xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001498269
007 cr mnu|||uuuuu
008 041209s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000544
Walther, Richard Ayres.
Construction of a late Pleistocene paleothermometer based on amino acid racemization in fossil Succinea shells
h [electronic resource] /
by Richard Ayres Walther.
[Tampa, Fla.] :
University of South Florida,
Thesis (M.S.)--University of South Florida, 2004.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 76 pages.
ABSTRACT: Racemization kinetics of amino acids, determined for the commonly occurring fossil gastropod Succinea, facilitates the ability to construct an accurate and precise paleothermometer to estimate paleotemperatures over specific time intervals during the last 150,000 years in parts of Central Europe. Racemization within the carbonate shell of Succcinea is induced at high temperatures over increasing intervals of time in the laboratory and measured for aspartic acid (asp), glutamic acid (glu), valine (val), and phenylalanine (phe), by reverse-phase liquid chromatography. The activation energy (Ea), frequency factor (A), and forward rate constant (k1) of the Arrhenius equation are determined from the racemization of specific amino acids over time.The Arrhenius parameters, combined with racemization data and independent age estimates of fossil Succinea shells, are used to solve for temperature in geologic samples. Succinea recovered from a loess sequence in western Germany, located around the town of Nussloch, has been chosen for amino acid paleothermometry calculations. Samples were collected from the Nussloch loess -- paleosol sequence in the summer of 2001. The sequence spans from greater than 130,000 years to the present, is dated by luminescence and radiocarbon methods, and has abundant published proxy paleoclimate data for comparison. Temperatures calculated for the bracketed time interval representing the last glacial maximum (25 20ka) averaged -5.3°C± 6.8°C using aspartic acid racemization data. Arrhenius parameters for aspartic acid racemization were the best constrained and provide temperature estimates consistent with previously published data.Paleotemperatures calculated for other bracketed intervals of time within the Succinea shells from Nussloch dated within the last 150,000 years exhibited values similar to previously published data with acceptable error.
Adviser: Oches, Eric A.
amino acid geochronology.
t USF Electronic Theses and Dissertations.
Construction of a Late Pleistocene Pa leothermometer Based on Amino Acid Racemization in Fossil Succinea Shells by Richard Ayres Walther A thesis submitted in partial fulfillment of the requirement s for the degree of Master of Science Department of Geology College of Arts and Sciences University of South Florida Major Professor: Eric A. Oches, Ph.D. Philip van Beynen, Ph.D. Peter Harries, Ph.D. Date of Approval: September 11, 2004 Keywords: Loess, Paleoclimate, Amino Acid Geochronology, Paleothermometer, Nussloch, Germany Copyright 2004, Richard Ayres Walther
Acknowledgements I would first like to thank all the loca l people that helped in the collection of samples and location of field sites duri ng research trips throughout central and eastern Europe. These people include, but are not limited to: Manfred Frechen of the Leibniz Institute for Applied Geo sciences, Ludwig Zoeller from Bayreuth, Germany, Manfred Loscher from Nussloc h, Germany, Slobodan Markovic from the University of Novi Sad, Novi Sad, Serbia & Montenegro, and Erzsebet Horvath from Eotvos University, Budap est, Hungary. We would have been lost without these people and appreciat ed their hospitality. I would also like to thank Kelly Moore, from the University of South Florida and Thomas Stevens, from the Univer sity of Massachusetts, the graduate students that made the resear ch trips with me and helped in sample collection. Dr. William McCoy from the University of Massachusetts is to be thanked for providing a wealth of knowledge and alwa ys getting us to our destinations safely. Dr. Eric Oches deserves my greatest appreciation for teaching me how to become a geologist and providing me the oppor tunity to see the world. I wouldnÂ’t be where I am today wit hout him, thanks Rick.
i Table of Contents List of Tables ii List of Figures iv Abstract v Chapter One Introduction 1 Chapter Two Methods 6 Lab Analysis 10 Chapter Three Amino Acid Racemization Kinetics 13 Arrhenius Parameter Determination 21 Weighted least Â– squares Determination of Arrhenius Parameters 24 Uncertainties in Arrhenius Parameters Determination 34 Chapter Four Amino Acid Paleothermometry Applied to a Loess/Paleosol Sequence in Nussloch, Germany. 37 Introduction 37 Geologic Setting and Stratigraphy 40 Sampling 41 Chronology 44 Paleotemperature Calculation 46 Discussion 58 Sources of error 62 Conclusions 63 References 66
ii List of Tables Table 2.1 Interlaboratory comparison sample standards (ILC-B) during the period of study. 12 Table 3.1 Kinetic study times and temperatures for Succinea samples. 14 Table 3.2 AMS-radiocarbon sample age estimates from the Miami River, Ohio. 15 Table 3.3 Initial D/L values for modern Succinea. 15 Table 3.4 Weighted least squares estimation of Arrhenius parameters for aspartic acid racemization in Succinea 26 Table 3.5 Error propagation in the calculation of Arrhenius parameters for aspartic acid racemization in Succinea 26 Table 3.6 Weighted least squares estimation of Arrhenius parameters for glutamic acid racemization in Succinea 27 Table 3.7 Error propagation in the calculation of Arrhenius parameters for glutamic acid racemization in Succinea 27 Table 3.8 Weighted least squares estimation of Arrhenius parameters for valine racemization in Succinea 28 Table 3.9 Error propagation in the calculation of Arrhenius parameters for valine racemization in Succinea 28 Table 3.10 Weighted least squares estimation of Arrhenius parameters for phenylalanine racemization in Succinea 29 Table 3.11 Error propagation in the calculation of Arrhenius parameters for phenylalanine racemization in Succinea 29 Table 3.12 Summary of weight ed least Â– squares determination of Arrhenius parameters for Succinea 34
iii Table 4.1 Radiocarbon and Thermoluminescence sample age estimates for the Nussloch section. 45 Table 4.2 Effective diagenetic te mperatures (EDTs) calculated from independently dated samples using aspartic acid D/L ratios in Succinea 50 Table 4.3 Effective diagenetic te mperatures (EDTs) calculated from independently dated samples using glutamic acid D/L ratios in Succinea 51 Table 4.4 Effective diagenetic te mperatures (EDTs) calculated from independently dated samples using valine D/L ratios in Succinea 52 Table 4.5 Effective diagenetic te mperatures (EDTs) calculated from independently dated samp les using phenylalanine D/L ratios in Succinea 53 Table 4.6 EDTs calculated for intervals of time bracketed by independently dated samples using aspartic acid D/L ratios in Succinea 54 Table 4.7 EDTs calculated for intervals of time bracketed by independently dated samples us ing glutamic acid D/L ratios in Succinea 55 Table 4.8 EDTs calculated for intervals of time bracketed by independently dated samples usi ng valine D/L ratios in Succinea 56 Table 4.9 EDTs calculated for intervals of time bracketed by independently dated samples using phenylalanine D/L ratios in Succinea 57
iv List of Figures Figure 2.1 Diagenetic relati onships of amino acids. 7 Figure 3.1 Plot of D/ L ratio vs. time for Succinea at 85C. 17 Figure 3.2 Plot of D/ L ratio vs. time for Succinea at 110C. 18 Figure 3.3 Plot of D/ L ratio vs. time for Succinea at 135C. 19 Figure 3.4 Succinea weighted regression Arrhenius plot for aspartic acid. 30 Figure 3.5 Succinea weighted regression Arrhenius plot for glutamic acid. 31 Figure 3.6 Succinea weighted regression Arrhenius plot for valine. 32 Figure 3.7 Succinea weighted regression Arrhenius plot for phenylalanine. 33 Figure 4.1 Map of the Nussloch sect ion and other Pleistocene loess profiles (Modified from Antoine et al., 2001). 40 Figure 4.2 Stratigraphy of sampl ed profiles at Nussloch Quarry (Modified from Moore, 2002). 43 Figure 4.3 EDTs calculated for intervals of time bracketed by independently dated samples. 56
v Construction of a Late Pleistocene Pa leothermometer Based on Amino Acid Racemization in Fossil Succinea Shells Richard Ayres Walther ABSTRACT Racemization kinetics of amino acids, determined for the commonly occurring fossil gastropod Succinea facilitates the ability to construct an accurate and precise paleothermometer to estimate pa leotemperatures over specific time intervals during the last 150,000 years in par ts of Central Europe. Racemization within the carbonate shell of Succcinea is induced at high temperatures over increasing intervals of time in the l aboratory and measured for aspartic acid (asp), glutamic acid (glu), valine (val ), and phenylalanine (phe), by reverse-phase liquid chromatography. The activation energy (Ea), frequency factor (A), and forward rate constant (k1) of the Arrhenius equation are determi ned from the racemization of specific amino acids over time. The Arrhenius parameters, combined with racemization data and independent age es timates of fossil Succinea shells, are used to solve for temperature in geologic samples. Succinea recovered from a loess sequence in western Germany, located around the town of Nussloch, has been c hosen for amino acid paleothermometry calculations. Samples were collect ed from the Nussloch loess Â– paleosol
vi sequence in the summer of 2001. The sequence spans from greater than 130,000 years to the present, is da ted by luminescence and radiocarbon methods, and has abundant published proxy paleoclimate data for comparison. Temperatures calculated for the bracke ted time interval representing the last glacial maximum (25 Â– 20ka) aver aged Â–5.3C 6.8C using aspartic acid racemization data. Arrhenius parameters for aspartic acid racemization were the best constrained and provide temperature estimates consistent with previously published data. Paleotemperatures calculat ed for other bracketed intervals of time within the Succinea shells from Nussloch dated within the last 150,000 years exhibited values similar to previous ly published data with acceptable error.
1 CHAPTER 1 INTRODUCTION Quantitative reconstructions of pas t temperature fluctuations within continental interiors are integral in determining the effect of regional environmental responses to changing atmos pheric conditions during the glacialinterglacial cycles of the middle to late Pleistocene. Improvements and new developments in amino-acid geochr onology and paleothermometry have made it possible to determine accurate and precis e temperature estimates for specific intervals of time during the last glacialinterglacial cycle and beyond. Amino acid racemization data, combined with impr ovements in dating methodology may improve understanding of temperature vari ations over the last 150,000 years, corresponding to marine oxyg enÂ–isotope stages 6 Â– 1. The objective of this research is to develop an accurate amino acid paleothermometer to reconstruct paleotempera tures over specific time intervals over the last 150,000 years in parts of Central Europe, using amino acid racemization kinetics, det ermined in the commonly occurring fossil terrestrial gastropod Succinea To achieve this, modern Succinea are used to determine racemization kinetics of amino acids preserved within th e carbonate shell, which is induced by high temperatures in the laboratory. Amino acid racemization within gastropod
2 shells increases over time at a co nstant rate and is primarily dependent on temperature and time. The temper ature dependent rate of amino acid racemization can be deterimined by exposing separate suites of Succinea to high temperatures over increasing intervals of time. The rate of amino acid racemization over time is used to determine Arrhenius parameters for selected m easured amino acids. The Arrhenius parameters, Ea and A, the activation energy and the frequency factor, respectively, are variables of the Arrheni us equation. The Arrhenius equation defines the temperature depende nt reaction rate of amino acid racemization, determined from the experimentally determi ned Arrhenius paramaters. Arrhenius parameters are determined from heating modern Succinea at known temperatures and subsequently determining the ensuing amino acid racemization rate, which is used to provi de solutions for the Arrhenius equation. Once solved, the Arrhenius equati on can be rearranged to determine paleotemperatures in fossilized Succinea which have experimentally predetermined amino-acid racimization rate s and radiocarbon dates. The Arrhenius equation and application is described in further det ail in Chapter 3. The terrestrial gastropod genus Succinea is chosen for experimentation due to its widespread distribution and abundance in the loess sedimentary record. The shell contains original proteins, which were formed in the biomineralization of the carbonate shell (Miller and Brigham-Gre tte, 1989). The shell forms in successive layers, isolat ing organics from degradation, resulting in
3 long periods of preservation. Amino ac ids comprise the preserved proteins and can be accurately and precisely measured using reverse Â– phase liquid chromatography. It is hypothesized that accurate paleot emperatures can be calculated from the measured extent of racemization of preserved amino acids, combined with independently determined ages of fossilized Succinea and the experimentally determined Arrhenius param eters of racemization. The specific objectives of this research are: 1) To determine Arrhenius parameters of racemization in several amino acids, including aspartic acid (asp), glutamic acid (glu), valine (val), and phenylalanine (phe) of modern Succinea through heating at controlled temperatures for specific time increments. 2) To measure the extent of racemization in Succinea shells collected from Late Pleistocene loess at Nu ssloch, Germany using reverse-phase liquid chromotography. 3) To determine independent ages of samples using radiocarbon and luminescence dating techniques.
4 4) To apply the derived amino acid paleothermometer to Succinea collected from the Nussloch, Germany, loess section in order to calculate paleotemperature estimates for selected intervals of time over the last 150,000 years. Amino acid geochronology and paleothermometry have been demonstrated to effectively estimate paleotemperatures ov er the last 30,000 years in the loess region of the Mississippi Valley, U.S. (Oches, et al., 1996). The use of similar, yet refined, me thods for developing an amino acid paleothermometer to calculate paleotemper ature values will be applied within the loess belt of central Europe. This research is aimed at developing a paleothermometer capable of reconstruc ting detailed paleotemperature for the last interglacial Â– glacial cycle in the loess region of Europe. By using the temperature-controlled amino acid racemization reaction within fossil assemblages of terrestrial gastropod shells preserved in loess, a uniform paleothermometer may be employed to quant ify the temperatur e changes within the study region. It is expected that this research wil l define an amino acid racimization based paleothermometer using the gastropod species Succinea which is helpful for future loess related paleoclima te studies. As a case study, a paleotemperature reconstruction using am ino acid paleothermometry is applied to the site at Nussloch, Germany due to the abundance of Succinea the
5 timespan of the loess profile, and abundance of previously documented stratigraphic and geochronological data (R ousseau et al., 2002; Antoine et al., 2001).
6 CHAPTER 2 METHODS Amino acid geochronology has been show n to provide important insight in the evaluation of chronology and paleot emperature data of fossiliferous sediments of the Quater nary Period (e.g. Oches and McCoy, 2001; Wehmiller and Miller, 2000). Amino acids comprise proteins of all living organisms. Gastropods, which are commonly preserv ed in loess, precipitate calcareous material onto structural protein membranes to form their shell, which protects the amino acid chains within the protei ns from geochemical degradation (McCoy, 1987). Amino acids, which are optically active molecules, are synthesized by organisms in the L-configuration (lev orotary) and begin to degrade through a series of complex diagenetic chemical r eactions (Mitterer, 1993). Once isolated from living tissue, L-amino acids undergo re versible, stereochemical inversion to their D-configuration (dex trorotary) enantiomer or isomer (Oches and McCoy, 2001). This process is called racemizati on when inversion of the L-amino acid takes place around a single chiral carbon to form a D-amino acid, which is its mirror image (Oches and Mc Coy, 2001). Racemization wil l continue until the D/L ratio reaches equilibrium, at a value of approximately 1.0. Some amino acids have more than one central chiral carbon at om. In this case, inversion occurs around the alpha-carbon to form a structurally distinct D-amino diastereoisomer
7 in a process called epimerization, in which equilibrium is reached at a D/L ratio of about 1.3 (Oches and McCoy, 2001)(Figure 2.1). The rate of racemization or epimererization of a particular amino acid in mollusk shells is mainly dependent on time, temperature and taxonomy. Amino acids in different taxa exhibit different ra tes of racemization due in part to the arrangement of the amino acids within peptide chains and also due to the rate of hydrolysis, i.e., whether they are internally bound, at a terminal position, or exist as free amino acids (Kriausakul and Mitterer, 1978). Racemization or epimerization progresses through time, until equilibrium is reached, which is mostly controlled by temper ature. Because large temperature fluctuations have occurred during the Quat ernary Period, it is difficult, if not impossible, to determine accurate numerical age estimates. Figure 2.1. Diagenetic relationships of amino acids. A) The racemization of L-amino acids around a single chiral carbon to form D-amino acids. B) The epimerization of L-amino acids with two chiral carbons produces D-amino acid diastereomers (Miller and Brigham-Grette, 1989)
8 However if independent age determinations are coupled with D/L ratios and kinetic models of amino acid racemization, relatively precise paleotemperatures can be calculated (McCoy, 1987). Several assumptions described by Mitterer (1993) and Wehmiller and Miller (2000) must hold constant in order to apply principles of amino acid geoc hronology and paleothermometry: 1) Fossils of the same genus, wit hin the same geographical region, should have experienced the same thermal history and thus should have the same rates of racemizati on for any particular amino acid; 2) Amino acids behave in a systematic and predictable manner, if the burial history and taxonomy of a fossil are held constant; 3) The system in which the amino acids were preserved has been closed to the inflow and outflow of amino acids since formation. In the past, the most commonly m easured amino acid reaction for paleotemperature studies has been the epi merization of the diastereoisomers Dalloisoleucine and L-isoleucine (A /I), using high-performance liquid chromatography (HPLC) with a post co lumn OPA derivatization method (e.g., Kaufman and Brigham-Grette, 1993; Oches et al., 1996). Though this is the traditional amino-acid pair for analysis, isoleucine epimerizes rather slowly,
9 exhibiting little or no resolvable change in the A/I ratio over a single glacial cycle within samples of the study r egion. It is therefore of li ttle utility in measuring the fine scale changes over the time interv al studied. Kaufman and Manley (1998) have developed a method using reverse-phase liquid chromatography (RPLC) to separate up to ten different amino acid DL isomers in fossil carbonates, which is the method currently in use at the Un iversity of South Florida Geology Department Amino Acid Geochronology Laboratory. The fastest racemizing amino acid reso lvable by RPLC is aspartic acid, at an order of magnitude great er than isoleucine epimerization (Goodfriend, 1992). Aspartic acid racemization will be t he focus of this paleotemperature investigation, providing the resolution required to identify temperature changes within individual glacial cycl es. (Note: due to the sa mple preparation process, asparagine is converted to aspartic acid, and the two compounds are measured together. However, this does not negativel y affect analysis or interpretation of aspartic acid racemization ratios). Other amino acids such as glutamic acid and valine, which racemize more slowly than aspartic acid, but faster than isoleucine, will also be measured to optimize the determination of paleotemperatures over independently dated intervals of time. Phenylalanine, which initially appears to racemize as fast as aspartic acid, will also be used in paleotemperature calculations to determine t he utility of this little studi ed amino acid. Measuring multiple amino acid pairs will provide the opportunity to test t he utility of several
10 amino acids in paleotemperat ure calculation along with a cross check of validity and verification of results. LAB ANALYSIS Samples are prepared in the Universi ty of South Florida Amino Acid Geochronology Laboratory following method s described by Kaufman and Manley (1998). For each sample, a suite of shells of the same genus is placed into a 20ml snap-cap vial and labeled. Each vial was filled with purified water and placed into a Fisher Scientific ultra sonicator to remove debris and foreign particles from each individual shell. Sonification is performed repeatedly and as long as needed for complete cleaning of t he shells. Further mechanical cleaning was performed as needed. Shells are put into the drying hood upon cleaning and removed for weighing when dried. Mini mum weight required for preparation and analysis is about 1mg. Subsamples, each comprised of a single shell, are individually weighed fo r further preparation. Each sub-sample from the sampling suite is dissolved in cold 7N HCl in the proportion of 1 ml of HCl per 50 mg of shell material. Samples are then heated under a nitrogen atmosphere at 110C for 6 hours for total acid hydrolysate analysis (Oches et al., 1996). The hydrolysis step is necessary to break down peptide bonds connecting chains of amino acids in order to measure the total amino acid popul ation (free and peptide bound states) in the sample (Oches et al., 1996). Samples are then dessicated in an evaporating unit in an
11 80C heating module, under nitrogen. Samp les were never desicated for more than two hours in order to minimize lab-induced racemization. Immediately before analysis, samp les are rehydrated with Â“0.01XÂ” rehydration solution containing 1.428 ml of 7 N HCl and 13.5 mg of L-homoArginine per liter of solution at a rati o of .01ml per 1mg of shell material. Analysis is performed at the Univer sity of South Florida Amino Acid Geochronology laboratory using reve rse phase liquid chromatography as described by Kaufman and Manley (1998) on an inte grated Hewlett-Packard HP1100 liquid chromatograph equipped with a quaternary pump and vacuum degasser, an auto-injector and autos ampler, and a HP1100A programmable fluorescence detector. Mixing of O PA (O-phthalaldialdehyde) and ILBC (Nisobutyryl-L-cysteine) with each sample reacts with the amino acids to produce fluorescent diastereomeric by-products, which are then injected onto the Hypersil reverse phase chromatography column. The separation of the Dand Lamino acids employs a C18 stationary phase and mobile phase channels A, B, and C. The stationary phase uses 5 m Hypersil BDS packed in a 250 x 4 mm stainless steel column held at 25 C. eluent A is 3.13g sodium acet ate, 0.1g sodium azide (to inhibit bacterial growth), and 0.5g EDTA, adjusted to pH 6.00 with 10% acetic acid. Fresh eluent A is prepared daily or as needed. Eluent B is optima grade methanol and eluent C is optima grade acetonitrile. Prior to analysis the column is flushed with 95% eluent B and 5% eluent C for a minimum of 20 minutes, and then at 95% eluent A and
12 5% eluent B for another 20 minutes. At this point the run is initiated at 95% A and 5% B. The mobile phase is modified at a uniform gradient to 23% B, 0.4% C at 31 minutes, 44% B, 5. 0% C at 83 minutes, reac hing 95% B, 5% C at 90 minutes. Operation was performed at a flow rate of 0.6 ml/minute, and sample injection volume was 1.5l for the HYD analysis. HP Chemstation software performed in strument control, data acquisition, and chromatographic peak integration. Although our method is capable of seperating the D and L isomers of t en different amino acids, we focused on aspartic acid, glutamic acid, valine, and phenylalanine as representing a wide range of racemization rates. ILC-B (interlaboratory comparison standards; Wehmiller 1984) was measured with each dayÂ’ s set of samples as a check of reproducibility within the equipment (Table 2.1). Table 2.1. Interlaboratory comparison sam ple standards (ILC-B) during the period of study. Aspartic Acid Glutamic Acid Valine Phenylalanine No. of Analysis 27 D/L ratio mean 0.699 0.428 0.477 0.624 Stdev. 0.032 0.016 0.011 0.029 C.V. (%) 4.628 3.807 2.396 4.618 C.V. is the Coefficient of Variation (std. dev./mean)x100
13 CHAPTER 3 AMINO ACID RACEMIZATION KINETICS Amino Acid paleotemperature calculation requires a model for the racemization kinetics of each amino ac id of interest for the gastropod genus studied, plus independent age estimates for sample. Racemization kinetics are acquired through heating experiments of the desired gastropod genera. Heating times and temperatures are modeled after Kaufman (2000), with multiple s ub-modern samples of the genus Succinea heated at three temperatures for increas ing periods of time. For example, 12 Succinea samples were heated at 85C, rangi ng from 0 to 300 days, 14 samples were heated at 110C for intervals of 0 to 60 days, and 14 samples were heated at 135C for intervals of 0 to 200 hour s. Each sample contained three subsamples, which allowed us to assess variability at each step. After heated samples were analyzed, and D/L ratios we re calculated so that Arrhenius parameters of racemization could be determined for each amino acid of interest. Arrhenius parameters are needed for the determination of paleotemperatures since time of burial and for intervals of time bracketed by independently dated samples. Tables 3.1 presents kinetic study times and temperatures for each set of samples.
14 Table 3.1. Kinetic study times and temperatures for Succinea samples. Lab # Temperature (C) time (hours) time (years) 0807-D 85 117.50 1.34E-02 0807-E 85 236.50 2.70E-02 0807-F 85 528.50 6.03E-02 0807-G 85 722.50 8.24E-02 0807-H 85 963.75 1.10E-01 0807-I 85 1202.75 1.37E-01 0807-J 85 1433.75 1.64E-01 0807-K 85 1796.00 2.05E-01 0808-AD 85 2440.50 2.78E-01 0808-AE 85 4970.25 5.67E-01 0808-Q 110 25.50 2.91E-03 0808-R 110 47.98 5.47E-03 0808-S 110 96.25 1.10E-02 0808-T 110 144.12 1.64E-02 0808-U 110 192.75 2.20E-02 0808-V 110 242.75 2.77E-02 0808-W 110 360.78 4.12E-02 0808-X 110 501.25 5.72E-02 0808-Y 110 667.50 7.61E-02 0808-Z 110 721.75 8.23E-02 0808-AA 110 1008.50 1.15E-01 0808-AB 110 1200.08 1.37E-01 0808-AC 110 1441.50 1.64E-01 0808-D 135 1.00 1.14E-04 0808-E 135 2.00 2.28E-04 0808-F 135 5.00 5.70E-04 0808-G 135 10.00 1.14E-03 0808-H 135 20.00 2.28E-03 0808-I 135 30.00 3.42E-03 0808-J 135 40.00 4.56E-03 0808-K 135 50.00 5.70E-03 0808-L 135 70.00 7.99E-03 0808-M 135 100.00 1.14E-02 0808-N 135 130.00 1.48E-02 0808-O 135 150.00 1.71E-02 0808-P 135 200.00 2.28E-02 1029-D 135 2.00 2.28E-04 1029-E 135 6.00 6.84E-04 1029-F 135 10.00 1.14E-03 1029-G 135 20.67 2.36E-03 1029-H 135 30.00 3.42E-03 1029-I 135 40.00 4.56E-03 1029-J 135 50.00 5.70E-03 1029-K 135 70.33 8.02E-03 1029-L 135 100.33 1.14E-02 1029-M 135 119.50 1.36E-02
15 Heating experiments require sub-moder n shells whose ages and postdepositional temperature histories can be reasonably estimated. For Succinea (Fal-0807, 0808) we used shells collect ed from the banks of the Miami River collected near Middletown, Ohio, wh ich were AMS-radiocarbon dated at approximately 1000 years B.P. after calib ration by Calpal, developed by the University of Cologne (Table 3.2). The mean annual temperature used for determination of the racemization reac tion rate constant for the unheated samples is 11.6C, which is the average of the mean average annual temperature at Dayton, Ohio (11 C) and Cincinnati, Ohio (12.2C). Table 3.2. AMS-radiocarbon sample age es timates from the Miami River, Ohio. U. of Arizona Field Number Genus Lab Number OIS 14C age (years) Error (years) Cal. Age Error (years) FAL-0807 Succinea AA47908 1 1,130 56 890 70 FAL-0808 Succinea AA47909 1 1,025 47 1,020 70 We measured initial D/L values (Do/Lo) on modern (live collected) Succinea shells in order to determine the amount of preparation-induced racemization (Table 3.3). Table 3.3. Initial D/L values for modern Succinea. FAL # ASP Do/Lo GLU Do/Lo VAL Do/Lo PHE Do/Lo 1037-AH1 0.064 0.021 0.023 0.015 1037-BH1 0.045 0.016 0.009 0.012 1037-CH1 0.054 0.018 0.010 0.012 MEAN 0.054 0.018 0.014 0.013 STDEV 0.010 0.003 0.008 0.002
16 Succinea shells were prepared for heating by placing up to three subsamples into separate 10 ml test tubes on a ~4 cm3 bed of sand. Approximately 2 ml of deionized water was added to each test tube to saturate the sand to simulate natural conditions. Temperatures of 85 C, 100 C, and 135 C were used to develop temperature kinetic curves for amino ac id racemization ratios to be determined through RPLC analysis for Succinea modeled after Manley et al. (2000) (Figures 3.1, 3.2, 3.3).
17 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000 0100020003000400050006000 time(hours)D/L ratio Aspartic Acid Glutamic Acid Valine Phenylalanine 0 1 2 3 4 0200040006000ln[(1+D/L)/(1-D/L)] Figure 3.1. Plot of D/L ratio vs. time for Succinea at 85C. Inset graph shows linear transformation of the data for the purpose of defining the upper limit of the initial linear approximation of the racemization reaction.
18 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000 0250500750100012501500 time(hours)D/L ratio Aspartic Acid Glutamic Acid Valine Phenylalanine 0 1 2 3 4 010002000ln[(1+D/L)/(1-D/L)] Figure 3.2. Plot of D/L ratio vs time for Succinea at 110C. Inset graph shows linear transformation of the data for the purpose of defining the upper limit of the initial linear approximation of the racemization reaction.
19 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000 0255075100125150 time (hours)D/L ratio Aspartic Acid Glutamic Acid Valine Phenylalanine 0 1 2 3 4 050100150ln[(1+D/L)/(1-D/L)] Figure 3.3. Plot of D/L value vs. time for Succinea at 135C. Inset graph shows linear transformation of the data for the purpose of defining the upper limit of the initial linear approximation of the racemization reaction.
20 The extent of amino acid ra cemization (D/L ratio) in Succinea increases with heating time in the four amino acids aspa rtic acid (Asp), glutamic acid (Gla), valine (Val), and phenylalanine (Phe) resolved by reverse phase liquid chromatography, as shown in figures 3.13.3. This basic premise is needed to model the racemization of an amino acid with respect to temperature and time. The plots of D/L values vs. time model the racemi zation progression of each amino acid at high temperatures. This simulates the increase of the amino acids over geologic time at low temperatures. (Manley et al., 2000). Only the initial linear approximation of the modeled race mization curves is valid for the paleotemperature calculations (Manley et al., 2000). A plot of ln[(1+D/L)/(1-D/L)] vs. ti me is shown for each genus at each heating time and is displayed in the left corn er of figures 3.1 Â– 3.3. Points that deviate from this linear transformation define the upper limit of each amino acid D/L value that can be used in A rrhenius parameter determination and paleotemperature calculation. For aspar tic acid, linear behavio r is exhibited for D/L ratios < about 0.4 at te mperatures 85 C, 110 C, and 135 C for all genera. Aspartic acid D/L values above 0.4 exhibi t much slower race mization rates and no longer exhibit first order linear kineti cs. This is expected, because other studies have shown that aspartic ac id racemization is complicated and unpredictable due to the interplay of am ino acid and polypeptide decomposition, hydrolysis, and racemization (Kriausakul and Mitterer, 1978). Linear behavior is demonstrated through transformation plot s for glutamic acid, valine, and
21 phenylalanine at all heating times for Succinea measured D/L ratios. Anomalous D/L values were removed from ln[(1+D/L) /(1-D/L)] plots for all measured amino acids if D/L values significantly deviated from the mean of other subsamples at a particular heating time or all subsamples at a given heating time displayed significant scatter (> 10%). Arrhen ius parameters are determined based on D/L values exhibiting first order linear kinet ics for each individual amino acid and for genus. Two different sets of kinetic experi ments were undertaken at 135 C using Succinea due to unexplained anomalies in mu ltiple amino acid racemization ratios for the firs t set: FAL-0808 (D-P). The second 135 C heating experiment set, using FAL-1029 (D-M), did not contain any anomalous data, though D/L ratios higher than 0.35 0.4 exhibi ted high error about the mean. Using ln[(1+D/L)/(1-D/L)] transfo rmation, it is demonstrated that aspartic acid follows first order linear kineti cs until a D/L ratio of 0.4 and are non-linear past that point. ARRHENIUS PARAMETER DETERMINATION Arrhenius parameters and subsequent paleotemperature estimation is determined following the methods described by McCoy (1987). The Lto Damino acid transformation is considered to be a first-order reversible reaction within the total acid hydrolysate and c an be approximated as a linear progression
22 for the initial phase of racemizati on (Schroeder and Bada, 1976; Williams and Smith, 1977): D Lk k 1 2 Where L and D represent the Land Dstereoisomers and k1 and k2 are the forward and reverse rate constants, res pectively, of the ra cemization reaction (Smith, et al., 1977). The rate expression fo r this reaction is: D k L k dt L d2 1 (1) which can be integrated to give equation (2). Two basic equations are used in the determination of Arrhenius parameters and paleotemper atures from the racemizati on kinetics of D/L amino acid ratios through time. These in clude the integrated instantaneous rate equation (2) and the Arrhenius equation (3 ) (Schroeder and Bada, 1976; Williams and Smith, (1977). t k K L D K L D L D K L D1 0 0 0 01 1 1 ln 1 1 ln (2)
23 D/L equals the ratio of the concentrations of Dand Lenantiomers of an amino acid. K' equals k2/k1 or the reciprocal of the equilibrium constant of the racemization, which equals 1.0 for amino acids measured in this study. D0/L0 is equal to D/L at t=0, which is the D/L ra tio of a modern sample and may not equal zero due to laboratory induced racemi zation during sample preparation, and t is the independently determined ag e of the sample. Temperature dependence of the rate constant (k1) is expressed by the Arrheni us equation (3) (McCoy, 1987; Oches, 1996). RT EaAe k1 (3) A is the frequency factor or entropy, Ea represents the energy of activation (cal mol-1 Kelvin -1), R is the gas constant (1.9872 cal Kelvin-1 mol-1) and T is temperature in Kelvin. Ea and A are the Arrhenius parameters of the racemization reaction. A linear re lationship can be established among the variables by taking the natural logarithm of both sides of the Arrhenius equation: ln k1 = ln A Â– Ea/RT (4) Values for k1 can be determined by substituting the D/L ratios of the analyzed heating experiment samples with the other known variables into equation (1), rearranged to solve for k1. Values of ln k1 were determined for each analyzed
24 amino acid pair for Succinea Ln k1 versus 1/T is then plotted for each analyzed amino acid for each genus, and the slope (= -Ea/R) and intercept (= ln A) of the regression line through the points dete rmines the Arrhenius parameters, Ea and A. WEIGHTED LEAST Â– SQUARES DETERMINATION OF ARRHENIUS PARAMETERS McCoy (1987) described a modified approach for solving the Arrhenius parameters, in which a weighted least s quares regression is fit to the data. Weightings are proportional to the inverse of the variance of ln k1. The variance of ln k1 is determined by the slope (M) of t he leastÂ–squares regression line plus the square root of the standard deviation of ln k1 multiplied by t he square root of the standard deviation of 1/T. This method also allows for a multivariate error analysis, where the uncertainties in measured paramet ers can be propagated through the Arrhenius equation in order to estimate uncertainty of Ea and A (Clifford, 1973). Generally, Arrhenius parameters are determined by a least-squares fit to the plot of ln k1 vs. 1/T (McCoy, 1987). Instead, weightings for the least Â– squares determinations were calculated by taking the inverse of the variance of ln k1 for each sample at each temperatur e, i.e. 11.6 C ( no heating, mean annual temp.), 85 C, 110 C, and 135 C. Once the weightings have been determined, the slope (M) and y-intercept (B) of the l east squares line may be calculated for
25 each measured amino acid in Succinea using the following equations (Clifford, 1973, p.41): i i i i i i i iY w X w w Y X w lsd M 1 (5) i i i i i i i i iY X w X w Y w X w lsd B21 (6) where X = 1/T, Y = ln k1 2 2 i i i i iX w w X w lsd i iY w21 and i = 1,n. The Arrhenius parameters ar e then calculated from the slope (M) and intercept (B) for each measured amino acid of each genus; following the method of McCoy (1987)(Tables 3.4 Â– 3.11)(Figures 3.4 3.7): Ea = -1.9872M, and A = eB
26 r 0.9995 Ea 28929.6 Std. Dev. Ea445.69 A 2.22E+18 Std. Dev. A 1.27E+18 slope -14557.99 epd 3.22E-39 intercept 42.24 c12 -5.57E-25 -2.88E-21 -1.09E-18 -7.29E-19 -1.82E-18 lsd 1.86E-02 c2 2.19E-09 8.99E-06 3.17E-03 2.00E-03 5.18E-03 wi 30.42 5.07 732.59 3.67 c1 1.42E-40 9.23E-37 3.73E-34 2.66E-34 6.40E-34 Tot. Var. ln k10.03 0.2 1.37E-03 0.27 dk1/dEa-2.39E-07 -6.82E-03 -9.06E-02 -8.74E-01 SUM= Std. Dev. 1/T 5.82E-11 0 0 6.96E-11 dk1/dA 6.09E-23 2.19E-18 3.11E-17 3.19E-16 Std. Dev. ln k11.81E-01 4.44E-01 3.69E-02 5.22E-01 wi 3.28E+04 0.19 0.39 0 Average ln k1-8.94 2.23 4.23 6.32 Tot. Std. k12.62E-05 5.17 2.59 382.02 Average k11.34E-04 9.96 68.86 636.74 Std. Dev. k12.40E-05 3.75 2.57 350.24 Average 1/T 0.00351 0.00279 0.00261 0.00245 Average k11.34E-04 9.96 68.86 636.74 T (Kelvin) 284.6 358 383 408Table 3.5. Error propagation in the calculation of A rrhenius parameters for aspartic acid racemization in Succinea Table 3.4. Weighted least squares estimation of A rrhenius parameters for aspartic acid racemization in Succinea
27 r 0.9998 Ea 28480.4 Std. Dev. Ea3389.65 A 1.21E+17 Std. Dev. A 5.08E+17 slope -14331.93 epd 6.82E-40 intercept 39.33 c12 -5.32E-24 -6.68E-21 -1.35E-19 -1.03E-18 -1.17E-18 lsd 4.77E-03 c2 1.13E-09 1.13E-06 2.14E-05 1.54E-04 1.77E-04 wi 519.06 4.18 4.31 3.1 c1 2.49E-38 3.94E-35 8.54E-34 6.94E-33 7.83E-33 Tot. Var. ln k11.93E-03 0.24 2.30E-01 0.32 dk1/dEa-2.87E-08 -6.97E-04 -8.88E-03 -8.26E-02 SUM= Std. Dev. 1/T 5.21E-11 5.49E-11 2.77E-07 3.42E-06 dk1/dA 1.35E-22 4.11E-18 5.61E-17 5.55E-16 Std. Dev. ln k14.00E-02 4.90E-01 4.80E-01 5.70E-01 wi 1.37E+06 2.34 0.27 0.02 Average ln k1-11.03 -0.39 1.78 4.06 Tot. Std. k17.29E-07 0.43 3.68 44.39 Average k11.63E-05 0.77 6.72 68.2 Std. Dev. k17.07E-07 0.45 3.75 41.08 Average 1/T 0.00351 0.00279 0.00261 0.00245 Average k11.63E-05 0.77 6.72 73.77 T (Kelvin) 284.6 358 383 408Table 3.7. Error propagation in the calculation of A rrhenius parameters for glutamic acid racemization in Succinea Table 3.6. Weighted least squares estimation of A rrhenius parameters for glutamic acid racemization in Succinea
28 r 0.9998 Ea 27077.9 Std. Dev. Ea4300.26 A 8.45E+15 Std. Dev. A 4.50E+16 slope -13626.15 epd 8.28E-38 intercept 36.67 c12 -3.32E-24 -4.26E-20 -1.21E-18 -1.58E-17 -1.71E-17 lsd 4.49E-05 c2 4.97E-11 5.06E-07 1.34E-05 1.65E-04 1.79E-04 wi 1.05 3.8 11.78 16.2 c1 2.23E-37 3.58E-33 1.09E-31 1.52E-30 1.63E-30 Tot. Var. ln k10.95 0.26 8.00E-02 0.06 dk1/dEa-2.40E-08 -3.50E-04 -3.93E-03 -3.26E-02 SUM= Std. Dev. 1/T 5.21E-11 5.82E-11 2.74E-07 2.33E-06 dk1/dA 1.61E-21 2.95E-17 3.54E-16 3.13E-15 Std. Dev. ln k19.80E-01 5.10E-01 2.90E-01 2.50E-01 wi 8.59E+04 4.12 0.87 0.15 Average ln k1-11.87 -1.01 1.23 3.13 Tot. Std. k11.16E-05 0.24 1.15 6.45 Average k11.11E-05 0.42 3.57 22.84 Std. Dev. k11.29E-05 0.31 1.24 6.78 Average 1/T 0.00351 0.00279 0.00261 0.00245 Average k11.11E-05 0.42 3.57 22.39 T (Kelvin) 284.6 358 383 408Table 3.9. Error propagation in the calculation of Arrhenius parameters for valine racemization in Succinea Table 3.8. Weighted least squares estimation of Arrhenius parameters for valine racemization in Succinea
29 r 0.9998 Ea 28786.6 Std. Dev. Ea1792.04 A 3.70E+17 Std. Dev. A 8.24E+17 slope -14485.99 epd 9.33E-40 intercept 40.45 c12 -5.68E-25 -5.35E-21 -1.70E-19 -1.20E-18 -1.38E-18 lsd 4.40E-04 c2 3.71E-10 2.78E-06 8.28E-05 5.47E-04 6.33E-04 wi 19.4 6.77 14.76 5.86 c1 8.69E-40 1.03E-35 3.51E-34 2.63E-33 2.99E-33 Tot. Var. ln k10.05 0.15 7.00E-02 0.17 dk1/dEa-5.13E-08 -1.39E-03 -1.82E-02 -1.73E-01 SUM= Std. Dev. 1/T 7.62E-11 5.82E-11 1.29E-06 3.42E-06 dk1/dA 7.08E-23 2.67E-18 3.75E-17 3.81E-16 Std. Dev. ln k12.30E-01 3.80E-01 2.60E-01 4.10E-01 wi 1.41E+05 1.44 0.25 0.02 Average ln k1-10.49 0.39 2.60 4.68 Tot. Std. k17.08E-06 0.69 4.00 54.97 Average k12.85E-05 1.61 13.96 117.55 Std. Dev. k16.42E-06 0.75 4.07 54.84 Average 1/T 0.00351 0.00279 0.00261 0.00245 Average k12.85E-05 1.61 13.96 125.16 T (Kelvin) 284.6 358 383 408Table 3.11. Error propagation in the calculation of A rrhenius parameters for phenylalanine racemization in Succinea Table 3.10. Weighted least squares estimation of A rrhenius parameters for phenylalanine racemization in Succin e
30 ln k1 = 42.2 -14558 / T Ea = 28930 446 cal/mol A = 2.22 x 1018 1.27 x 1018r = 0.9995-10 -8 -6 -4 -2 0 2 4 6 8 0.00240.00260.00280.0030.00320.00340.0036 1/T (kelvin)ln k1 Figure 3.4. Succinea weighted regression Arrhenius plot for aspartic acid.
31 ln k1 = 39.3 -14332 / T Ea = 28480 3340 cal/mol A = 1.21 x 1017 5.08 x 1017r = 0.9998-12 -10 -8 -6 -4 -2 0 2 4 6 0.00240.00260.00280.00300.00320.00340.0036 1/T (kelvin)ln k1 Figure 3.5. Succinea weighted regression Arrhenius plot for glutamic acid.
32 ln k1 = 36.7 -13626 / T Ea = 27077.88 4300 cal/mol A = 8.45 x 1015 4.50 x 1016r = 0.9998-14 -12 -10 -8 -6 -4 -2 0 2 4 6 0.00240.00260.00280.00300.00320.00340.0036 1/T (kelvin)ln k1 Figure 3.6. Succinea weighted regression Arrhenius plot for valine.
33 ln k1 = 40.5 -14486 / T Ea = 28787 1792 cal/mol A = 3.70 x 1017 8.24 x 1017r = 0.9998-12 -10 -8 -6 -4 -2 0 2 4 6 8 0.00240.00260.00280.00300.00320.00340.0036 1/T (kelvin)ln k1 Figure 3.7. Succinea Weighted regression Arrhenius plot for phenylalanine.
34 A summary of the determined Arrhenius parameters is listed in table 3.12. Table 3.12. Summary of weighted least Â– squares determination of Arrhenius parameters for Succinea. Amino Acid Ea Stdev. Ea A Stdev. A r Aspartic Acid 28929.63 445.69 2.22E+18 1.27E+18 0.9995 Glutamic Acid 28480.42 3389.65 5.08E+17 5.08E+17 0.9998 Valine 27077.88 4300.26 8.45E+15 4.50E+16 0.9998 Phenylalanine 28786.55 1792.04 3.70E+17 8.24E+17 0.9998 UNCERTAINTIES IN ARRHENIUS PARAMETER DETERMINATION Uncertainties in Arrhenius paramet er estimations, determined through the weighted leastÂ–squares me thod, can be assessed through multivariate error analysis outlined by Clifford (1973). A nother set of weightings is calculated based on the variance of k1, which is determined by the exponent of the average ln k1 minus the exponent of the average ln k1 plus the square root of the total variance of ln k1. The errors are propagated from the uncertainty in the forward rate constants, k1, to the Arrhenius parameters Ea and A modeled after the methods of McCoy (1987). See Clifford (1973) for a detailed explanation and derivation: 2 1 2 12 2 1 1c c c c Ea (7) 2 1 2 12 2 1 2c c c c A (8)
35 where ,2 1 1dA dk w ci i ,2 1 2a i idE dk w c ,1 1 12a i i idE dk dA dk w c i ik w1 21. When 9872 1 exp1T E dA dka and 9872 1 exp 9872 11T E T A dE dka a Temperature uncertainties are minima l in laboratory-heated samples with a maximum uncertainty of 1 C. Effe ctive diagenetic temperatures for submodern geologic samples used in Arr henius parameter determination is estimated to be the current mean annual air temperature at the sample locality. Therefore, temperature uncer tainty in those samples is much greater. McCoy (1987) provides a thorough ex planation of the error analysis and concludes that the combined errors of the Arrhenius par ameters are much la rger than the error associated with temperature estimation for the samples used in Arrhenius parameter determination. Uncertainty in the D/L amino acid race mization ratios is expressed as the standard deviation of the mean, with the mean incl uding about three subsamples at each temperature interval. Th is error is usually < 5%; samples with larger errors are not used for quantificati on of Arrhenius parameters. D/L values
36 of collected gastropod samples from specific loess stratigraphic layers consistently display even less error. Do/Lo values for the live collected Succinea which is used in the constant within the integrated instant aneous rate equation, have measured error of 1%.
37 CHAPTER 4 AMINO ACID PALEOTHERMOMETRY APPLIED TO A LOESS/PALEOSOL SEQUENCE IN NUSSLOCH, GERMANY INTRODUCTION On continents, quantitative temperatur e data are important in determining climatic responses to environmental c hanges. To determine variabilities in climate change, a sedimentary sequence able to register climatic oscillations is needed. Due to its high accumulation ra tes and nearly continuous deposition, loess sequences are ideal for recordi ng changes in climate throughout the Quaternary (Kukla, 1977). Previous paleoclimatic studies in central Europe have been performed based on modern analogue interpretations of fossil assemblages collected from lake sediment cores and loess profiles. These data are limited by our knowledge of the present environment al ranges and preservation of the fossil floral and faunal assemblages. Studies include 1) pol len (e.g., Woillard 1978; Guiot et al., 1989, 1992), 2) insect remains (e.g., Guiot et al., 1993; Ponel, 1995), and 3) mollusks (e.g., Rousseau, 1991; Moine et al., 2002). While these methods provide quantitative estimate s of paleotemperatures, they all have limitations. The biggest problem is the lack of a modern analogue with which to fossil assemblages of full glacial environment s (Guiot et al., 1989, 1992). Another
38 problem is that different taxonom y of fossil assemblages can produce significantly different paleotemperature es timates (e.g. Guiot et al., 1993). Also, pollen data are plagued by a limited avail ability of suitable sampling sites. Relatively few lake sediments cores spanni ng the entire last glacial-interglacial cycle are available from central Europe (Guiot et al., 1993). Current paleotemperature reconstruc tions based on the forementioned methods, suggest that Europe was about 12 Â– 22 C and 8 Â– 12 C cooler in February and August, respectively, duri ng the last glacial maximum (LGM) compared to present, in the broad region co rresponding to the loess belt (Frenzel et al., 1992). COHMAP paleoclimate models predict much cooler temperatures across Europe for the LGM of 16 Â– 32 C and 4 Â– 16 C colder than present in January and July, respectively (Kutz bach et al., 1993). It is clear that paleoclimate models need further refinem ent to adequately model the past climate and proxy data needs. Paleocli mate studies utilizing amino acid racemization kinetics for paleotemperature calculations may provide the needed data that other methods cannot offer. Present climate in the loess belt of the study region exhibits a generally moist continental climate. A pplying this unique biogeochemical paleothermometer across this region will help determine the degree to which the Fennoscandian Ice Sheet and North Atlantic sea ice expansion during the last glacial maximum contributed to more st rongly continental climates. General climate trends during the middle Â– late Pleistocene for the region can also be
39 resolved and compared to current paleoclimat ic data from lake sediment cores, faunal, floral and oxygen isotope data to check validity and enhance climate modeling. Finally, a la rge database of amino acid geochronological data, independent chronological dating, ef fective temperature data and kinetic parameters of amino acid racemization will be created that will supplement and refine methods of amino acid geochr onology and paleothermo metry for future investigations. A loess sequence in western Germ any, located near the town of Nussloch, along the Nekar River, has been chosen for our amino acid paleothermometry calculations. Sample s were collected from the Nussloch loess-paleosol sequence, for this study, in the summe r of 2001. The sequence represents deposition from greater than 120,000 years B. P. to the persent, as dated by luminescence and radiocarbon methods, and has been a site of considerable research, which allows for co mparisons to our findings. Within the Nussloch loess paleosol sequence, infer ences of middle to late Pleistocene paleoclimate data have been derived through 13O of organic matter (Hatt et al., 1999, 2001), magnetic suscept ibility (Rousseau et al., 2002) and sedimentology (Antoine et al., 2001), although actual paleotemperature values have not been determined. 13O measurements of the loess profile at Nussloch by Hatt et al., (1999) exhibit general trends of increasing and decreasing values similar to increasing and decreasing values of 18O in the GISP 2 ice core when compared chronologically over the past ~70 ka. Nu merical estimates of paleotemperature
40 are not calculated, but general cool ing and warming trends have been suggested from the 13O data at Nussloch (Hatt et al., 1999). GEOLOGIC SETTING AND STRATIGRAPHY The Nussloch loess-paleosol section is located on the east margin of the Rhine graben, along the Neka r River valley at 49 21 N, 8 43 E within an active Limestone quarry in the town of Nussloch, Germany (Figure 4.1). This loess paleosol sequence is extremely well developed, with distinct intervals of glacial, interstadial, and interglacial deposits. The sequence is ~26 m thick, has been correlated to marine oxygen isot ope stage (OIS) 6 (possibly OIS10) to OIS 1 and is underlain by Triassic car bonates (Zller and Lscher, 1999) (Figure 4.2). The base of the sequenc e, representing OIS 6 or older, contains loess and Figure 4.1. Map of the Nussloch section a nd other Pleistocene loess profiles (Modified from Antoine et al., 2001 ) .
41 reworked loessic sands and silts. OIS 5e is correlated to a parabranerde soil, OIS 5c is represented by a gray forest soil, and the Mosbach Humus Zone (interpreted as a chernozem) represents OIS 5a. The top of this chernozem marks the boundary between OIS 5 and OIS 4 (Zller and Lscher, 1999). Loess occurs at the base of OIS 4 with the Â“Nussloch soilÂ” formed above it. Above the Nussloch soil is the Niedereschbacher Zone or reworked loessic sands, depending on the location within the exposure. The base of OIS 3 is marked by the Grselberger Boden, a weak brown soil with tundra-gley features. This is fo llowed by a thick section of loess. The entire sequence of loess is intermingl ed with tundra-gley soils and the Lohner Boden, which is interpreted as an arctic meadow soil (Zller and Lscher, 1999). OIS 2, the last glacial maximum, is represented by over 8 m of loess, intermingled with five weak tundra-gley paleosols. OIS 2 also contains the Eltville Tephra, which is dated at other localities to about 17 kya (Frechen, 1999). SAMPLING Sampling was performed on three differ ent profiles within the Nussloch loess exposure and at a nearby nature pres erve (Figure 4.2). 1-2 kg bags of sediment were collected from many leve ls within of the profile, targeting each loess unit and a few weak paleosols. Bags of collected sediment were sieved and rinsed with water to isolate desired gastropod shells from their sediment matrix. The separated gastropods were placed into vials labeled with the
42 corresponding field number based on st ratigraphic assignment. Vials were transported back to the Ami no Acid geochronology Laboratory at the University of South Florida for further cleaning, pr eparation and analysis. Only collected samples containing desired gastropod s hells used for paleothermometry are displayed on Figure 4.2.
43 Figure 4.2. Stratigraphy of sampled pr ofiles at Nussloch Quarry (Modified from Moore, 2002). Profiles A and B are from the limestone quarry. Profile C is from an exposure at a nature preserve approximately 2 miles away. OIS correlations are from Hatte et al., (2001). TL age estimates are from Zoller and Semmel (2001). Radiocarbon age estimates are from this study and were measured at University of Arizona NSF-AMS Radiocarbon Laboratory.
44 CHRONOLOGY Gastropod shells collect ed from the Nussloch stratigraphic section were independently dated by AMS radiocarbon dating of she lls and correlation with previously published luminescenc e age estimates (Table 4.1). Luminescence dating was performed on loess sediment surrounding the samples, which is an ideal material for applying this method, as these sediments tend to be age zeroed during transport and burial and have the potential for yielding a high resolution chronostratigr aphy over the last 150,000 years (Zller and Wagner, 1990). The Nussloch loess profile has been previously dated through luminescence methods, providing already detailed chronologies (Zller and Semmel 2001; Zller et al., 1988). Dating of gastropod shells was done through accelerator mass spectrometry (AMS) radiocarbon dating at the NSF-University of Arizona AMS facility. AMS radiocarbon dating of car bonate fossil shells is reliable to about 35,000 years ago, allowing for paleothermome try through the LGM. Reliability of the chronology is imperative, so radioc arbon age estimates are obtained from the same suite of shells used for amino acid racemization analysis (Table 4.1). Our radiocarbon ages correlate with radiocarbon age estimates and luminescence ages from the Nussloch section present ed by Lang et al. (2003). TL ages correlated well with 14C ages until about 30 ka. For older samples 14C ages are consistently younger than the accepted age based on TL dating and stratigraphy (Zller et al., 1988; Lang et al., 2003). This probably reflects the reliable upper-
45 age limit of 14C measured in shell material. T he same results were observed in this study, with all radiocar bon ages in profile B (Figur e 4.2) considered younger than actual and also in Sample # 010617-3 from profile A (Figure 4.2). Arrhenius parameters of amino acid racemization are based on calendar years, requiring accurate calibration of radiocarbon ages. Calibration was achieved by CalPal software from the University of Cologne for all samples younger than 35,000 years. Calibrated ages are recorded on Table 4.1 and are incorporated into the paleotemperature ca lculations in Tables 4.2 through 4.9. Table 4.1. Radiocarbon and Thermo luminescence sample age estimates from the Nussloch section. Field Number Genus Oxygen Isotope Stage (OIS) 14C Age Error Calibrated Age Error 100617-3 Trichia 4 36,480 870 38,310 950 010617-6 Succinea 3 33,620 580 36,490 1,820 010617-7 Succinea 3 32,670 550 35,170 1,110 010617-9a Pupilla 3 26,540 270 28,020 620 010617-9b Succinea 3 24,610 220 25,990 770 010617-9c Trichia 3 24,230 210 25,840 890 010617-11a Pupilla 2 24,780 220 26,080 750 010617-11b Succinea 2 23,950 240 25,690 970 010617-11c Trichia 2 24,280 300 25,860 880 010617-12a Pupilla 2 24,240 330 25,830 890 010617-12b Succinea 2 23,300 210 24,480 180 010617-12c Trichia 2 22,760 200 24,100 240 010618-14 Succinea 2 19,060 100 20,520 320 010619-1 Succinea 4 38,600 1,100 39,870 500 010619-2 Pupilla 4 32,060 590 34,610 1,490 010619-3 Succinea 4 36,280 790 38,240 930
46 PALEOTEMPERATURE CALCULATION Paleotemperature estimates der ived from the i ndependently dated Succinea samples collected at Nussloch, Ge rmany are based on the previously determined Arrhenius parameters of racemi zation for aspartic acid, glutamic acid, valine, and phenylalanine. Equations (2) and (3) can be combined and rearranged to provide equation 9, which is used to estimate the effective diagenetic temperature of the entire postdepositional hi story of a sample: At L D L D L D L D R E To o o o a2 1 1 ln 1 1 ln ln (9) T (in Kelvin), is the ef fective diagenetic temperat ure (EDT), which is the temperature experienced by a fossil, int egrated over time, since deposition. The temperature history of a sample may be complex, with temperature fluctuations ranging from days to tens of thousands of years. This may affect the rate of racemization, which is increasingly a ccelerated with increased temperature, related to the activation energy (Ea). The measured activation energies of the analyzed amino acids ( Asp, Glu, Val, and Phe ) are relatively high, with Ea values ranging from 27 Â– 28 kcal mol-1 (in Succinea ), resulting in increasing
47 racemization rates with increasing temper ature. Hence, the time spent by a sample at high temperatures is dispropor tionately more important in increasing D/L ratios than the time spent at lo wer temperatures. This may produce an effective diagenetic temperat ure greater than the m ean temperature that the sample actually experienced (Wehmiller, 1977). High rates of loess deposition and rapid burial are solutions to this problem, which appear to be prevalent during glacial periods (Pye, 1987). The gastropod sample experiences little time at increased temperatures and most of its burial history at moderate, nonvariable temperatures attributed to rapid and thick loess deposition. The temperature for an interval of time bracketed by independent age estimates can be calculated (McCoy, 1987): k A R E Ta t tln1 2 (10) where 1 2 1 1 2 2/ t t t k t k k 1 2 2 1' 1 ln t t K a a (11) and 1 1 1 1 1' 1 1 L D K L D a (12) and 2 2 2 2 2' 1 1 L D K L D a (13)
48 In these equations, t2 and t1 refer to the ages of the older and younger samples, respectively. D2/L2 and D1/L1 represent the older and younger amino acid racemization ratios, respectively. Uncertainties in the Arrhenius par ameters, the equili brium constant, measured D/L ratios, and i ndependent age estimates ar e propagated through the temperature equations to determine t he magnitude of uncertainty of the paleotemperature estimates, following the method established by McCoy (1987). A precision of 1C can be achieved by minimizing the uncertainties during paleotemperature and temperatures for brac keted intervals of time calculations (McCoy, 1987). The entire postdepositiona l temperature according to each amino acid measured was determined us ing equation 9 (Ch. 2) (Tables 4.2 Â– 4.5). Temperatures for an interval of time bracketed by independent age estimates for each amino acid measured were determi ned using equation 10 (Ch. 2) (Tables 4.6 Â– 4.9) and the difference in effect ive diagenetic temperature (EDT) between the younger and older intervals was calcul ated from the previous two estimates (Tables 4.6 Â– 4.9). The latter differs bec ause D/L ratios represent the integrated temperature since sample burial, and that older intervals experienced the same subsequent EDT changes as the younger intervals (Kaufman, 2003). The change in temperature ( T) is determined from the formula T = EDT1 Â– T(t2 Â– t1), where EDT1 is the effective diagenetic temper ature of the younger age to present and T(t2 Â– t1) is the temperature interval from a later age to the younger age
49 representing the end member of EDT1. The T represents the change in temperature from one time interval to another and determines if temperatures are increasing or decreasing from one time interval to another.
50 Table 4.2. Effective diagenetic temperatures (EDTs) calculated from independently dated samples using aspartic acid D/L ratios in Succinea Error Do/Lo = 0.054 0.010 Ea = 28929.63 445.69 A = 2.22E+18 1.27E+8 r = 0.9995 Older Asp EDT Field # D/L error Age (ka) error (C) error profile A 010617-1 0.769 0.033 150,000 5,000 010617-3 0.391 0.017 60,000 5,000 -4.9 1.4 010617-6 0.321 0.025 36,490 1,820 -3.7 1.4 010617-7 0.370 0.018 35,170 1,110 -2.6 1.3 010617-9 0.338 0.009 25,990 770 -1.6 1.3 010617-11 0.356 0.017 25,690 970 -1.3 1.3 010617-12 0.347 0.013 24,480 180 -1.2 1.3 010618-14 0.324 0.009 20,520 320 -0.7 1.3 profile B 010619-2 0.447 0.012 70,000 5,000 010619-1 0.404 0.013 39,870 500 010619-3 0.371 0.018 38,240 930 -3.0 1.3 profile C 010620-1 0.538 0.016 140,000 5,000
51 Table 4.3. Effective diagenetic temperatures (EDTs) calculated from independently dated samples using glutamic acid D/L ratios in Succinea. error Do/Lo = 0.018 0.003 Ea = 28480.42 3389.646 A = 1.21E+17 5.08E+17 r = 0.9998 Older Glu EDT Field # D/L error t (ka) error (C) error profile A 010617-1 0.432 0.034 150,000 5,000 2.1 10.5 010617-3 0.133 0.026 60,000 5,000 -0.1 10.7 010617-6 0.090 0.018 36,490 1,820 0.0 10.7 010617-7 0.084 0.007 35,170 1,110 -0.2 10.6 010617-9 0.081 0.012 25,990 770 1.1 10.6 010617-11 0.073 0.004 25,690 970 0.4 10.6 010617-12 0.080 0.004 24,480 180 1.4 10.5 010618-14 0.076 0.006 20,520 320 1.9 10.5 profile B 010619-2 0.161 0.011 70,000 5,000 0.3 10.6 010619-1 0.119 0.005 39,870 500 1.4 10.5 010619-3 0.113 0.007 38,240 930 1.2 10.5 profile C 010620-1 0.216 0.015 140,000 5,000 -1.6 10.7
52 Table 4.4. Effective diagenetic Temperatures (EDTS) ca lculated from independently dated samples using valine D/L ratios in Succinea error Do/Lo = 0.014 0.008 Ea = 27077.88 4300.262 A = 8.45E+15 4.50E+16 r = 0.9998 Older Val EDT Field # D/L error t (ka) error (C) error profile A 010617-1 0.402 0.051 150,000 5,000 2.2 14.1 010617-3 0.088 0.017 60,000 5,000 -2.1 14.4 010617-6 0.044 0.005 36,490 1,820 -4.4 14.5 010617-7 0.053 0.006 35,170 1,110 -2.7 14.4 010617-9 0.045 0.002 25,990 770 -2.3 14.4 010617-11 0.048 0.006 25,690 970 -1.7 14.4 010617-12 0.041 0.003 24,480 180 -2.7 14.5 010618-14 0.053 0.009 20,520 320 0.3 14.3 profile B 010619-2 0.161 0.011 70,000 5,000 0.8 14.2 010619-1 0.119 0.005 39,870 500 2.1 14.1 010619-3 0.113 0.007 38,240 930 1.9 14.1 profile C 010620-1 0.216 0.015 140,000 5,000 -1.2 14.3
53 Table 4.5. Effective diagenetic temperatures (EDTs) calculated from indepently dated samples using phenylalanine D/L ratios in Succinea. error Do/Lo = 0.013 0.002 Ea = 28786.55 1792.036 A = 3.70E+17 8.24E+17 r = 0.9998 Older Phe EDT Field # D/L error t (ka) error (C) error profile A 010617-1 0.835 0.053 150,000 5,000 4.4 5.4 010617-3 0.088 0.017 60,000 5,000 -5.1 5.8 010617-6 0.186 0.015 36,490 1,820 1.8 5.5 010617-7 0.176 0.020 35,170 1,110 1.6 5.5 010617-9 0.156 0.010 25,990 770 2.5 5.5 010617-11 0.160 0.009 25,690 970 2.7 5.5 010617-12 0.151 0.009 24,480 180 2.6 5.5 010618-14 0.146 0.008 20,520 320 3.4 5.5 profile B 010619-2 0.282 0.012 70,000 5,000 0.7 5.5 010619-1 0.244 0.017 39,870 500 2.8 5.5 010619-3 0.213 0.013 38,240 930 2.3 5.5 profile C 010620-1 0.438 0.019 140,000 5,000 -0.3 5.5
54 Error 3.9 12.3 4.7 8.4 5.9 (C) 3.7 2 7.6 9.6 11 TError 1.3 1.3 1.3 1.3 1.3 (C) -0.7 -1.2 -0.7 -1.3 -2.6 EDT1 Error 3.7 12.3 4.5 8.3 5.8 T(t2-t1) (C) -4.5 -3.2 -8.3 -10.8 -13.6 Error 0.009 0.013 0.009 0.017 0.018 D1/L1 0.324 0.347 0.324 0.356 0.37 Error 320 180 320 970 1110 Age1 20520 24480 20520 25690 35170 Error 0.013 0.017 0.009 0.018 0.017 Error 445.69 1.27E+18 D2/L2 0.347 0.356 0.338 0.37 0.391 28929.63 2.22E+18 0.9995 Error 180 970 770 1110 5000 Ea = A = r = Age2 24480 25690 25990 35170 60000Table 4.6. EDTs calculated for intervals of time bracketed by independently dated samples using aspartic acid D/L ratios in Succinea T = EDT1-T(t2-t1), where EDT1 is from the younger age to presen t and T(t2-t1) is the interval temperature between the two ages.
55 Error 17.1 90.6 19.5 15.5 15.3 15 16.5 15.1 (C) 4.6 8.8 5.5 3.1 -0.3 -3.4 -2.4 3 TError 10.5 10.54 10.5 10.6 10.6 10.7 10.5 10.5 (C) 1.88 1.35 1.88 0.45 -0.2 -0.1 1.24 1.36 EDT1 Error 13.5 90 16.4 11.3 11 10.5 12.6 10.8 T(t2-t1) (C) -2.7 -7.4 -3.6 -2.7 0.1 3.3 3.6 -1.7 Error 0.006 0.004 0.006 0.004 0.007 0.026 0.007 0.005 D1/L1 0.076 0.08 0.076 0.073 0.084 0.133 0.113 0.119 Error 320 180 320 970 1110 5000 930 500 Age1 20520 24480 20520 25690 35170 60000 38240 39870 Error 0.004 0.012 0.012 0.007 0.026 0.034 0.005 0.011 Error 3389.65 5.08E+17 D2/L2 0.08 0.081 0.081 0.084 0.133 0.432 0.119 0.161 28480.42 1.206E+17 0.9998 Error 180 770 770 1110 5000 20000 500 5000 Ea = A = r = Age2 24480 25990 25990 35170 60000 150000 39870 70000 T = EDT1-T(t2-t1), where EDT1 is from the younger age to presen t and T(t2-t1) is the interval temperature between the two ages. Table 4.6. EDTs calculated for intervals of time bracketed by independently dated samples using glutamic acid D/L ratios in Succinea
56 Error 20.8 21 23 20.5 20.1 20.3 21 (C) -9.1 5.4 5.6 -1.5 -6.1 -6.1 8.4 TError 14.5 14.3 14.4 14.4 14.4 14.1 14.1 (C) -2.7 0.3 -1.7 -2.7 -2.1 1.94 2.05 EDT1 Error 15 15.4 18 14.6 14.1 14.6 15.6 T(t2-t1) (C) 6.5 -5.2 -7.3 -1.3 4 8 -6.3 Error 0.003 0.003 0.006 0.006 0.017 0.008 0.016 D1/L1 0.041 0.041 0.048 0.053 0.088 0.077 0.072 Error 180 320 970 1110 5000 930 500 Age1 24480 20520 25690 35170 60000 38240 39870 Error 0.006 0.002 0.006 0.017 0.051 0.006 0.006 Error 4300.26 4.5E+16 D2/L2 0.048 0.045 0.053 0.088 0.402 0.089 0.089 27077.88 8.45E+15 0.9998 Error 970 770 1110 5000 20000 500 5000 Ea = A = r = Age2 25690 25990 35170 60000 150000 39870 70000 T = EDT1-T(t2-t1), where EDT1 is from the younger age to presen t and T(t2-t1) is the interval temperature between the two ages. Table 4.6. EDTs calculated for intervals of time bracketed by independently dated samples using valine D/L ratios in Succinea
57 Error 15.7 10.4 10.1 10.6 13.6 8 8.5 8.3 (C) 8.7 -2 6.4 6.4 11.7 -11.3 -7.1 7.6 TError 5.5 5.5 5.5 5.5 5.5 5.8 5.5 5.5 (C) 3.4 2.6 3.4 2.7 1.6 -5.1 2.3 2.8 EDT1 Error 14.7 8.9 8.5 9.1 12.4 5.5 6.5 6.3 T(t2-t1) (C) -5.4 4.6 -3.1 -3.7 -10.1 6.3 9.4 -4.8 Error 0.008 0.009 0.008 0.009 0.02 0.015 0.013 0.017 D1/L1 0.146 0.151 0.146 0.16 0.176 0.186 0.213 0.244 Error 320 180 320 970 1110 5000 930 500 Age1 20520 24480 20520 25690 35170 60000 38240 39870 Error 0.009 0.009 0.01 0.02 0.015 0.053 0.017 0.012 Error 4300.26 4.5E+16 D2/L2 0.151 0.16 0.156 0.176 0.186 0.835 0.244 0.282 27077.88 8.45E+15 0.9998 Error 180 970 770 1110 5000 20000 500 5000 Ea = A = r = Age2 24480 25690 25990 35170 60000 150000 39870 70000 T = EDT1-T(t2-t1), where EDT1 is from the younger age to presen t and T(t2-t1) is the interval temperature between the two ages. Table 4.6. EDTs calculated for intervals of time bracketed by independently dated samples using phenylalanine D/L ratios in Succinea
58 DISCUSSION The effective diagenetic te mperature (EDT) of each Succinea sample from the Nussloch section is calculated from equation (9) and independent age estimates (Table 4.1) based on individual A rrhenius parameters for aspartic acid, glutamic acid, valine, and phenylalanine. The sample depth and stratigraphic placement within the Nussloch section is represented in Figure 4.2, and correlation with marine oxygen-isotope stages 1-6 are shown (Hatt et al., 2001). Results of EDTs calculated for intervals of time bracketed by independently dated samples of Succinea collected from the Nussloch loess section, based on aspartic acid, glutamic acid, valine, and phenylalanine are summarized on figure 4.3. Aspartic acid EDT values were comput ed for D/L ratios < 0.4. Aspartic acid values above this ratio no longer di splay linearity (Figure 3.1-3.3), making them unusable. This limits paleotemperat ure estimates, based on aspartic acid D/L ratios, to 60,000 5000 years B.P. fo r this study. This is acceptable, considering that the three remaining ami no acids exhibit a linear D/L relationship over the full range of time represent ed at Nussloch, based on heating kinetics (figures 3.1-3.3).
59 -15 -10 -5 0 5 10 15 020,00040,00060,00080,000100,000120,000140,000160,000 Time (kya)Temperature (C) Aspartic Acid Glutamic Acid Valine Phenylalanine Figure 4.3 Results of EDTs calculated for inte rvals of time bracketed by independently dated samples of Succinea collected from the Nussloch loess section, based on aspartic acid, glutamic acid, valine, and phenylalanine
60 All EDTs based on the measured amino acid D/L ratios displa7y progressively warmer temperatures from 60 kya to 20kya, which is not consistent with marine OIS curves (The date for field number 010617-1 is estimated from stratigraphic position and is not consi dered accurate to determine temperature data). Marine OIS curves exhibit a c ooling trend beginning around 60kya, which increases to 35kya. This may be a result of meas uring the entire postdepositional temperature hi story of sample, where the time spent at relatively high temperatures is of far greater importance than time spent at lower temperatures (Oches et al., 1996). The rate of racemization is increasingly accelerated with increasing temperature, which results in higher effective temperature exposure of samples than the mean annual temperature experienced by the samples (Wehmiller, 1977). This explains why younger samples exhibit higher temperatur es, because they have spent a longer percentage of their history at warmer te mperatures, increasing their calculated EDTs. EDT values derived from aspartic acid D/L ratios were consistently lower than EDT values from glutamic acid and phenylalanine, though not valine. EDTs with respect to aspartic acid D/L rati os of three independently dated samples between 24,480 kya and 25,990 kya are wit hin 0.4C of one another (Table 4.2), displaying good agreement. T he EDT error of the three samples is 1.3C, which is less than 1% error.
61 The EDT of the same three samples between 24,480 kya and 25,990 kya computed with glutamic acid D/L values are within 1C of each other with about 4% error (Table 4.2). This is also an acceptable level of precision. The 24,480 kya and 25,990 kya dated sample s are also within 1C of each other for valine, with an error of 5% (T able 4.3). The increased error associated with glutamic acid and valin e are expected, when considering that these two amino acids racemize much slower t han aspartic acid and phenylalanine, making them less precise in recording changes in temperature over time. This is exhibited in valine with a total D/L ratio change of 0. 007 for the three samples, whereas aspartic acid had a D/L ratio diffe rence of 0.018 for the three samples. Phenylalanine had a 0.2C difference for the 24,480 kya and 25,990 kya dated samples, with an error of 2% (Tabl e 4.4). This low error is expected according to the previous statement. The three mean age estimates of 25,990 kya, 25,690 kya, and 24,480 kya are basically equivalent when accounting for radiocarbon dating error and should have similar EDT values for all four meas ured amino acids. From Tables 4.1 4.4, the highest calculated EDT for these dates is 2.7C from phenylalanine D/L ratios and the lowest is -2.7C from va line D/L values; roughly a 5.4C difference, which is within the error estimates. The effective diagenetic temperatures calculated for intervals of time bracketed by independently dated Succinea samples is determined from
62 equation (10)(Tables 4.5 4.8), based on t he Arrhenius parameters of aspartic acid, glutamic acid, valine, and phenylalanine. Comparing Succinea samples from 25, 990 20,520 kya for the four measured amino acids pairs, an average EDT for the bracketed interval of -5C is calculated. Aspartic acid D/L ratios produce a value of -8.3C with an error of 2%, glutamic acid D/L ratios produce a va lue of -3.6 with an error of 6%, valine D/L ratios produce a value of -5.2 wit h an error of 6%, and phenylalanine D/L ratios produce a value of -3.1 with an erro r of 3%. Errors increase due to an increase of errors being propagated through the equation with two EDTs, compared to one EDT wit h equation (9). An average temperature of -5C 4% fo r the bracketed interval for the last glacial maximum is consistent with estimates by Frenzel et al., (1992), Kutzbach et al., (1993), and Moine et al., (2002). Th is is 15C colder than present annual temperature averages of ~10.5C for the study area (GHCN 2 beta). SOURCES OF ERROR There are many sources of error to take into account when determining paleotemperatures. Errors revolve around independent age estimates, D/L ratios, and especially Arrhenius parameters (Tables 6a-d). Most of the error in paleotemperature calculations is due to Ar rhenius parameter error. This error can be reduced by refining amino acid D/L values measured in the kinetic
63 experiments used for k1 determination. Amino acid D/L error measured from kinetic experiments can be further constrained by: 1) Using more sub-samples per heati ng time to reduce the amino acid ratio standard deviation 2) Incorporate more heating time sample intervals per heating temperature 3) Refine the reverse phase liquid chromatography process for increased precision and accuracy. Errors involved in independent ag e determination can be reduced by independently dating numerous sub-sample s from the same sample to reduce the standard deviation. CONCLUSIONS Arrhenius parameters were determined from racemization kinetics of the amino acids: aspartic acid, glutamic acid, valine, and phenylalanine preserved within sub-modern Succinea A large suite of Succinea shells, heated at a range of temperatures over varying time intervals, with multip le sub-samples, provided acceptable racemization kinetics. D/L values incr eased over time at c onstant temperature with less than 5% error, providing a good fit regression line. Succinea
64 racemization data, translated to ln k1 vs. 1/T values, also provided a good fit regression line for determining Arrhenius parameters used in paleotemperature calculation. Arrhenius parameters determined from Succinea racemization kinetics yield paleotemperature values with reasonable error among measured amino acids. Amino acid paleothermometry based on Arrhenius parameters combined with independent age estimates applied to the Succinea shells from the Nussloch loess section provides paleotemperature values c onsistent, within the error, of paleotemperature values of other methods (Frenzel et al., 1992, Kutzbach et al., 1993, and Moine et al., 2002). Applying the paleotemperature equations (9 and 10) to four amino acid D/ L ratios of the same genus, Succinea, provides paleotemperature estimates consistent within the error within the same bracketed, independent ly dated, interval time and also consistency within each individual measured amino acid when co mparing separate samples of about the same time period. Bracketed intervals of time from a pproximately 26,000 kya Â– 20,520 kya, representing the last glacial maximum, display average te mperatures of -5.3 C 6.8 C, using aspartic acid arrheni us parameters. Average bracketed temperatures for the same time period using arrhenius parameters yield values of -4.3 C 39.3 C for glutamic acid, -2 .0 C 16.1 C for valine, and -1.3 C 10.7 C for phenylalanine. The paleotemperat ure error of the la tter three amino acids is high, but provides an average te mperature consistent within the last
65 glacial maximum. Aspartic acid paleot emperature estimates provide the least amount of error with values also consis tent with expected results for the last glacial maximum.
66 REFERENCES An, Z., Liu, T., Porter, S.C., Kukla, G. Wu, X., and Hua, Y. (1990). The long term paleomonsoon variation recorded by t he loess paleosol sequence in central china. Quaternary International 7/8 91-95. An, Z., Porter, S.C., Kukla, G., Xaio, J. L. (1991). Late Quat ernary dust flow on the Chinese Loess Plateau. Catena 18 125-132. Boenigk, W., and Frechen, M. (2001). The loess record in sections at KoblenzMetternich and Tnchesburg in the Middle Rhine Area. Quaternary International 76/77 201-209. Bradley, R.S. (1985). Paleoclimatol ogy: Reconstructing Climates of the Quaternary. Harcourt Academic Press, 613p. Clifford, A.A. (1973). Mult ivariate error analysis. New York, John Wiley and Sons. Ding, Z., Rutter, N.W., and Lui, T.S. (1993) Pedostratigraphy of Chinese loess deposits and climatic cycles in the last 2.5 Ma. Catena 20 73-91. Frechen, M. (1992). Systematic thermolu minescence dating of two loess profiles from the Middle Rhine area (F .R.G.). Quaternary Science Reviews, v. 11, p. 93101. Frechen, M., Zander, A., Cilek, V., and Lo ek, V. (1999). Loess chronology of the last interglacial/glacial cycle in Bohemia and Moravia, Czech Republic. Quaternary Science Reviews 18 1467-1493. Frechen, M., van Vliet-Lano, B., and van den Haute, P. ( 2001). The Upper Pleistocene loess record at Harmignies /BelgiumÂ—High resolu tion terrestrial archive of climate forcing. Palaeogeography, Palaeoclima tology, Palaeoecology 173 175-195. Goodfriend, G.A. (1992). Rapi d racemization of aspartic acid in mollusk shells and potential for dating over recent cent uries. Nature, v. 357, p. 399-401.
67 Kaufman, D. and Manley, W. (1998). A new procedure fro determining D/L amino acid ratios in fossils usin g reverse phase liquid chromatography. Quaternary Science Reviews 17 987-1000. Kriausakul, N. and Mitterer, R.M. (1978). Isoleucine epimerization in peptides and proteins: kinetic factors and a pplication to fossil proteins. Science 201 1011-1014. Kukla, G. (1975). Loess stratigraphy of Central Europe. In: Butzer, K.W. and Isaac, G.L. (eds.), After the Australopi thecines. Monton, the Hagine, 99-188. Kukla, G. (1977). Pleistocene Land-S ea Correlations. I. Europe. Earth-Science Reviews 13 307-374. Kukla, G. (1987). Pleistocene Climates in China and Europe compared to oxygen isotope record. Palaeoecology of Africa 18 37-45. Kukla, G., An, Z., Melice, J.L., Gavin, J., and Xiao, J.L. (1990). Magnetic Susceptibility record of Chinese Loess. Trans. Royal Society of Edinburgh: Earth Sciences 81 263-288. Kutzbach, J.E., Guetter, P. J., Behling, P.J., and Selin, R. (1993). Simulated climatic changes: results of the COHM AP climate-model ex periments. In: Wright, H.E., Kutzbach, J.E., Webb, T., R uddiman, W.F., Street-Perrott, F.A., and Bartlein, P.J., (Eds.). Global Climat e Since the Last Glacial Maximum. University of Minnesota Pr ess, Minneapolis, p. 24-93. Lowe and Walker (1984). Reconstructing Quaternary Environm ents. Prentice Hall, 446p. Maher, B.A., Thompson, R. (1995). Pal eorainfall reconstructions from pedogenic magnetic susceptibility variati ons in the Chinese Loess. Quaternary Research 44 383-391. McCoy, W.D. (1987). The Precisi on of Amino Acid Geochronology and Paleothermometry. Quaternary Science Reviews 6 43-54. Miller, G.H., and Brigham-Grette, J. (1989). Amino Acid Geochronology: Resolution and Precision in Carbonate Fossils. Quaternary International 1 111128. Mitterer, J.E. (1993). The di agenesis of proteins and amino acids in fossil shells. In: Engel, M.H. and Macko, S.A. (eds.) Organic Geochemistry, Plenum Press, 739-753.
68 Oches, E.A., McCoy, W., and Clark, P. (1996). Amino Acid Estimates of latitudinal temperature gradients and geochronology of loess deposition during the last glaciation, Mississippi Valley, United States. GSA Bulletin 108 892-903. Oches, E.A., and McCoy, W.D. (2001). Historical developments and recent advances in amino acid geochronology appli ed to loess research: examples from North America, Europe and Chi na. Earth-Science Reviews. Pye, K. (1987). Aeolian Dust And Dust Deposits. Academic Press, 334p. Reinders, J., and Hambach, U. (1995). A geomagnetic event recorded in loess deposits of the Tnchesberg (Germany): identification of the Blake magnetic polarity episode. Geophysics Journal International 122 407-418. Rousseau, D.D., Antoine, P., Hatt, C. Lang, A., Zller, L., Fontugne, M., Ben Othman, D., Luck, J.M., Moine, O., Labonne, M., Bentaleb, I., and Jolly, D. (2002). Abrupt millennial climatic changes from Nussloch (Germany) Upper Weichselian eolian records dur ing the Last Glaciation. Quaternary Science Reviews 21 1577-1582. Schroeder, R.A., and Bada. J.L. (1976) A review of the geochemical applications of the amino ac id racemization reaction. Earth Science Reviews 12 347-391. Wehmiller, J.F., and Miller, G.H. (2000). Aminostratigraphic Dating Methods in Quaternary Geology. In: Quaternary Geology: Methods and Applications. AGU reference shelf 4, 187-222. Smalley, I., and Smalley, V. (1983). Loess material and loess deposits: formation, distribution and consequences. In: M. Brookfield and T. Ahlbrandt, Eolian Sediment and Processes, 51-68. Elsevier, Amsterdam. Williams, K.M., and Smith, G. G. (1977). A critical eval uation of the application of amino acid racemization to geochronology and geothermometry. Origins of Life 8 91-144. Zller, L., and Wagner, G. (1990). Thermo luminescence dating of loess Â– recent developments. Quaternary Inter national, v. 7/8, p. 119-128.