95 Group 4 1A 3.60 2A 3.37 1/96 2.87 21/98 3.84 144 A.2.1 References Drobniak, A., Mastalerz, M., 2006. Chemical evolution of Miocene wood: Example from the Belchatow brown coal deposit, central Poland. International Journal of Coal Geology 66, 157– 178. 145 A.3 Supplementary Information for Chapter 5 A.3.1 Extended methods We quantified 95 wood samples (37 from 73 sediment samples ranging 11 – 77 g and 58 individual wood pieces directly sampled from split core surfaces). Sample distribution across the 19 Myr sedimentary record depended on core recovery and visible presence of wood fragments. Sections of turbiditic units (sub-unit Ta through Td of the Bouma sequence) with coarse grains and other units with black-colored particles were selected for sampling. A.3.1.1 Bulk wood elemental and isotopic analyses For bulk wood TOC and  13 C analyses, 50 to 100 g of ground wood was treated with 4% HCl at 85º C for 1 h to remove carbonate, before measurement on a modified EuroEA3028-HT elemental analyzer coupled to a GV Instruments IsoPrime continuous-flow isotope mass spectrometer. TOC + indicates correction for loss during acidification. Although samples were freeze-dried before picking and grinding wood samples, water absorption from a humid atmosphere may be an additional explanation for lower than expected OC%, in addition to matrix dilution and degradation during burial. Calibration to the LVSEC/VPDB isotopic scale was achieved with two internal standards in mineral matrix and a third standard run unknown to assess accuracy. The  13 C results were accurate and precise to better than 0.3‰. A.3.1.2 Lignin methoxyl hydrogen isotopic analyses For lignin methoxyl isotopic analyses, 20 to 80 mg of powdered wood was reacted with 0.1 – 0.15 mL of hydroiodic acid (55% HI) in the 2 mL GC vials with Al crimp caps with PTFE/rubber TF2 septum gas tight septa. Samples were heated at 120º C for 30 min and shielded from light. Vials were cooled and held at ambient conditions ~22º C for at least 30 min to allow 146 the iodomethane product to equilibrate into headspace. Samples were neutralized via the injection of 0.15 mL 5 M KOH through the septum with additional 0.05 – 0.1 mL until neutralized (as detected by color change). Liquid – liquid extraction was performed to partition the iodomethane into isooctane, the organic phase. 250 L of isooctane was injected into the vial through the septum after neutralization. The mixture was vigorously shaken for 30 sec, centrifuged to facilitate separation into organic and aqueous layers. The organic phase containing iodomethane was extracted by syringe (1 – 3 times) and the fractional volume recovery was recorded. The D methoxyl was determined at USC using a Trace GC connected via a GC–Isolink pyrolysis furnace (at 1400C), passing through a cold trap, and a Conflo IV interface to Delta VPlus IRMS. The GC was fitted with a ZB-5 ms column (30 m  0.25 mm  1 m) and for the liquid method, 1-2 mL of isooctane containing iodomethane was injected using a gas tight syringe into a split/splitless (SSL) inlet operated at constant temperature (200C) in splitless mode. The flow rate was at 3 mL min -1 with the oven temperature starting at 33C, held for 2.5 min, followed by a 20C min -1 ramp up to 130C which was held for 0.6 min to remove the isooctane. The GC Isolink backflush multi-functional valve controller (MFVC) was heated to 65º C, necessary to aid expulsion of isooctane. Four peaks of hydrogen reference gas bracket the iodomethane analyte peak during the GC- IRMS run. One of the initial peaks was used for standardization of the isotopic analyses, while the other three bracketing peaks were treated as unknowns (precision averaged 0.4‰, 1, n = 208). The known reference peak was set by comparing with an external standard (A6mix) obtained from A. Schimmelmann, Indiana University, Bloomington, containing 15 n-alkane 147 compounds (C 16 to C 30 ), with D values spanning –9 to –254‰. The RMS error determined by replicate measurements of the standard across the course of analyses was 4.2‰. Data were then normalized to the Vienna Standard Mean Ocean Water / Standard Light Antarctic Precipitation (VSMOW/SLAP) isotopic scale by comparison with both the A6mix and an external standard of 99.7% purity CH 3 I analyzed by offline combustion and analysis by dual inlet IRMS with D values –95.6‰, 1 1.6‰, n = 6 (the D value was analyzed and supplied by A. Schimmelmann, Indiana University, Bloomington). Three additional in-house standards (USC Lignin, Bamboo, Poplar (-254‰, -187‰, and -330‰, respectively; available from S. Feakins upon request) were run to monitor for consistency, following the principle of like substrate to sample. A.3.1.3 Lignin phenol characterization Lignin derived phenols were released using copper (II) oxide (CuO) oxidation method (Hedges and Mann, 1979; Ma et al. 2018). Briefly, wood samples (20-100 mg) were mixed with 0.5-1 g CuO, 200 mg ammonium iron (II) sulfate hexahydrate [Fe(NH 4 ) 2 (SO 4 ) 2 · 6H 2 O] and 20 mL of nitrogen-purged NaOH solution (2 M) in teflon-lined bombs. All bombs were flushed with nitrogen in the headsapce for 10 min and heated at 150°C for 2.5 h. The lignin oxidation products (LOPs) were spiked with a surrogate standard (ethyl vanillin), acidified to pH 1 with 6 M HCl, and kept in the dark for 1 h. After centrifugation (2500 rpm, 30 min), LOPs were liquid- liquid extracted from the clear supernatant with ethyl acetate, and concentrated under nitrogen. LOPs were derivatized with N,O-bis-(trimethylsilyl) trifluoroacetamide (BSTFA) and pyridine (70°C, 1 h) to yield trimethylsilyl (TMS) derivatives before quantification. Lignin phenols were identified and quantified on a Trace 1310 gas chromatograph coupled to an ISQ mass spectrometer (Thermo Fisher Scientific, USA) using a DB-5MS column (30 m × 0.25 mm i.d., film thickness, 0.25 μm). The oven temperature was held at 65°C for 2 min, increased from 65 to 148 300°C at a rate of 6°C min -1 with final isothermal hold at 300°C for 20 min. Helium was used as carrier gas (0.8 mL min −1 ). The mass spectrometer was operated in the electron impact mode (EI) at 70 eV and scanned from 50 to 650 daltons. Quantification was achieved by comparing with surrogate standards to account for compound loss during extraction procedures. External quantification standards were used to normalize the response factor for different lignin phenols separately. Vanillyl (vanillin, acetovanillone, vanillic acid), syringyl (syringaldehyde, acetosyringone, syringic acid), and cinnamyl (p-coumaric acid, ferulic acid) phenols were summarized to represent lignin phenols. A.3.2 Extended supplementary data discussion A.3.2.1 Organic carbon content in wood Wood fragments yielded OC% from 3.9 to 81.4% (mean 33.5%, 1 = 14.0, n = 257) and C/N from 18 – 189 (mean 57, 1 = 24, n = 146, Table A.8). In tropical and temperate living trees, wood OC is typically within the range of 40 – 55% (3) and C/N is typically > 20 (Thomas and Martin 2012). Low OC (<40%, n=161) together with low C/N (<20, n=1) in a few buried wood fragments may be due to dilution by mineral matrix embedded within fragments. Minerals such as biotite can relatively enrich N compared to C, lowering OC% and C/N without affecting δ 13 C wood . We found a correlation between OC% and C/N (r 2 = 0.56, p<0.05), but no correlations between OC% and δ 13 C wood nor N% with δ 13 C wood (Fig. A.5). Moreover, water adsorption may be a driver of low OC without lowering of the C/N ratio from that expected from living plants or without altering δ 13 C. Measurements with CO 2 pulses <10x blank were excluded (n = 13), N 2 peaks were low for most samples (<10x blank) which may bias C/N results to high values but we did not find a strong dependence of N% and C/N to the N peak area. Measurements with N 2 pulses <2x blank were excluded (n=23). Contamination by marine OC (low C/N) is assumed to 149 be negligible within coarse grain turbiditic layers that contain minimal non-wood OC. One sample horizon (U1455C 43R 2W 16-97, 9.7 Ma) yielded evidence for charred wood with all 3 aliquots having ~80% OC. A.3.2.2 Variations in atmospheric carbon isotopic composition We do not have samples representing the MMCO and thus we expect negligible influence from  13 C atm variations to the Neogene record of buried wood. There is some variability of Neogene atmospheric composition, and such variations can be inferred from the carbon isotopic composition of benthic foraminifera with 1‰ peak in  13 C atm around the Miocene Climatic Optimum (France-Lanord et al., 2016). However, we do not have any samples from the MMCO therefore we assume that the Neogene variations have a minor influence on our samples. In contrast Pleistocene glacial cycles can drive up to 3‰ variations in  13 C atm (France-Lanord et al., 2016), though this is small compared to the 19‰ variability seen in the Pleistocene samples and other sources of variability including more C 3 and C 4 variability in plant type and shifting loci of erosion with glacier advance and retreat are expected to be the dominant variables. 150 Fig. A.1. Maps of Exp354 IODP cores analyzed in this study. (A) Map of the Himalayan erosion system showing the G – B rivers, the location of Expedition 354 transect at 8°N (boxed region). The Bengal Fan sediment isopachs (blue lines; in kilometers) are simplified from (Curray et al., 2002) and represent the total sedimentary and metasedimentary rocks above the oceanic basalt as interpreted from seismic reflection and refraction data. Colors denote elevation, lowlands (0-200 m) are indicated in shades of green. Figure modified from (France-Lanord C, et al. 2016). (B) Six cores sampled in this study overlaid on interpreted line drawing of Profile GeoB97-020/027, showing the later stage channel-levee systems, regional unconformities, and faults. Modified from (Curray et al., 2002). VE = vertical exaggeration. TW = two-way travel time. Horizontal scale = CMP, CMP distance = 20 m. Naming of channels follows (Curray et al., 2002). 151 Fig. A.2. Wood fragments (blue symbols) and barren samples (open symbols), displayed in approximate sampling location on the Exp 354 stratigraphic columns (France-Lanord et al., 2016). 152 Fig. A.3. Occurrence and isotopic composition of wood in IODP Expedition 354 Sites U1450, U1451, U1452, U1453, U1454, U1455 in the Bengal Fan (color of symbol denotes Site – see legend on (C). (A) Total wood mass per sample in log scale from two datasets: those wood fragments picked from shipboard visual inspection of split core, and fragments sieved and picked from 10 – 80 g of sediments post- expedition (B) Carbon isotopic composition of wood ( 13 C wood ). (C) Hydrogen isotopic composition of lignin methoxyl (D methoxyl ). 153 Fig. A.4. Ratios and concentrations of major lignin phenol groups in Bengal Fan buried wood fragments. Phenol concentrations of (A) vanillyl (V) group (B) syringyl (S) group and (C) cinnamyl (C) group. [V], [S], and [C] concentrations decline with age, yet the larger decrease in S compared to V results in the decline in S/V ratios with age shown in Fig. 3, interpreted in terms of progressive degradation based on the data in panels A-E here. Acid to aldehyde ratios for (D) the V (Vd/Vl) and S group phenols (Sd/Sl). Acid/aldehyde ratios increase with age revealing progressive degradation. Overall the samples are remarkably well preserved, with samples within the last 12 Myr, falling within the values <0.4 typical of fresh wood (green shading). (F) Ratio of C to V phenols (C/V). Low C/V ratios (<0.05) confirm that the samples are sourced from woody not leafy tissues. 154 Fig. A.5. Bulk organic carbon properties of Bengal Fan wood fragments: TOC vs. carbon isotopic composition. Low OC implies matrix dilution or water adsorption. Three samples with elevated TOC outside of the range of living plant tissues (~80%) likely represent charcoal. Fig. A.6. Monthly D precip plotted against elevation from a selection of Global Network of Isotopes in Precipitation (GNIP) stations in the G – B catchment (Gangotri, Gomukh, Dobrani, Maneri, Uttarkashi, Tehri, Devprayag, Rishikesh, Roorkee, Nainital, New Delhi, Lhajung, Lhasa, Lucknow, Allahabad, Patna, Dinajpur, Guwahati, Dhaka, Sylhet, Shillong, Barisal, Chuadanga, Kolkata, Satkhira – see Table A.10) (8). Grey circles indicate individual data with sizes scaled to precipitation amount. At every altitude, there is a large isotopic spread (>100 ‰) due to seasonal variability with the Indian Summer Monsoon. 155 Elevation trend in weighted mean annual precipitation D (magenta circles) is -13.7‰/km (r 2 = 0.55, p <0.0001) based on ordinary least squares linear regression (line), showing 95 % confidence interval (shading). Fig. A.7. Modern C 3 plant survey of woody branch tissue δ 13 C values across an elevation transect in Arunachal Pradesh, India and Central Nepal. We observe an overall tendency to 13 C-enrichment with elevation of +0.8‰ km -1 (r 2 = 0.22, p = 0.00). Plant types have distinct elevation ranges, and are shown for illustrative purposes, although we find no systematic difference between plant type (fern, graminoid (bamboo), gymnosperm tree, angiosperm tree/shrub), with the exception of two heather samples that are notably 13 C-enriched relative to other angiosperms at that elevation, including other shrubs. One C 4 grass at 1.5 km was excluded from the C 3 elevation trend presented here (Table A.11). 156 Fig. A.8. Box plots comparing δ 13 C of modern plant survey across an elevation transect in Arunachal Pradesh, India and Central Nepal (Table A.11) and GNIP δD precip data (Table A.10) with Exp 354 buried wood isotope results. (A) Modern plant survey of woody tissue δ 13 C values (red, by elevation in Fig. A.7) and the pre-Industrial corrected values of the data (orange) are shown with the buried wood results (brown). Box plots of pre-Industrial and buried wood are comparable with regards to the median, 1 st and 3 rd quartiles and the range. Although our survey of modern C 4 plants has been minimal, C 4 plants have clearly contributed woody material to the fan. (B) GNIP δD precip data (stations listed in Fig. A.6, Table A.10) (cyan) are shown with estimated δD precip based on δD methoxyl results (blue) using published 2  methoxyl/precip (Table A.9). Outliers are in black crosses and represent data that fall above and below 1.5 times the interquartile range from the median (outside 2.698). The methoxyl-based estimates from buried wood samples are consistent with precipitation in the catchment, although extremely D-depleted precipitation (outliers) are not represented. 157 Table A.6. Bengal Fan wood mass recovered. Table continues across multiple pages. U1450 A 4 F 1 W 96 99 21.16 0.27 0 76.245 sieved 23.26 3.3 U1450 A 4 F 1 W 100 101 21.20 0.27 0 137.206 surface pick U1450 A 8 F 1 W 125 128 40.25 0.30 0 4.570 sieved 28.39 0.2 U1450 A 14 F 1 W 38 41 67.78 0.32 0 0.000 sieved U1450 A 18 F CC W 4 7 90.94 0.39 0 0.000 sieved U1450 A 22 H 1 W 33 36 104.83 0.44 0 43.415 sieved 32.93 1.3 U1450 A 28 F 1 W 6 9 132.96 0.60 0 101.075 sieved 38.16 2.6 U1450 A 32 F 1 W 32 35 152.22 0.70 0 5.720 sieved 28.55 0.2 U1450 A 38 F 1 W 77 80 181.17 1.13* 1 0.000 sieved U1450 A 42 F 2 W 45 48 201.30 1.24* 1 0.000 sieved U1450 A 48 F 1 W 15 18 228.05 1.40 1 0.000 sieved U1450 A 54 F 1 W 3 6 256.43 1.56 1 0.000 sieved U1450 A 58 F 1 W 7 10 275.47 1.68 1 0.000 sieved U1450 A 64 F 1 W 10 13 304.00 1.86 1 48.565 sieved 42.01 1.2 U1450 A 76 F 1 W 50 53 361.40 2.24 2 0.000 sieved U1450 A 84 F 1 W 8 11 398.78 2.52 2 0.000 sieved U1450 A 86 F 1 W 20 23 408.40 2.59 2 0.000 sieved U1450 A 94 F 1 W 28 31 446.48 2.91 2 0.000 sieved U1450 A 98 F 1 W 40 43 465.60 3.08 3 0.705 sieved 42.29 0.0 U1450 A 106 F 1 W 54 57 503.74 3.44 3 3.610 sieved 42.79 0.1 U1450 A 117 F 1 W 83 86 551.73 3.95 3 13.060 sieved 23.53 0.6 U1450 A 120 X 2 W 25 28 562.57 4.08 4 2.365 sieved 32.79 0.1 U1450 A 120 X 2 W 59 59 562.91 4.08 4 55.320 surface pick U1450 A 122 X 1 W 44 44 580.04 4.29 4 20.297 surface pick U1450 A 122 X 1 W 136 136 580.96 4.30 4 78.378 surface pick U1450 A 123 X CC W 48 48 590.66 4.42 4 73.921 surface pick U1450 B 16 R 3 W 77 78 747.03 6.86 6 35.691 surface pick U1450 B 18 R 1 W 99 102 764.29 7.19 7 712.912 sieved 42.62 16.7 U1450 B 18 R 1 W 101 101 764.31 7.19 7 138.345 surface pick U1451 A 4 H 1 W 138 139 25.38 0.27 0 44.301 surface pick U1451 A 13 F 2 W 18 18 71.37 0.75 0 93.457 surface pick U1451 A 16 F 1 W 130 133 85.10 1.19* 1 0.000 sieved U1451 A 19 F 2 W 105 108 100.45 1.25* 1 0.000 sieved U1451 A 27 H 1 W 72 75 160.22 4.92 4 0.000 sieved U1451 A 31 F 1 W 50 53 198.00 6.18 6 0.000 sieved U1451 A 41 F 1 W 107 110 246.07 6.75 6 0.000 sieved U1451 A 45 F 1 W 57 59 264.57 6.96 6 0.000 sieved U1451 A 54 F 1 W 89 92 307.59 7.44 7 0.000 sieved U1451 A 56 F 2 W 3 6 317.25 7.46 7 4209.163 sieved 48.74 86.4 U1451 A 56 F 2 W 5 5 317.25 7.46 7 175.692 surface pick U1451 A 58 F 1 W 62 65 326.32 7.48 7 212.890 sieved 49.76 4.3 U1451 A 58 F 1 W 78 78 326.48 7.48 7 71.409 surface pick U1451 A 66 F 1 W 13 13 363.83 8.54 8 2.811 surface pick U1451 A 66 F 1 W 60 60 364.30 8.55 8 2.632 surface pick U1451 A 66 F 1 W 91 91 364.61 8.55 8 2.861 surface pick U1451 A 66 F 1 W 92 95 364.62 8.55 8 0.000 sieved U1451 A 74 F 2 W 61 64 403.81 8.72 8 0.000 sieved U1451 A 80 F 1 W 20 23 430.40 8.75 8 0.000 sieved U1451 A 100 F 1 W 57 60 525.77 10.22 10 0.000 sieved U1451 B 3 X 3 W 134 134 556.04 10.42 10 53.380 surface pick U1451 B 5 X CC W 32 35 571.42 10.51 10 0.000 sieved U1451 B 9 X CC W 3 3 606.93 10.74 10 3.016 surface pick Method of isolation Wood / sediment (mg/g) Sediment mass (g) Wood mass recovered (mg) Site Hole Core Type Sect A/W Myr bin Top offset (cm) Bottom offset (cm) Depth CSF-A (mbsf) Age (Ma) 158 U1451 B 11 X CC W 29 34 621.31 10.83 10 0.000 sieved U1451 B 12 X 1 W 34 34 629.64 10.88 10 1.566 surface pick U1451 B 16 R 2 W 23 26 657.11 11.06 11 0.000 sieved U1451 B 19 R 3 W 36 40 687.81 11.25 11 1788.422 sieved 52.28 34.2 U1451 B 19 R 3 W 39 39 687.84 11.25 11 53.280 surface pick U1451 B 20 R 3 W 33 36 697.96 11.29 11 10.930 sieved 30.20 0.4 U1451 B 21 R 1 W 40 43 705.10 11.49 11 226.930 sieved 17.41 13.0 U1451 B 21 R 1 W 41 41 705.01 11.49 11 59.213 surface pick U1451 B 21 R 1 W 50 54 705.00 11.49 11 21.490 sieved 21.81 1.0 U1451 B 22 R 2 W 3 9 715.93 11.80 11 19.032 surface pick U1451 B 22 R 2 W 24 30 716.14 11.81 11 14.749 surface pick U1451 B 22 R 2 W 48 52 716.38 11.82 11 19.775 sieved 36.87 0.5 U1451 B 23 R 1 W 48 48 724.68 11.86 11 49.669 surface pick U1451 B 23 R 1 W 86 86 725.06 11.86 11 236.028 surface pick U1451 B 25 R 1 W 93 97 744.75 11.87 11 347.810 sieved 77.28 4.5 U1451 B 25 R 1 W 94 94 744.74 11.87 11 83.946 surface pick U1451 B 26 R CC W 3 7 753.82 11.88 11 1.345 sieved 36.65 0.0 U1451 B 28 R 1 W 64 66 773.44 11.89 11 0.321 surface pick U1451 B 28 R 2 W 35 38 774.56 11.89 11 5.207 surface pick U1451 B 29 R 2 W 131 134 784.98 11.90 11 0.000 sieved U1451 B 31 R 2 W 55 57 803.47 12.30 12 73.425 sieved 20.11 3.7 U1451 B 31 R 2 W 56 56 803.48 12.30 12 43.496 surface pick U1451 B 32 R 1 W 36 36 811.96 12.50 12 9.433 surface pick U1451 B 33 R 1 W 45 48 821.75 12.74 12 57.615 sieved 42.89 1.3 U1451 B 33 R 2 W 11 14 822.48 12.75 12 258.320 sieved 36.32 7.1 U1451 B 34 R 1 W 68 70 831.68 13.00 13 9.375 sieved 10.97 0.9 U1451 B 34 R CC W 7 7 832.61 13.01 13 458.568 surface pick U1451 B 37 R 3 W 11 13 862.78 13.43 13 637.710 sieved 40.42 15.8 U1451 B 43 R 2 W 34 39 919.74 14.41 14 0.000 sieved U1451 B 43 R 3 W 30 30 920.50 14.42 14 50.803 surface pick U1451 B 45 R 1 W 12 15 938.12 14.74 14 8.580 sieved 27.12 0.3 U1451 B 45 R 2 W 50 50 939.34 14.76 14 15.061 surface pick U1451 B 47 R 1 W 12 13 952.12 15.45 15 0.000 sieved U1451 B 48 R 1 W 33 36 957.83 15.85 15 0.000 sieved U1451 B 49 R 1 W 10 14 967.30 16.00 16 0.000 sieved U1451 B 53 R 1 W 8 12 1006.2 16.58 16 0.083 sieved 30.44 0.0 U1451 B 53 R 1 W 20 22 1006.18 16.58 16 8.620 sieved 22.27 0.4 U1451 B 53 R CC W 0 5 1006.31 16.59 16 3893.673 surface pick U1451 B 53 R CC W 0 6 1006.31 16.6 16 557.4 surface pick U1451 B 55 R 1 W 25 28 1025.85 16.88 16 0.000 sieved U1451 B 56 R 1 W 46 49 1035.76 17.03 17 0.205 sieved 34.35 0.0 U1451 B 56 R 2 W 36 40 1036.54 17.04 17 41.785 sieved 20.96 2.0 U1451 B 57 R 2 W 108 111 1041.77 17.12 17 7.185 sieved 26.33 0.3 U1451 B 58 R 1 W 50 56 1045.50 17.18 17 382.532 sieved 74.29 5.1 U1451 B 58 R 3 W 100 102 1048.50 17.22 17 0.000 sieved U1451 B 59 R 1 W 143 144 1056.23 17.34 17 2.530 sieved 29.09 0.1 U1451 B 60 R 3 W 39 41 1068.57 17.52 17 65.055 sieved 10.95 5.9 U1451 B 60 R 4 W 63 66 1067.44 17.51 17 0.145 sieved 31.34 0.0 U1451 B 63 R 5 W 122 125 1100.09 18.43 18 88.695 sieved 35.17 2.5 U1451 B 69 R 5 W 45 48 1148.03 35.10 35 0.000 sieved U1452 B 4 H 2 W 149 149 30.20 0.3 0 7.138 surface pick U1452 C 5 H 6 W 56 56 38.06 0.3 0 62.150 surface pick U1452 C 1 H 5 W 130 130 6.80 0.2 0 119.315 surface pick U1453 A 2 H 3 W 141 141 13.41 0.2 0 32.076 surface pick U1453 A 16 F 3 W 32 32 84.62 0.4 0 31.864 surface pick U1453 A 25 F 1 W 19 19 123.79 0.5 0 54.413 surface pick U1454 B 5 F CC W A,B,C,D,E,F,G,H,I 34.43 0.05 0 5999.00 surface pick 159 Table A.7. Bengal Fan wood lignin phenol concentrations. Table continues across multiple pages. U1454 B 32 F 2 W 60 60 158.95 1.00 1 13.402 surface pick U1455 C 2 H 6 W 41 42 16.52 0.22 0 0.446 surface pick U1455 C 3 H 2 W 78 79 20.48 0.22 0 236.130 surface pick U1455 C 4 F 2 W 46 46 25.16 0.23 0 8.352 surface pick U1455 C 7 F 1 W 19 26 37.49 0.25 0 16.001 surface pick U1455 C 7 F 1 W 76 82 38.06 0.25 0 1.429 surface pick U1455 C 7 F 2 W 9 10 38.89 0.25 0 61.262 surface pick U1455 C 28 F 2 W 78 79 368.78 5.84 5 2.184 surface pick U1455 C 39 F 1 W 47 48 418.67 6.55 6 8.552 surface pick U1455 C 39 F 2 W 105 110 420.75 6.58 6 9.079 surface pick U1455 C 43 R 1 W 67 69 773.67 9.67 9 4.397 surface pick U1455 C 43 R 2 W 16 97 774.46 9.67 9 3.421 surface pick U1455 C 44 R 1 W 0 2 782.80 9.75 9 3099.687 surface pick U1455 C 44 R 1 W 4 4 782.84 9.75 9 28.386 surface pick U1455 C 44 R 1 W 4 8 782.84 9.75 9 1.295 sieved 30.67 0.0 U1455 C 44 R 3 W 79 86 786.52 9.78 9 9.216 surface pick U1455 C 45 R 3 W 25 25 795.68 9.87 9 2.916 surface pick U1455 C 49 R 2 W 0 90 833.21 10.27 10 65.317 surface pick U1455 C 56 R 1 W 93 94 901.23 11.27 11 8.247 surface pick U1455 C 59 R 1 W 37 38 929.77 11.83 11 204.363 surface pick U1455 C 60 R CC W 2 3 942.37 12.11 12 341.622 surface pick U1450 B 18 R 1 W 99 102 764.29 7.19 0.83 0.01 0.38 0.31 10.78 8.93 U1451 A 56 F 2 W 3 6 317.25 7.46 1.19 0.01 0.36 0.16 17.39 20.77 U1451 A 58 F 1 W 62 65 326.32 7.86 1.62 0.02 0.31 0.22 6.05 9.78 U1451 B 19 R 3 W 36 40 687.81 11.25 0.60 0.02 0.41 0.37 7.11 4.25 U1451 B 21 R 1 W 40 43 705.10 11.41 0.68 0.01 0.35 0.32 13.61 9.29 U1451 B 23 R 1 W 86 86 725.06 11.55 0.70 0.01 0.88 0.78 1.48 1.04 U1451 B 25 R 1 W 93 97 744.75 11.87 0.57 0.02 0.39 0.41 6.80 3.88 U1451 B 33 R 2 W 11 14 822.48 12.50 0.46 0.01 0.41 0.40 10.91 5.01 U1451 B 34 R CC W 7 7 832.61 13.01 0.44 0.01 0.44 0.40 5.34 2.36 U1451 B 37 R 3 W 11 13 862.78 13.43 0.45 0.02 0.47 0.41 2.48 1.13 U1451 B 53 R CC W 0 5 1006.31 16.59 0.34 0.02 0.49 0.49 1.90 0.65 U1451 B 58 R 1 W 50 56 1045.50 17.18 0.38 0.01 0.53 0.49 2.44 0.93 U1454 B 5 F CC W A 0.05 0.00 0.02 0.21 0.25 29.10 0.06 U1454 B 5 F CC W B 0.05 0.00 0.01 0.29 0.16 23.77 0.08 U1454 B 5 F CC W C 0.05 4.07 0.04 0.16 0.16 17.27 70.23 U1454 B 5 F CC W D 0.05 1.68 0.02 0.19 0.22 34.27 57.66 U1454 B 5 F CC W E 0.05 0.00 0.02 0.24 0.22 32.05 0.08 U1454 B 5 F CC W F 0.05 0.00 0.01 0.37 0.20 42.62 0.16 U1454 B 5 F CC W G 0.05 0.00 0.01 0.32 0.30 43.36 0.08 U1454 B 5 F CC W H 0.05 0.00 0.02 0.29 0.21 33.57 0.08 U1454 B 5 F CC W I 0.05 0.79 0.01 0.26 0.16 33.68 26.60 U1455 C 44 R 1 W 0 2 782.80 9.75 0.51 0.00 0.29 0.33 3.78 1.92 U1455 C 60 R CC W 2 3 942.37 12.11 0.38 0.01 0.51 0.44 1.26 0.48 C/V Vd/Vl Sd/Sl V (mg/g) S (mg/g) S/V Site Hole Core Type Section W/A Top Offset (cm) Bottom Offset (cm) Depth CSF-A (mbsf) Age (Ma) 160 Compound concentrations (g/g) Bd 3-OH Bd 3,4-OH Bd DHA Vg Sg pBl pBn pBd P phenols (pBl, pBn, pBd) 81.43 253.48 ND 336.23 2549.91 1681.96 611.79 190.06 512.55 1314.40 46.51 192.05 ND 440.75 4010.28 5437.79 156.15 52.89 82.31 291.35 23.38 214.59 35.08 625.05 1147.58 891.14 491.58 195.92 394.00 1081.50 56.69 365.18 ND 441.24 1870.16 789.45 462.90 181.14 421.01 1065.05 43.29 272.69 49.90 399.01 2317.77 849.86 447.52 175.32 583.35 1206.20 13.79 83.85 11.10 102.95 209.38 11.22 34.26 21.16 49.39 104.82 85.34 368.00 ND 454.58 1682.78 690.02 760.34 247.82 609.23 1617.39 28.61 333.56 52.00 384.68 2256.97 629.81 349.47 161.08 566.40 1076.95 24.76 375.18 ND 782.27 1884.33 713.90 100.95 46.40 86.78 234.13 55.65 379.52 ND 542.55 760.19 303.10 353.24 118.65 389.12 861.02 68.49 217.95 ND 259.64 473.77 115.91 205.31 93.27 217.62 516.20 146.50 380.94 ND 588.60 637.69 193.48 417.69 162.62 473.70 1054.00 23.06 61.37 26.89 21.43 3114.27 4.80 1302.00 447.35 323.86 2073.21 54.12 127.32 ND 46.32 3615.90 5.90 992.67 371.39 192.37 1556.43 40.94 151.43 100.40 254.51 2794.05 5878.69 287.23 87.74 161.68 536.65 53.88 130.02 98.40 133.53 4943.43 5649.04 208.13 63.67 82.14 353.94 41.32 72.77 32.26 34.64 3839.70 5.96 2410.24 990.43 654.27 4054.94 57.67 116.65 16.48 28.78 4452.51 4.94 2415.47 702.25 791.98 3909.69 42.34 116.09 21.33 43.93 5425.66 5.63 2461.06 963.06 751.44 4175.56 29.76 82.42 23.81 35.51 4373.62 5.39 1633.49 618.69 425.64 2677.81 33.29 80.83 72.25 50.77 4844.31 2674.14 125.48 44.05 48.18 217.71 16.23 228.91 ND 272.03 652.64 237.77 44.74 20.02 40.72 105.48 27.47 198.46 ND 74.20 223.77 19.34 102.67 46.14 103.04 251.85 Compound concentrations (g/g) Vl Vn Vd V phenols (Vl, Vn, Vd) Sl Sn Sd S phenols (Sl, Sn, Sd) pCd Fd C phenols (pCd, Fd) 6234.93 2156.99 2390.80 10782.72 5666.85 1514.44 1751.97 8933.26 29.04 112.95 141.99 10317.09 3383.73 3689.31 17390.13 14215.07 4256.50 2296.55 20768.12 51.07 95.97 147.04 3654.31 1268.57 1125.40 6048.28 6405.97 1945.92 1431.34 9783.24 26.84 108.47 135.30 4031.30 1441.88 1639.05 7112.23 2625.79 668.53 959.14 4253.45 30.75 76.08 106.83 8015.16 2786.06 2813.09 13614.31 5646.65 1842.76 1804.81 9294.23 34.64 146.54 181.18 639.87 283.43 560.79 1484.09 485.67 173.59 377.58 1036.84 3.09 10.84 13.94 3819.89 1472.72 1502.97 6795.59 2180.86 804.06 894.66 3879.58 55.82 67.55 123.36 6050.50 2385.26 2476.70 10912.46 2808.95 1085.87 1117.52 5012.35 25.00 131.03 156.04 2945.30 1090.30 1302.26 5337.86 1424.73 367.83 563.08 2355.65 5.45 48.27 53.71 1331.21 520.35 623.65 2475.21 638.85 224.38 262.27 1125.49 13.30 28.39 41.70 986.06 434.25 480.23 1900.53 347.04 133.09 171.05 651.19 10.67 24.57 35.24 1239.59 548.52 651.01 2439.13 501.15 187.04 246.51 934.69 13.59 22.53 36.12 20361.54 4512.62 4227.38 29101.53 22.24 33.56 5.49 61.29 49.66 476.14 525.80 14382.24 5220.39 4166.03 23768.66 28.27 42.37 4.55 75.19 22.97 287.06 310.03 12446.89 2819.93 2004.16 17270.98 48594.25 13762.96 7874.45 70231.66 51.40 608.18 659.58 23766.98 5945.14 4553.24 34265.36 36568.87 13211.22 7877.54 57657.63 25.44 621.44 646.88 21490.77 5493.69 5065.32 32049.78 33.27 36.71 7.38 77.37 99.86 500.06 599.92 25999.04 7017.35 9603.57 42619.95 82.18 61.76 16.06 160.00 91.22 345.74 436.96 27366.57 7344.62 8652.49 43363.68 13.87 64.35 4.23 82.44 80.12 543.54 623.66 21428.65 5886.80 6250.55 33565.99 28.67 41.54 6.02 76.22 63.33 510.81 574.14 21805.05 6154.25 5717.76 33677.05 18722.52 4841.58 3037.37 26601.47 36.67 331.96 368.63 2336.55 761.61 680.34 3778.50 1166.53 364.81 389.13 1920.47 1.61 15.83 17.43 660.72 260.96 336.13 1257.80 281.38 78.24 124.52 484.13 4.31 6.27 10.58 161 Lignin dimer concentrations (g/g) VlVl VlVn VlVd VnVd VdVd Vm2Sl Vm2Sn Vm2Sd Sm2Sl Sm2Sn Lignin dimers 102.49 54.76 78.71 20.56 15.11 10.78 8.75 5.29 18.86 4.14 319.45 220.06 109.89 123.83 29.50 20.37 29.69 34.73 11.35 81.26 15.14 675.81 63.40 40.68 45.20 17.70 9.04 13.35 20.63 5.68 28.42 7.26 251.36 45.93 26.45 44.89 12.05 10.00 6.58 4.39 3.38 7.79 1.92 163.39 130.44 78.77 112.38 32.83 27.50 22.49 24.24 11.60 29.98 8.49 478.71 14.14 8.78 20.65 5.62 9.39 4.61 8.57 5.00 5.13 1.85 83.74 29.18 18.64 27.90 7.56 6.36 5.12 3.22 3.06 6.07 1.51 108.62 86.73 54.61 83.83 25.40 21.51 17.83 17.31 9.95 16.06 5.36 338.60 31.43 18.64 30.54 8.81 7.63 5.39 3.97 2.53 3.83 0.97 113.74 9.37 6.74 11.33 3.67 3.18 1.95 1.11 1.07 1.62 0.57 40.61 9.19 6.14 10.76 3.30 3.11 1.67 1.13 0.68 0.97 0.37 37.33 6.81 4.60 8.28 2.91 2.40 1.59 0.93 0.73 0.84 0.33 29.41 996.46 489.40 505.87 112.26 66.95 63.63 17.43 18.39 3.50 2.02 2275.90 1021.01 648.97 517.83 116.44 51.63 63.50 15.55 24.95 2.84 1.04 2463.76 182.43 77.12 72.89 16.91 8.33 15.57 92.87 5.87 241.84 38.62 752.46 573.68 259.75 224.17 46.42 22.88 40.92 134.36 9.74 213.87 44.38 1570.16 1178.24 605.81 585.07 128.77 61.04 91.88 23.27 24.63 4.69 2.11 2705.51 1639.61 942.61 1264.02 250.29 199.22 198.00 61.58 75.09 11.74 3.65 4645.82 1586.72 854.16 715.53 191.03 125.01 155.07 13.81 97.88 17.37 28.61 3785.20 1190.38 627.87 728.78 151.99 95.28 92.60 25.76 28.67 5.35 2.04 2948.73 693.54 386.24 383.76 97.23 45.49 54.48 74.31 17.92 130.60 30.55 1914.12 75.34 43.95 45.61 13.61 9.55 9.62 12.87 2.68 7.97 2.04 223.22 18.93 10.10 17.51 4.23 4.72 3.31 3.12 0.90 2.08 0.52 65.40 Ratios P/(V+S) DHA/V BCAs/V Dimers/VSC Vd/Vl Sd/Sl pBd/pBl VdVd/VlVl pCd/Fd pBl/P pBn/P pBd/P 0.07 0.03 0.00 0.02 0.38 0.31 0.84 0.15 0.26 0.47 0.14 0.39 0.01 0.03 0.00 0.02 0.36 0.16 0.53 0.09 0.53 0.54 0.18 0.28 0.07 0.10 0.01 0.02 0.31 0.22 0.80 0.14 0.25 0.45 0.18 0.36 0.09 0.06 0.01 0.01 0.41 0.37 0.91 0.22 0.40 0.43 0.17 0.40 0.05 0.03 0.00 0.02 0.35 0.32 1.30 0.21 0.24 0.37 0.15 0.48 0.04 0.07 0.01 0.03 0.88 0.78 1.44 0.66 0.29 0.33 0.20 0.47 0.15 0.07 0.01 0.01 0.39 0.41 0.80 0.22 0.83 0.47 0.15 0.38 0.07 0.04 0.01 0.02 0.41 0.40 1.62 0.25 0.19 0.32 0.15 0.53 0.03 0.15 0.01 0.01 0.44 0.40 0.86 0.24 0.11 0.43 0.20 0.37 0.24 0.22 0.02 0.01 0.47 0.41 1.10 0.34 0.47 0.41 0.14 0.45 0.20 0.14 0.02 0.01 0.49 0.49 1.06 0.34 0.43 0.40 0.18 0.42 0.31 0.24 0.02 0.01 0.53 0.49 1.13 0.35 0.60 0.40 0.15 0.45 0.07 0.00 0.00 0.08 0.21 0.25 0.25 0.07 0.10 0.63 0.22 0.16 0.07 0.00 0.00 0.10 0.29 0.16 0.19 0.05 0.08 0.64 0.24 0.12 0.01 0.01 0.00 0.01 0.16 0.16 0.56 0.05 0.08 0.54 0.16 0.30 0.00 0.00 0.00 0.02 0.19 0.22 0.39 0.04 0.04 0.59 0.18 0.23 0.13 0.00 0.00 0.08 0.24 0.22 0.27 0.05 0.20 0.59 0.24 0.16 0.09 0.00 0.00 0.11 0.37 0.20 0.33 0.12 0.26 0.62 0.18 0.20 0.10 0.00 0.00 0.09 0.32 0.30 0.31 0.08 0.15 0.59 0.23 0.18 0.08 0.00 0.00 0.09 0.29 0.21 0.26 0.08 0.12 0.61 0.23 0.16 0.00 0.00 0.00 0.03 0.26 0.16 0.38 0.07 0.11 0.58 0.20 0.22 0.02 0.07 0.01 0.04 0.29 0.33 0.91 0.13 0.10 0.42 0.19 0.39 0.14 0.06 0.02 0.04 0.51 0.44 1.00 0.25 0.69 0.41 0.18 0.41 162 Table A.8 Bengal Fan wood organic carbon concentration and carbon isotope measurements. Table continues across multiple pages. 1 2 3 mean  U1450 A 4 F 1 W 96 99 21.16 0.27 23.8 26.7 27.3 26.0 1.9 U1450 A 4 F 1 W 100 101 21.20 0.27 24.6 40.6 37.3 34.2 8.5 U1450 A 8 F 1 W 125 128 40.25 0.30 19.1 22.8 25.0 22.3 3.0 U1450 A 22 H 1 W 33 36 104.83 0.44 13.2 12.2 12.7 0.7 U1450 A 98 F 1 W 40 43 465.60 3.08 35.6 38.3 28.0 34.0 5.3 U1450 A 106 F 1 W 54 57 503.74 3.44 33.4 13.8 41.6 29.6 14.3 U1450 A 117 F 1 W 83 86 551.73 3.95 21.1 27.0 29.6 25.9 4.3 U1450 A 120 X 2 W 25 28 562.57 4.08 29.8 47.1 53.9 43.6 12.4 U1450 A 120 X 2 W 59 59 562.91 4.08 44.1 34.1 34.3 37.5 5.7 U1450 A 122 X 1 W 44 44 580.04 4.29 51.2 51.7 49.3 50.7 1.3 U1450 A 122 X 1 W 136 136 580.96 4.30 51.2 51.7 49.3 50.7 1.3 U1450 A 123 X CC W 48 48 590.66 4.42 36.7 40.4 30.4 35.9 5.1 U1450 B 16 R 3 W 77 78 747.03 6.86 52.4 55.5 54.1 54.0 1.6 U1450 B 18 R 1 W 99 102 764.29 7.19 15.8 30.6 29.5 25.3 8.2 U1450 B 18 R 1 W 101 101 764.31 7.19 32.9 43.9 37.2 38.0 5.5 U1451 A 4 H 1 W 138 139 25.38 0.27 46.6 45.6 45.3 45.8 0.7 U1451 A 13 F 2 W 18 18 71.37 0.75 46.2 41.4 44.5 44.0 2.4 U1451 A 56 F 2 W 3 6 317.25 7.46 31.8 21.6 32.6 28.7 6.2 U1451 A 56 F 2 W 5 5 317.25 7.46 42.8 26.6 42.0 37.1 9.1 U1451 A 58 F 1 W 62 65 326.32 7.48 19.6 19.6 U1451 A 58 F 1 W 78 78 326.48 7.48 19.4 20.6 18.2 19.4 1.2 U1451 A 66 F 1 W 13 13 363.83 8.54 47.3 24.1 44.5 38.6 12.7 U1451 A 66 F 1 W 60 60 364.30 8.55 32.3 39.6 43.0 38.3 5.5 U1451 A 66 F 1 W 91 91 364.61 8.55 10.2 27.2 18.7 12.0 U1451 B 3 X 3 W 134 134 556.04 10.42 34.0 44.5 38.1 38.9 5.3 U1451 B 9 X CC W 3 3 606.93 10.74 40.7 48.4 43.8 44.3 3.9 U1451 B 12 X 1 W 34 34 629.64 10.88 40.1 45.8 41.4 42.5 3.0 U1451 B 19 R 3 W 36 40 687.81 11.25 22.4 16.1 13.7 17.4 4.5 U1451 B 19 R 3 W 39 39 687.84 11.25 50.4 44.7 41.5 45.5 4.5 U1451 B 20 R 3 W 33 36 697.96 11.29 33.4 24.6 24.7 27.6 5.1 U1451 B 21 R 1 W 40 43 705.10 11.49 23.01 19.42 21.2 2.5 U1451 B 21 R 1 W 41 41 705.01 11.49 50.3 42.3 52.0 48.2 5.2 U1451 B 21 R 1 W 50 54 705.00 11.49 37.2 33.1 18.7 29.7 9.7 U1451 B 22 R 2 W 3 9 715.93 11.80 18.5 23.5 24.1 22.1 3.1 U1451 B 22 R 2 W 24 30 716.14 11.81 47.7 44.0 44.8 45.5 2.0 U1451 B 22 R 2 W 48 52 716.38 11.82 34.3 35.1 20.4 29.9 8.3 U1451 B 23 R 1 W 48 48 724.68 11.86 25.0 21.7 19.3 22.0 2.9 U1451 B 23 R 1 W 86 86 725.06 11.86 13.0 8.9 13.4 11.8 2.5 U1451 B 25 R 1 W 93 97 744.75 11.87 13.1 25.9 35.1 24.7 11.0 U1451 B 25 R 1 W 94 94 744.74 11.87 49.3 34.4 37.0 40.2 8.0 U1451 B 26 R CC W 3 7 753.82 11.88 42.3 46.2 30.2 39.6 8.3 U1451 B 28 R 1 W 64 66 773.44 11.89 43.0 43.7 45.3 44.0 1.2 U1451 B 28 R 2 W 35 38 774.56 11.89 53.3 53.5 51.4 52.7 1.1 OC wt % (relative to wood) Depth CSF-A (mbsf) Age (Ma) A/W Top offset (cm) Bottom offset (cm) Site Hole Core Type Sect 163 1 2 3 weighted mean mean s 1 2 3 mean  -22.9 -23.3 -22.6 -23.0 -23.0 0.3 sieved 0.085 -22.7 -26.2 -26.6 -25.5 -25.2 2.2 50.94 55.17 66.72 57.61 8.17 surface pick -30.4 -27.7 -29.1 -29.0 -29.1 1.3 26.9 31.4 29.3 29.17 2.24 sieved 0.004 -27.5 -23.7 -25.6 2.7 sieved 0.017 -26.3 -25.5 -25.2 -25.7 -25.7 0.5 44.8 39.8 42.31 3.55 sieved 0.001 -23.5 -22.8 -23.5 -23.4 -23.2 0.4 34.8 35.0 34.91 0.16 sieved 0.002 -23.6 -20.3 -23.5 -22.4 -22.5 1.9 37.5 31.5 34.51 4.25 sieved 0.014 -26.7 -26.9 -26.1 -26.5 -26.6 0.4 30.7 56.2 53.8 46.90 14.09 sieved 0.003 -25.7 -25.7 -26.0 -25.8 -25.8 0.2 44.48 53.78 38.85 45.71 7.54 surface pick -23.2 -23.4 -23.9 -23.5 -23.5 0.4 30.98 30.33 43.96 35.09 7.69 surface pick -21.8 -21.1 -18.9 -20.6 -20.6 1.5 80.96 64.83 75.05 73.62 8.16 surface pick -20.5 -21.1 -23.8 -21.7 -21.8 1.8 66.14 55.37 30.18 50.56 18.46 surface pick -27.7 -27.3 -27.4 -27.5 -27.5 0.2 54.87 71.46 93.80 73.38 19.54 surface pick -29.9 -27.8 -28.0 -28.3 -28.6 1.2 46.1 46.07 sieved 0.423 -28.9 -27.9 -28.1 -28.3 -28.3 0.6 33.71 45.01 65.32 48.01 16.01 surface pick -26.0 -26.0 -26.2 -26.0 -26.0 0.1 97.87 62.15 87.75 82.59 18.41 surface pick -28.4 -25.4 -26.2 -26.7 -26.7 1.5 58.14 66.76 81.87 68.92 12.01 surface pick -27.3 -27.9 -27.7 -27.6 -27.7 0.3 66.7 66.69 sieved 2.475 -25.0 -25.0 -25.1 -25.0 -25.0 0.0 56.68 110.22 83.45 37.86 surface pick -27.5 -27.5 -27.5 sieved 0.084 -25.7 -25.6 -25.7 -25.6 -25.6 0.1 35.02 38.77 36.89 2.65 surface pick -25.3 -25.7 -25.0 -25.3 -25.3 0.3 53.38 85.50 69.44 22.72 surface pick -26.2 -26.4 -23.8 -25.4 -25.5 1.4 55.20 61.32 39.88 52.13 11.04 surface pick -28.5 -28.2 -28.3 -28.3 0.3 31.13 31.13 surface pick -26.9 -26.8 -27.2 -26.9 -26.9 0.2 48.47 44.02 47.97 46.82 2.44 surface pick -27.1 -26.4 -26.7 -26.7 -26.7 0.4 49.34 77.89 62.96 63.40 14.28 surface pick -26.1 -24.8 -27.0 -25.9 -26.0 1.1 49.25 45.81 53.85 49.64 4.03 surface pick -29.6 -26.8 -27.7 -28.2 -28.0 1.4 33.1 34.3 33.72 0.82 sieved 0.595 -26.3 -26.3 -27.0 -26.5 -26.5 0.4 87.04 60.30 54.66 67.33 17.30 surface pick -26.1 -26.2 -26.4 -26.2 -26.2 0.2 31.5 32.9 29.1 31.16 1.92 sieved 0.010 -27.2 -25.9 -26.6 -26.5 0.9 44.8 44.82 sieved 0.277 -26.9 -27.2 -26.9 -27.0 -27.0 0.2 82.10 50.35 95.51 75.99 23.19 surface pick -27.7 -28.3 -28.5 -28.1 -28.2 0.5 50.6 50.9 30.7 44.10 11.60 sieved 0.029 -25.9 -26.1 -25.6 -25.9 -25.9 0.3 38.50 39.83 46.80 41.71 4.46 surface pick -26.7 -25.9 -25.6 -26.1 -26.1 0.6 51.13 72.77 66.37 63.42 11.12 surface pick -24.9 -25.3 -25.6 -25.2 -25.3 0.4 58.5 52.9 55.70 3.99 sieved 0.016 -26.5 -26.4 -26.8 -26.6 -26.6 0.2 45.78 40.53 41.43 42.58 2.81 surface pick -24.4 -24.5 -24.4 -24.4 -24.4 0.0 36.11 17.68 36.96 30.25 10.90 surface pick -26.8 -26.8 -27.3 -27.0 -27.0 0.3 28.3 36.1 43.8 36.02 7.75 sieved 0.111 -25.5 -25.2 -25.6 -25.4 -25.4 0.2 100.25 63.08 71.07 78.13 19.56 surface pick -27.5 -27.7 -24.9 -26.9 -26.7 1.6 56.5 52.5 54.48 2.80 sieved 0.001 -26.1 -24.8 -24.7 -25.2 -25.2 0.8 surface pick -27.0 -26.9 -26.6 -26.9 -26.9 0.2 80.45 71.09 68.83 73.45 6.16 surface pick C/N  13 C wood (‰) OC % (relative to sed) Method of isolation 164 U1451 B 31 R 2 W 55 57 803.47 12.30 34.6 34.2 14.4 27.7 11.5 U1451 B 31 R 2 W 56 56 803.48 12.30 4.6 3.9 5.8 4.8 1.0 U1451 B 32 R 1 W 36 36 811.96 12.50 21.9 13.1 9.3 14.8 6.5 U1451 B 33 R 1 W 45 48 821.75 12.74 43.7 46.3 45.0 1.9 U1451 B 33 R 2 W 11 14 822.48 12.75 18.0 15.9 17.0 1.5 U1451 B 34 R 1 W 68 70 831.68 13.00 22.2 56.0 15.5 31.2 21.7 U1451 B 34 R CC W 7 7 832.61 13.01 22.5 18.4 20.4 2.9 U1451 B 37 R 3 W 11 13 862.78 13.43 23.8 26.8 25.3 2.1 U1451 B 43 R 3 W 30 30 920.50 14.42 54.9 50.9 46.2 50.6 4.4 U1451 B 45 R 1 W 12 15 938.12 14.74 38.2 36.0 40.0 38.1 2.0 U1451 B 45 R 2 W 50 50 939.34 14.76 43.0 46.4 44.7 44.7 1.7 U1451 B 53 R 1 W 8 12 1006.2 16.58 36.0 36.0 U1451 B 53 R 1 W 20 22 1006.18 16.58 34.3 37.7 19.0 30.4 10.0 U1451 B 53 R CC W 0 5 1006.31 16.59 40.9 23.0 31.9 12.6 U1451 B 53 R CC W 0 6 1006.3 16.6 10.7 13.1 18.1 13.9 3.8 U1451 B 56 R 1 W 46 49 1035.76 17.03 49.4 13.5 31.5 25.3 U1451 B 56 R 2 W 36 40 1036.54 17.04 21.4 21.2 22.4 21.7 0.6 U1451 B 57 R 2 W 108 111 1041.77 17.12 44.2 39.9 42.1 3.0 U1451 B 58 R 1 W 50 56 1045.50 17.18 17.1 23.9 10.4 17.2 6.8 U1451 B 59 R 1 W 143 144 1056.23 17.34 31.4 27.8 31.4 30.2 2.0 U1451 B 60 R 3 W 39 41 1068.57 17.52 12.0 18.9 29.6 20.2 8.8 U1451 B 60 R 4 W 63 66 1067.44 17.51 36.5 36.5 U1451 B 63 R 5 W 122 125 1100.09 18.43 20.5 24.0 26.3 23.6 2.9 U1452 B 4 H 2 W 149 149 30.20 0.3 37.2 33.5 32.9 34.5 2.4 U1452 C 5 H 6 W 56 56 38.06 0.3 46.3 45.7 46.4 46.1 0.4 U1452 C 1 H 5 W 130 130 6.80 0.2 9.8 5.3 7.6 3.2 U1453 A 2 H 3 W 141 141 13.41 0.2 27.0 34.5 56.8 39.4 15.5 U1453 A 16 F 3 W 32 32 84.62 0.4 35.4 43.5 39.0 39.3 4.1 U1453 A 25 F 1 W 19 19 123.79 0.5 23.8 18.5 17.7 20.0 3.3 U1454 B 5 F CC A 34.43 0.05 U1454 B 5 F CC B 0.05 U1454 B 5 F CC C 0.05 U1454 B 5 F CC D 0.05 U1454 B 5 F CC E 0.05 U1454 B 5 F CC F 0.05 U1454 B 5 F CC G 0.05 U1454 B 5 F CC H 0.05 U1454 B 5 F CC I 0.05 U1454 B 32 F 2 W 60 60 158.95 1.00 36.2 25.7 38.6 33.5 6.9 U1455 C 2 H 6 W 41 42 16.52 0.22 24.2 18.8 21.5 3.8 U1455 C 3 H 2 W 78 79 20.48 0.22 15.9 24.9 20.4 6.4 U1455 C 4 F 2 W 46 46 25.16 0.23 47.1 46.0 46.4 46.5 0.5 U1455 C 7 F 1 W 19 26 37.49 0.25 41.9 10.5 37.6 30.0 17.0 U1455 C 7 F 1 W 76 82 38.06 0.25 46.2 45.5 13.7 35.1 18.5 U1455 C 7 F 2 W 9 10 38.89 0.25 45.1 37.7 44.9 42.6 4.2 U1455 C 28 F 2 W 78 79 368.78 5.84 37.6 42.9 28.9 36.4 7.1 U1455 C 39 F 1 W 47 48 418.67 6.55 41.4 45.1 42.1 42.9 2.0 U1455 C 39 F 2 W 105 110 420.75 6.58 7.3 37.7 37.7 21.5 U1455 C 43 R 1 W 67 69 773.67 9.67 38.1 46.1 47.5 43.9 5.1 U1455 C 43 R 2 W 16 97 774.46 9.67 78.8 79.9 81.4 80.1 1.3 U1455 C 44 R 1 W 0 2 782.80 9.75 19.4 31.7 20.3 23.8 6.8 U1455 C 44 R 1 W 4 4 782.84 9.75 56.3 56.2 54.9 55.8 0.8 U1455 C 44 R 1 W 4 8 782.84 9.75 42.0 19.0 30.8 30.6 11.5 U1455 C 44 R 3 W 79 86 786.52 9.78 35.1 39.5 20.7 31.8 9.8 U1455 C 45 R 3 W 25 25 795.68 9.87 33.6 38.7 32.3 34.9 3.4 U1455 C 49 R 2 W 0 90 833.21 10.27 51.3 45.1 51.3 49.2 3.6 U1455 C 56 R 1 W 93 94 901.23 11.27 46.0 48.6 43.5 46.0 2.6 U1455 C 59 R 1 W 37 38 929.77 11.83 15.6 12.1 13.9 2.5 U1455 C 60 R CC W 2 3 942.37 12.11 9.3 10.6 10.0 0.9 165 -26.5 -26.9 -27.3 -26.8 -26.9 0.4 49.8 53.0 29.6 44.14 12.73 sieved 0.101 -28.0 -27.9 -28.6 -28.2 -28.1 0.4 surface pick -27.3 -27.1 -27.2 -27.2 -27.2 0.1 37.24 24.04 26.04 29.11 7.11 surface pick -26.6 -26.8 -26.7 -26.7 0.2 36.4 36.42 sieved 0.060 -26.5 -26.2 -26.4 -26.4 0.2 sieved 0.121 -28.3 -24.6 -27.0 -25.9 -26.6 1.9 41.1 58.2 49.62 12.08 sieved 0.027 -27.0 -27.0 -27.0 -27.0 0.0 54.91 52.99 53.95 1.36 surface pick -27.2 -26.9 -27.0 -27.0 0.2 33.6 30.4 32.02 2.28 sieved 0.400 -24.9 -24.5 -24.3 -24.6 -24.5 0.3 78.87 73.15 95.51 82.51 11.62 surface pick -26.6 -25.9 -25.9 -26.1 -26.1 0.4 34.6 39.5 36.8 36.98 2.48 sieved 0.012 -25.1 -25.2 -25.1 -25.1 -25.1 0.1 43.50 42.84 49.15 45.16 3.47 surface pick -27.2 -27.2 -27.2 33.9 33.94 sieved 0.000 -25.3 -25.9 -26.4 -25.8 -25.9 0.6 45.7 48.4 47.08 1.91 sieved 0.012 -24.4 -26.7 -25.2 -25.5 1.7 33.13 31.80 32.46 0.95 surface pick -26.1 -26.1 -25.7 -26.0 -26.0 0.2 30.86 31.63 31.24 0.55 surface pick -26.8 -26.7 -26.8 -26.7 0.1 33.9 33.91 sieved 0.000 -26.1 -24.9 -26.1 -25.7 -25.7 0.7 21.9 30.0 25.3 25.75 4.05 sieved 0.043 -24.1 -24.3 -24.2 -24.2 0.1 49.6 44.9 47.28 3.31 sieved 0.011 -28.4 -25.1 -26.2 -26.5 -26.6 1.7 27.0 26.99 sieved 0.088 -28.9 -27.1 -26.6 -27.5 -27.5 1.2 28.6 42.2 39.4 36.71 7.19 sieved 0.003 -27.4 -25.6 -28.2 -27.2 -27.1 1.3 40.1 43.3 41.66 2.28 sieved 0.120 -27.7 -27.7 -27.7 41.6 41.60 sieved 0.000 -27.7 -26.1 -25.7 -26.4 -26.5 1.1 32.3 27.5 25.6 28.45 3.49 sieved 0.060 -27.6 -26.7 -24.0 -26.2 -26.1 1.9 49.78 30.25 42.05 40.69 9.83 surface pick -26.5 -26.6 -26.8 -26.6 -26.6 0.1 108.39 105.24 92.08 101.91 8.65 surface pick -14.4 -13.9 -14.2 -14.1 0.4 surface pick -24.3 -26.3 -27.0 -26.2 -25.9 1.4 42.07 50.36 49.16 47.20 4.48 surface pick -25.1 -24.4 -25.3 -24.9 -24.9 0.5 99.41 84.72 92.06 10.38 surface pick -28.2 -28.4 -27.9 -28.2 -28.2 0.2 49.56 27.11 37.29 37.98 11.24 surface pick -26.1 -25.9 -26.0 0.2 surface pick -26.2 -26.0 -26.1 0.1 surface pick -28.3 -27.9 -28.1 0.3 surface pick -29.0 -29.5 -29.3 0.3 surface pick -25.6 -25.7 -25.7 0.1 surface pick -26.1 -26.2 -26.2 0.1 surface pick -25.6 -25.6 -25.6 0.0 surface pick -25.8 -25.9 -25.8 0.1 surface pick -25.9 -26.0 -26.0 0.1 surface pick -29.3 -29.1 -28.6 -29.0 -29.0 0.4 63.39 37.20 58.38 52.99 13.90 surface pick -11.5 -11.4 -11.5 -11.5 0.1 39.49 45.31 42.40 4.12 surface pick -13.5 -12.6 -12.9 -13.0 0.6 31.59 52.27 41.93 14.63 surface pick -12.1 -12.3 -12.1 -12.2 -12.2 0.1 57.16 52.82 57.26 55.75 2.54 surface pick -13.8 -15.2 -14.1 -14.1 -14.3 0.7 48.87 23.54 46.67 39.69 14.03 surface pick -13.4 -13.4 -26.2 -15.1 -17.7 7.4 59.84 56.79 58.32 2.16 surface pick -12.7 -13.2 -13.3 -13.1 -13.1 0.3 75.32 55.63 53.58 61.51 12.01 surface pick -27.0 -26.5 -26.5 -26.7 -26.7 0.3 58.59 50.36 50.00 52.98 4.86 surface pick -25.3 -25.0 -25.7 -25.3 -25.3 0.4 78.15 99.46 88.80 15.07 surface pick -27.5 -27.1 -27.2 -27.3 0.2 56.95 56.95 surface pick -26.9 -26.4 -26.8 -26.7 -26.7 0.3 43.96 67.01 58.59 56.52 11.66 surface pick -29.0 -28.3 -28.2 -28.5 -28.5 0.4 188.58 188.6 surface pick -29.7 -28.9 -29.3 -29.2 -29.3 0.4 surface pick -25.3 -25.5 -25.7 -25.5 -25.5 0.2 103.74 99.98 96.13 99.95 3.80 surface pick -28.6 -27.2 -26.2 -27.5 -27.3 1.2 45.6 46.3 45.94 0.52 sieved 0.001 -27.2 -25.8 -25.9 -26.3 -26.3 0.8 35.88 58.21 40.46 44.85 11.80 surface pick -26.8 -26.7 -27.0 -26.8 -26.8 0.2 29.20 46.08 32.90 36.06 8.87 surface pick -26.5 -25.8 -25.7 -26.0 -26.0 0.4 59.41 79.83 99.86 79.70 20.23 surface pick -28.8 -28.4 -25.1 -27.5 -27.4 2.1 59.27 59.12 70.32 62.90 6.42 surface pick -26.5 -26.4 -26.5 -26.4 0.1 34.51 21.46 27.98 9.23 surface pick -27.7 -27.3 -27.5 -27.5 0.3 20.89 22.63 21.76 1.24 surface pick 166 Table A.9. Bengal Fan wood lignin methoxyl hydrogen isotope measurements Site Hole Core Type Section W/A Top offset (cm) Bottom offset (cm) Aliquot Depth CSF-A (mbsf) Age (Ma) D methoxyl (‰)  n estimated D precip (‰) U1450 A 4 F 1 W 100 101 21.16 0.271 -207 3 4 12 U1450 A 120 X 2 W 59 59 562.91 4.084 -272 0 3 -71.0 U1450 A 122 X 1 W 44 44 580.04 4.289 -236 0 3 -25.6 U1450 A 122 X 1 W 136 136 580.96 4.301 -232 1 2 -20.2 U1450 A 123 X CC W 48 48 590.66 4.422 -231 1 3 -19.5 U1450 B 16 R 3 W 77 78 747.03 6.859 -272 1 3 -71.2 U1450 B 18 R 1 W 99 102 764.29 7.192 -272 2 3 -71.0 U1450 B 18 R 1 W 101 101 764.31 7.192 -266 1 3 -63.5 U1451 A 4 H 1 W 138 139 25.38 0.266 -261 1 3 -56.8 U1451 A 13 F 2 W 18 18 71.37 0.753 -249 1 3 -41.6 U1451 A 56 F 2 W 5 5 317.25 7.460 -278 1 3 -79.3 U1451 A 56 F 2 W 3 6 317.25 7.460 -274 0 2 -73.7 U1451 A 58 F 1 W 62 65 326.32 7.482 -236 4 6 -25.4 U1451 A 58 F 1 W 78 78 326.48 7.482 -287 1 3 -90.5 U1451 B 3 X 3 W 134 134 556.04 10.416 -250 0 3 -43.0 U1451 B 19 R 3 W 36 40 687.81 11.253 -267 2 6 -64.5 U1451 B 21 R 1 W 50 54 705.00 11.491 -243 1 3 -34.5 U1451 B 21 R 1 W 41 41 705.01 11.492 -271 0 3 -69.7 U1451 B 21 R 1 W 40 43 705.10 11.494 -262 5 5 -58.7 U1451 B 22 R 2 W 3 9 715.93 11.804 -284 4 3 -86.6 U1451 B 22 R 2 W 48 52 716.38 11.817 -254 1 3 -48.7 U1451 B 23 R 1 W 48 48 724.68 11.856 -250 1 3 -43.5 U1451 B 23 R 1 W 86 86 725.06 11.856 -262 2 3 -58.6 U1451 B 25 R 1 W 94 94 744.74 11.870 -228 1 3 -15.5 U1451 B 25 R 1 W 93 97 744.75 11.870 -234 2 2 -23.5 U1451 B 31 R 2 W 55 57 803.47 12.300 -242 1 3 -32.7 U1451 B 33 R 1 W 45 48 821.75 12.737 -277 0 3 -78.3 U1451 B 33 R 2 W 11 14 822.48 12.754 -252 1 4 -46.1 U1451 B 34 R CC W 7 7 832.61 13.013 -274 0 3 -73.6 U1451 B 37 R 3 W 11 13 862.78 13.430 -218 2 6 -2.7 U1451 B 43 R 3 W 30 30 920.50 14.422 -233 1 3 -21.1 U1451 B 45 R 2 W 50 50 939.34 14.758 -183 4 3 42.2 U1451 B 53 R CC W 0 5 1006.31 16.585 -261 2 2 -57.5 U1451 B 53 R CC W 0 6 1006.31 16.585 -260 4 3 -56.4 U1451 B 56 R 2 W 36 40 1036.54 17.041 -274 1 3 -74.1 U1451 B 58 R 1 W 50 56 1045.50 17.176 -274 0 3 -73.7 U1451 B 60 R 3 W 39 41 1068.57 17.523 -267 1 3 -65.3 U1452 C 5 H 6 W 56 56 38.06 0.286 -268 2 3 -66.9 U1453 A 2 H 3 W 141 141 13.41 0.234 -226 1 3 -13.1 U1453 A 16 F 3 W 32 32 84.62 0.430 -278 1 3 -79.4 U1453 A 25 F 1 W 19 19 123.79 0.493 -213 3 3 4.4 U1454 B 5 F CC W A 0.050 -295 2 3 -101.1 U1454 B 5 F CC W B 0.050 -294 3 3 -99.1 U1454 B 5 F CC W C 0.050 -209 4 3 8.7 U1454 B 5 F CC W D 0.050 -219 3 3 -3.3 U1454 B 5 F CC W E 0.050 -294 4 3 -99.6 U1454 B 5 F CC W F 0.050 -300 2 3 -107.8 U1454 B 5 F CC W G 0.050 -290 0 3 -94.7 U1454 B 5 F CC W H 0.050 -270 2 4 -68.7 U1454 B 5 F CC W I 0.050 -213 2 3 4.1 U1454 B 32 F 2 W 60 60 158.95 1.000 -175 4 3 52.3 U1455 C 3 H 2 W 78 79 20.48 0.222 -163 4 3 67.9 U1455 C 7 F 1 W 19 26 37.49 0.248 -247 2 3 -39.5 U1455 C 7 F 2 W 9 10 38.89 0.249 -230 1 3 -18.0 U1455 C 44 R 1 W 0 2 782.80 9.749 -274 1 4 -74.1 U1455 C 44 R 1 W 4 4 782.84 9.749 -300 0 3 -107.2 U1455 C 49 R 2 W 0 90 833.21 10.272 -275 1 3 -74.7 U1455 C 59 R 1 W 37 38 929.77 11.833 -259 1 3 -54.9 U1455 C 60 R CC W 2 3 942.37 12.113 -271 3 3 -70.1 167 Table A.10 Kml spreadsheet of locations referenced in this study. Table continues across multiple pages. Latitude Longitude Name Icon Icon scale 27.8950 86.8250 LHAJUNG 156 1 30.9261 78.9403 GOMUKH 156 1 29.7000 91.1333 LHASA 156 1 30.9967 78.9403 GANGOTRI 156 1 30.9461 78.6881 DOBRANI 156 1 29.4000 79.4600 NAINITAL 156 1 25.5700 91.8800 SHILLONG 156 1 30.7453 78.4467 MANERI 156 1 30.7292 78.4467 UTTARKASHI 156 1 30.3528 78.4833 TEHRI 156 1 30.1406 78.5967 DEVPRAYAG 156 1 30.1122 78.3025 RISHIKESH 156 1 29.8678 77.8939 ROORKEE 156 1 28.5800 77.2000 NEW DELHI 156 1 26.8747 80.9389 LUCKNOW 156 1 25.4500 81.7333 ALLAHABAD 156 1 25.5736 85.0703 PATNA 156 1 26.1908 91.7953 GUWAHATI 156 1 25.6200 88.6600 DINAJPUR 156 1 23.9528 90.2792 DHAKA 156 1 22.7978 88.3717 KOLKATA 156 1 22.7000 90.3600 BARISAL 156 1 22.7167 89.0833 SATKHIRA 156 1 24.9100 91.8453 SYLHET 156 1 23.6400 88.8500 CHUADANGA 156 1 8.0070 87.6708 U1450 91 0.5 8.0070 88.7417 U1451 91 0.5 8.0070 87.1817 U1452 91 0.9 8.0070 86.7983 U1453 91 0.9 8.0068 85.8498 U1454 91 0.9 8.0070 86.2833 U1455 91 0.5 27.097 92.583 Arunachal Pradesh (India) modern plant transect (low end) 186 0.8 27.708 91.718 Arunachal Pradesh (India) modern plant transect (high end) 186 0.8 28.10 85.19 Central Nepal modern plant transect (low end) 186 0.8 28.819 83.849 Central Nepal modern plant transect (high end) 186 0.8 24.044 89.040 Lowland C4 plant 186 0.8 168 Description LHAJUNG

Altitude = 4420 m
Sample period = 1974 - 1975
GOMUKH

Altitude = 3800 m
Sample period = 2004 - 2006
LHASA

Altitude = 3649 m
Sample period = 1986 - 1992
GANGOTRI

Altitude = 3053 m
Sample period = 2004 - 2006
DOBRANI

Altitude = 2050 m
Sample period = 2004 - 2006
NAINITAL

Altitude = 1953 m
Sample period = 1995
SHILLONG

Altitude = 1598 m
Sample period = 1969 - 1978
MANERI

Altitude = 1150 m
Sample period = 2005 - 2006
UTTARKASHI

Altitude = 1140 m
Sample period = 2004 - 2006
TEHRI

Altitude = 640 m
Sample period = 2004 - 2006
DEVPRAYAG

Altitude = 465 m
Sample period = 2004 - 2006
RISHIKESH

Altitude = 356 m
Sample period = 2005 - 2006
ROORKEE

Altitude = 274 m
Sample period = 2003 - 2006
NEW DELHI

Altitude = 212 m
Sample period = 1961 - 2007
LUCKNOW

Altitude = 128 m
Sample period = 2003 - 2004
ALLAHABAD

Altitude = 98 m
Sample period = 1980
PATNA

Altitude = 60 m
Sample period = 2003 - 2005
GUWAHATI

Altitude = 54 m
Sample period = 2003 - 2004
DINAJPUR

Altitude = 35 m
Sample period = 2014 - 2016
DHAKA

Altitude = 14 m
Sample period = 2009 - 2016
KOLKATA

Altitude = 6 m
Sample period = 2004 - 2006
BARISAL

Altitude = 10 m
Sample period = 2013 - 2016
SATKHIRA

Altitude = 10 m
Sample period = 2015 - 2016
SYLHET

Altitude = 20 m
Sample period = 2009 -2016
CHUADANGA

Altitude = 20 m
Sample period = 2014 - 2016
U1450

Depth recovered = 1499.3 m
Bottom age = Late Miocene
Core recovery = 51%
U1451

Depth recovered = 1763.4 m
Bottom age = Late Oligocene
Core recovery = 51%
U1452

Depth recovered = 267 m
Bottom age = Late Pleistocene
Core recovery = 80%
U1453

Depth recovered = 215.7 m
Bottom age = Early Pleistocene
Core recovery = 88%
U1454

Depth recovered = 243.6 m
Bottom age = Pleistocene
Core recovery = 84%
U1455

Depth recovered = 956.8 m
Bottom age = Middle Miocene
Core recovery = 57%
169 Altitude (m) Sample period Depth recovered (m) Bottom age Core recovery mean stem  13 C 1 n 4420 1974 - 1975 3800 2004 - 2006 3649 1986 - 1992 3053 2004 - 2006 2050 2004 - 2006 1953 1995 1598 1969 - 1978 1150 2005 - 2006 1140 2004 - 2006 640 2004 - 2006 465 2004 - 2006 356 2005 - 2006 274 2003 - 2006 212 1961 - 2007 128 2003 - 2004 98 1980 60 2003 - 2005 54 2003 - 2004 35 2014 - 2016 14 2009 - 2016 6 2004 - 2006 10 2013 - 2016 10 2015 - 2016 20 2009 -2016 20 2014 - 2016 1499.3 Late Miocene 51 1763.4 Late Oligocene 51 267 Late Pleistocene 80 215.7 Early Pleistocene 88 243.6 Pleistocene 84 956.8 Middle Miocene 57 -28.6 1.6 41 -24.3 5.1 15 -24.3 5.1 15 170 Table A.11 Arunachal Pradesh (India) and Central Nepal elevation transects to survey of woody plant carbon isotopic composition. Table continues across multiple pages. Sample name Country Latitude (°N) Longitude (°E) Altitude (km) Wood  13 C (‰) Plant part Arunachal Pradesh (India) Elevation Transect AR 106 India 27.097 92.583 0.49 -29.6 stem AR 106 India 27.097 92.583 0.49 -29.0 branch AR 106 India 27.097 92.583 0.545 -32.8 branch AR 106 India 27.097 92.583 0.545 -27.6 stem AR 106 India 27.097 92.583 0.545 -27.0 branch AR 105 India 27.106 92.556 0.842 -31.4 branch AR 105 India 27.106 92.556 0.842 -27.1 branch AR 34_2 India 27.193 92.574 1.061 -29.3 branch AR 38 India 27.193 92.574 1.061 -31.8 branch AR 103 India 27.204 92.561 1.128 -28.8 branch AR 103 India 27.221 93.577 1.128 -28.5 branch AR 34 India 27.163 92.560 1.154 -29.6 stem AR 104 India 27.176 92.571 1.478 -29.0 branch AR 104 India 27.176 92.571 1.478 -28.3 branch AR 101 India 27.206 92.384 1.488 -28.1 branch AR 50 India 27.376 92.235 1.58 -29.4 branch AR 50_2 India 27.376 92.235 1.58 -30.7 branch AR 80 India 27.634 91.720 1.58 -29.5 branch AR 80 India 27.634 91.720 1.58 -29.5 branch AR 98 India 27.204 92.414 1.6 -26.6 branch AR 41 India 27.371 92.236 1.626 -29.6 branch AR 76 India 27.590 91.701 2.085 -27.2 branch AR 60 India 27.580 91.972 2.2 -29.2 stem AR 68 India 27.580 91.792 2.305 -28.4 stem AR 68 India 27.580 91.792 2.305 -27.5 branch AR 82 India 27.708 91.718 2.382 -29.7 branch AR 63 India 27.572 91.873 2.529 -26.1 branch AR 96_2 India 27.448 92.109 2.59 -27.4 branch AR 62 India 27.578 91.863 2.68 -29.8 stem AR 96 India 27.447 92.120 2.85 -27.1 branch AR 57 India 27.571 91.989 2.88 -29.7 stem AR 92 India 27.458 92.113 3.132 -29.1 branch AR 92 India 27.458 92.113 3.132 -29.0 branch AR 91 India 27.475 92.110 3.527 -28.9 branch AR 90 India 27.484 92.109 3.7 -27.5 branch AR 89 India 27.493 92.105 4.013 -26.9 branch AR 55 India 27.515 92.091 4.043 -27.7 branch AR 55 India 27.531 93.108 4.043 -27.4 branch AR 56 India 27.515 92.091 4.043 -27.3 branch AR 87 India 27.495 92.105 4.05 -25.1 branch Central Nepal Elevation Transect PB 36 Nepal 28.475 83.637 1.15 -27.1 branch "Bamboo" 1500 Nepal 28.10 85.19 1.5 -12.7 stem Tsilaone 1500 Nepal 28.10 85.19 1.5 -29.0 branch PB 21 Pine Nepal 28.590 83.647 1.93 -25.4 branch Bamboo 2000-2500 Nepal 28.14 85.19 2.25 -24.5 stem PB11 Pine Nepal 28.718 83.664 2.6 -22.4 branch Bamboo 3000 Nepal 28.17 85.19 3 -28.7 stem Rhodo 3000 Nepal 28.17 85.19 3 -26.0 branch Bamboo 3500 Nepal 28.20 85.20 3.5 -28.9 stem Heather 3500 Nepal 28.20 85.20 3.5 -25.1 stem PB 4 Nepal 28.819 83.849 3.5 -26.8 branch Rhodo 3500 Nepal 28.20 85.20 3.5 -26.7 branch Heather 4000 Nepal 28.22 85.18 4 -23.9 stem Rhodo 4000 Nepal 28.22 85.18 4 -25.4 branch Bangladesh, River Bank Sample BR 423 Bangladesh 24.044 89.040 0.02 -12.4 stem 171 Family/Genus/Species/Common name if known Lifeform Classification Photosynthetic pathway Notes bamboo grass monocot C3 fern non-vascular C3 fern non-vascular C3 bamboo grass monocot C3 fern non-vascular C3 100m further south fern non-vascular C3 tree angiosperm/dicot C3 fern non-vascular C3 fern non-vascular C3 on terrace, 50m from the river Rhododendron spp. shrub/tree angiosperm/dicot C3 fern non-vascular C3 bamboo grass monocot C3 in very small valleys of the river shrub/tree angiosperm C3 fern non-vascular C3 Pinus wallichiana tree gymnosperm C3 small tree 50m from the river Pinus wallichiana tree gymnosperm C3 1m above the spring fern non-vascular C3 fern non-vascular C3 fern Rhododendron spp. shrub/tree angiosperm/dicot C3 Pinus wallichiana tree gymnosperm C3 tree angiosperm/dicot C3 young growth on side at 45⁰ Rhododendron spp. shrub/tree angiosperm/dicot C3 bamboo grass monocot C3 bamboo in talweg bamboo grass monocot C3 fern non-vascular C3 Pinus wallichiana tree gymnosperm C3 on the valley slope, 200-300 m above Zemithang Pinus wallichiana tree gymnosperm C3 on the valley slope Pinus wallichiana tree gymnosperm C3 bamboo grass monocot C3 herbs in talweg Pinus wallichiana tree gymnosperm C3 bamboo grass monocot C3 bamboo; height = 2-3 m fern non-vascular C3 tree gymnosperm C3 Pinus wallichiana tree gymnosperm C3 5m tall Pinus wallichiana tree gymnosperm C3 Pinus wallichiana tree gymnosperm C3 Pinus wallichiana tree gymnosperm C3 Sela Pass pine plantation Rhododendron spp. shrub/tree angiosperm/dicot C3 Pinus wallichiana tree gymnosperm C3 Pinus wallichiana tree gymnosperm C3 Below Sela Pass tree angiosperm/dicot C3 bamboo-like appearance grass monocot C4 tree angiosperm/dicot C3 Pinus wallichiana tree gymnosperm C3 bamboo grass monocot C3 Pinus wallichiana tree gymnosperm C3 bamboo grass monocot C3 Rhododendron spp. shrub/tree angiosperm/dicot C3 bamboo grass monocot C3 Cassiope fastigiata (heather) shrub angiosperm/dicot C3 tree angiosperm/dicot C3 Rhododendron spp. shrub/tree angiosperm/dicot C3 Cassiope fastigiata (heather) shrub angiosperm/dicot C3 Rhododendron spp. shrub/tree angiosperm/dicot C3 herb monocot/dicot? C4 172 Table A.12 Compilation of published bulk wood carbon isotopic compositions in tropical and temperate forests. Table continues below. Tropical This study Arunachal Pradesh, India and Nepal tropical to subtropical montane forest 27º N 91-93º E Ometto et al., 2006 Amazon Basin tropical rainforest 3º 06'07" S 60º 01'30" W Martinelli et al., 1998 Rondonia, Brazil tropical forest 8º 45' S 63º 23' W Ellis et al., 2011 Mekong basin tropical (various) 11º 35' N 104º 56' E von Fischer and Tieszen, 1995 Luquillo, Puerto Rico subtropical rain forest 18º 18' N 65º 47' W von Fischer and Tieszen, 1995 Luquillo, Puerto Rico lower montane rain forest 18º 18' N 65º 47' W von Fischer and Tieszen, 1995 Luquillo, Puerto Rico lower montane wet forest 18º 18' N 65º 47' W von Fischer and Tieszen, 1995 Luquillo, Puerto Rico subtropical wet forest 18º 18' N 65º 47' W Fichtler et al., 2000 multiple tropical locations tropical (various) n.a. n.a. Temperate Gori et al., 2012 SE Alps cold temperate forest 46º N 11º E Duffy et al., 2017 UK temperate 51º N 1º W º This correction accounts for the increasing depletion in  13 C atm since pre-Industrial period ( 13 C atm = 6.41) to the sampled year.  13 C atm data from 1994 to 2007 is from Mauna Loa and pre-Industrial  13 C atm is based on (Friedli et al., 1986) *The I-corr  13 Cwood shows carbon isotopic composition of wood corrected for the Suess effect, by incorporating the correction amount to the mean  13 Cwood. Location Reference Longitude Latitude Biome min max mean  n correctionº I-corr*  13 C wood (‰) ~2004 branch -32.8 -25.1 -28.6 1.6 41 1.8 -26.7 1999 - 2002 not specified n.a. n.a. -28.3 1.5 24 1.7 -26.6 ~1997 trunk -31.6 -25.4 -28.3 1.7 56 1.5 -26.7 2006 not specified -33.2 -27.2 -30.2 1.7 17 1.8 -28.3 ~1994 not specified n.a. n.a. -28.9 0.4 4 1.4 -27.5 ~1994 not specified n.a. n.a. -27.4 0.6 4 1.4 -26.0 ~1994 not specified n.a. n.a. -28.5 1.1 4 1.4 -27.1 ~1994 not specified n.a. n.a. -29.6 0.5 4 1.4 -28.2 <2000 trunk -30.3 -23.1 -26.1 1.5 12 1.6 -24.4 tree ring 2003 trunk n.a. n.a. -23.8 0.4 3 1.8 -22.1 tree ring 2007 trunk n.a. n.a. -24.9 0.7 1 - already corrected º This correction accounts for the increasing depletion in  13 C atm since pre-Industrial period ( 13 C atm = 6.41) to the sampled year.  13 C atm data from 1994 to 2007 is from Mauna Loa and pre-Industrial  13 C atm is based on (Friedli et al., 1986) *The I-corr  13 Cwood shows carbon isotopic composition of wood corrected for the Suess effect, by incorporating the correction amount to the mean  13 Cwood.  13 C wood (‰) tree part Sampling period 173 A.3.3 References Hedges JI, Mann DC (1979) The characterization of plant tissues by their lignin oxidation products. Geochim Cosmochim Acta 43:1803–1807. Ma T, et al. (2018) Divergent accumulation of microbial necromass and plant lignin components in grassland soils. Nat Commun 9:3480. Thomas SC, Martin AR (2012) Carbon content of tree tissues: A synthesis. Forests 3:332–352. Blum M, et al. (2018) Allogenic and Autogenic Signals in the Stratigraphic Record of the Deep- Sea Bengal Fan. Sci Rep 8:7973. France-Lanord C, et al. (2016) Proceedings of the International Ocean Discovery Program. Int Ocean Discov Progr 354. doi:10.14379/iodp/proc.354.101.2016. Curray JR, Emmel FJ, Moore DG (2002) The Bengal Fan: Morphology, geometry, stratigraphy, history and processes. Mar Pet Geol 19(10):1191–1223. Meyers PA, Ishiwatari R (1993) Lacustrine organic geochemistry-an overview of indicators of organic matter sources and diagenesis in lake sediments. Org Geochem 20:867–900. IAEA/WMO (2015) Global Network of Isotopes in Precipitation. The GNIP Database. Available via. https://nucleus.iaea.org/wiser. Available at: http://www.iaea.org/water. Friedli H, Lötscher H, Oeschger H, Siegenthaler U, Stauffer B (1986) Ice core record of the 13 C/ 12 C ratio of atmospheric CO 2 in the past two centuries. Nature 324:237–238. Ometto JPHB, et al. (2006) The stable carbon and nitrogen isotopic composition of vegetation in tropical forests of the Amazon Basin, Brazil. Biogeochemistry 79:251–274. Ellis EE, Keil RG, Ingalls AE, Richey JE, Alin SR (2012) Seasonal variability in the sources of particulate organic matter of the Mekong River as discerned by elemental and lignin analyses. J Geophys Res Biogeosciences 117:1–15. Fischer JC Von, Tieszen LL (1995) Carbon isotope characterization of vegetation and soil organic in Subtropical Forest in Luquillo, Puerto Rico. Biotropica 27:138–148. Gori Y, et al. (2013) Carbon, hydrogen and oxygen stable isotope ratios of whole wood, cellulose and lignin methoxyl groups of Picea abies as climate proxies. Rapid Commun Mass Spectrom 27:265–275. Duffy JE, et al. (2017) Short-lived juvenile effects observed in stable carbon and oxygen isotopes of UK oak trees and historic building timbers. Chem Geol 472:1–7.

Abstract (if available)

Abstract The carbon and hydrogen isotopic compositions of plant compounds are important proxies used to reconstruct past vegetation and precipitation. However, many stages between biosynthesis and sample analysis can potentially alter the original environmental isotopic signal of interest and obscure the geological implications. Experimental approaches are used in this thesis to constrain isotopic fractionations in plant wax compounds caused by changes in leaf physiology and laboratory protocols. In chapter 2, I use a genetically modified model organism to isolate the effect of changing stomatal density on the carbon and hydrogen isotopic compositions of plant nalkanes. Chapter 3 investigates the rate of hydrogen isotope exchange during phthalic acid methylation, which is an important step in the isotopic analysis of plant n-alkanoic acids. Chapters 4 and 5 focus on development and application of a relatively new biomarker approach: isotopic analysis and quantification of methoxyl groups within lignin. In chapter 4, a liquid injection method is used to quantify and determine hydrogen isotopic composition of iodomethane evolved from a suite of lignin-bearing samples. Results show that concentrations of methoxyl groups capture changes in wood diagenesis and catagenesis. Phenol and wood samples were surveyed with potential to serve as isotopic standards. Chapter 5 analyzes lignin characteristics and the carbon and hydrogen isotopic composition of bulk wood and lignin methoxyl groups of wood fragments found in the Bengal Fan spanning the last 19 million years. This thesis presents fundamental isotope biogeochemistry efforts to constrain plant biomarkers as proxies for paleoclimate and to study an overlooked wood contribution to the organic carbon burial in the global carbon cycle.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Evolution of the Indian Monsoon and rise of C₄ photosynthesis in the Miocene and Pliocene

PDF

Altitude effect on tree wood carbon isotopic composition in humid tropical forests

PDF

Antarctic climate variability from greenhouse to icehouse world

PDF

From tree tops to river runoff: tracing plant wax biomarkers across the Peruvian Andes and Amazon

PDF

Lithium isotopes as a carbon cycle proxy: examining the effects of high-temperature weathering, understanding brachiopod archives, and validating Cenozoic records

PDF

Late Pleistocene changes in winter moisture source in the coastal southwest United States

$Isotopic insights to human-impacted hydrologic systems: from urban watersheds to hydraulically fractured basins$

PDF

Isotopic insights to human-impacted hydrologic systems: from urban watersheds to hydraulically fractured basins

PDF

From deep lakes to deserts: Plio-Pleistocene paleoenvironmental transitions

PDF

Germanium and silicon isotope geochemistry in terrestrial and marine low-temperature environments

PDF

Recent variability in the hyrdological cycle of tropical Asia from oxygen isotopes of tree celulose

PDF

Paleoenvironmental and paleoecological trends leading up to the end-Triassic mass extinction event

PDF

Flowstone ideograms: deciphering the climate messages of Asian speleothems

PDF

Carbonate dissolution at the seafloor: fluxes and drivers from a novel in situ porewater sampler

PDF

The early Triassic recovery period: exploring ecology and evolution in marine benthic communities following the Permian-Triassic mass extinction

PDF

Concentration and size partitioning of trace metals in surface waters of the global ocean and storm runoff

$Germanium-silicon fractionation in a continental shelf environment: insights from the northern Gulf of Mexico$

PDF

Germanium-silicon fractionation in a continental shelf environment: insights from the northern Gulf of Mexico

PDF

Critical zone response to perturbation: from mountain building to wildfire

PDF

Electrochemical pathways for sustainable energy storage and energy conversion

PDF

Anaerobic iron cycling in an oxygen deficient zone

PDF

A Maxent-based model for identifying local-scale tree species richness patch boundaries in the Lake Tahoe Basin of California and Nevada

Asset Metadata

Creator Lee, Hyejung (author)

Core Title Isotopic fractionations in plant biomarker molecules with application to paleoclimate

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Geological Sciences

Publication Date 05/10/2019

Defense Date 03/22/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Bengal Fan,carbon cycle,carbon isotopes,fossil wood,GC-IRMS,GC-MS,hydrogen isotopes,lignin,lignin methoxy,OAI-PMH Harvest,organic geochemistry,paleoclimate,plant wax

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Feakins, Sarah J. (committee chair), Corsetti, Frank A. (committee member), Williams, Travis J. (committee member)

Creator Email hyejunghyejung211@gmail.com,hyejungl@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-157395

Unique identifier UC11660340

Identifier etd-LeeHyejung-7437.pdf (filename),usctheses-c89-157395 (legacy record id)

Legacy Identifier etd-LeeHyejung-7437.pdf

Dmrecord 157395

Document Type Dissertation

Format application/pdf (imt)

Rights Lee, Hyejung

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA