PP51C-1502
Using Maximum Autocorrelation Factors (MAFS) to Identify Common Trends in Proxy Data: A Tree-Ring Case Study
Proxy records of climate and environmental variability are most reliable when they are based on more than one sample per site. Therefore, prior to any type of process-modeling of the proxies themselves, it is often necessary to combine multiple samples into a single chronology. In doing so, it may be useful to develop numerical techniques that clearly identify trends that are in common among the individual samples. To this end, data are commonly analyzed using methods closely related to principal components analysis, although such methods are not specifically optimized for time trend detection. Less common alternative analyses use variants of MAF, maximal autocorrelation factor analysis. We describe MAF optimality properties that are specific to time trend extraction for proxy data, and illustrate MAFs trend extraction possibilities in low signal- to-noise situations by applying them to ring-width measurements taken on a total of 92 western juniper (Juniperus occidentalis Hook.) samples collected at two sites within the Walker River basin, on the eastern slopes of the Sierra Nevada near the California-Nevada boundary. Although the tree-ring dataset spans 2300 years, from 300 BC to AD 2001, the analysis had to be conducted on periods of overlap between individual samples. The resulting temporal trends are presented, and their interpretation in terms of biological and climatic processes is discussed.
PP51C-1503
SYNCHRON: A Forward Model of Tree-Ring Chronology Structure
The signal recorded in tree-ring based reconstructions of climate has been filtered not only through the interaction of the tree with its environment, but also through the process of sampling, measurement and analysis applied to the tree rings. Simulation models may be used to examine the operation of both "filters" and to identify potential sources of error and bias in such reconstructions. For the first "filter", models such as the Vaganov-Shashkin model simulate the climatic control of the processes of cell division, expansion and maturation leading to the production of tree rings. Such models can capture the major features of the climate response of a single "ideal tree". Here we demonstrate the usefulness of a new model, named SYNCHRON, designed to simulate the operation of the second "filter". It permits the exploration of the sensitivity of dendroclimatic reconstructions to the various methods used to extract common climate signal from collections of tree-ring data, the degree of replication and the specific distribution of samples of various ages through the reconstructed period. Starting with a perfectly known climate signal, the consequences of sample structure and data processing techniques representative of those reported in the current literature may be tested by creating very large numbers of synthetic tree-ring chronologies and reconstructions with realistic properties. The resulting reconstructions can then be compared with the input climate signal throughout the reconstruction period, unlike the real world situation. In each case our starting point is the sample structure and data processing techniques used in a particular study. SYNCHRON is used to explore the range of reconstructions that could be produced in such a case by repeated re-sampling of the population of synthetic tree-ring series to vary sample depth and age-structure distribution through time. We will present examples based on recent data sets that have contributed to climate reconstructions.
PP51C-1504
Reconstructing Climate Using non de-trended Stable Carbon Isotope Ratios From Pine Trees in Northern Finland
In a 2007 paper (Gagen et al., 2007) we presented a stable carbon isotope tree ring chronology covering the period AD 1640-2002. The series was constructed using Pinus sylvestris trees sampled from close to the northern limit for the species, in Finnish Lapland. This chronology was used to develop a reconstruction of summer temperatures, based on the correlation between the carbon isotope series and July/August mean temperatures (r = 0.72). In Gagen et al., (2007) we tested the hypothesis that stable carbon isotope series from tree rings might not contain any long-term age-related trends and thus require no statistical de-trending and we used a non- detrended stable carbon isotope series to reconstruct summer temperatures since AD 1640. However, we concluded that the sample replication in the proxy series was probably too low to capture lower frequency variability. Here we present an extension to that study via an AD 1150-2006 chronology, developed near to the original site, using sub-fossil trees extracted from lakes. We describe a novel sampling method which was used to rapidly process more than ninety individual trees for stable carbon isotope dendroclimatology with the aim of retaining both high and lower frequency climatic variability.
PP51C-1505
Applying Multivariate Calibration Techniques to a Multiple-Proxy Paleoclimate Problem
Interpreting paleoclimate data is often ambiguous as the proxies involved usually depend on more than one environmental variable, i.e. not just temperature. One method used regularly to overcome this problem is paired analysis of two proxies (mostly, δ18O and Mg/Ca), however, to produce a more robust statistical analysis and reduce error in the model, calibration studies with several proxies and several explanatory variables are needed (thus generating a multivariate calibration problem). Four methods of multivariate calibration will be discussed: classical prediction, a linear least-squares approach assuming error in the dependent (proxy) variables; Bayesian calibration, in which Baye's rule is applied to classical prediction to give an 'improved' estimator; inverse calibration, with x (environmental variables) regressed on y (proxies) (opposite to classical prediction, with error assumed in the independent variables) and, finally, functional regression, in which error is assumed on both the independent (environmental) and dependent (proxy) variables. Analysis of a synthetic dataset is used to highlight the differences between each method. Multivariate calibration is then applied to a real problem from the paleoceanographic literature to illustrate the merits of each multivariate technique. Coral data from New Caledonia is analysed using each technique to evaluate the use of Sr/Ca, δ18O and δ13C records to estimate past sea- surface temperature and salinity. A comparison is also made between reconstructions using the multivariate method and standard techniques to assess the applicability of multivariate calibration to further paleoclimate applications.
PP51C-1506
Pollen-Climate Calibration, Characterization of Statistical Uncertainty, and Forward Modeling for Integration Into Bayesian Hierarchical Climate Reconstruction
A strict process model for pollen as a climate proxy is currently not approachable beyond localized spatial scales; more generally, the canonical model for vegetation-pollen registration itself requires assimilation of empirically-derived information. In this paper, a taxonomically "reduced-space" climate-pollen forward model is developed, based on the performance of a parallel inverse model. The goal is inclusion of the forward model in a Bayesian climate reconstruction framework, following a 4-step process. (1) Ratios of pollen types calibrated to temperature are examined to determine if they can equal or surpass the skill of multi-taxonomic calibrations using the modern analog technique (MAT) optimized with receiver operating characteristic (ROC) analysis. The first phase of this examination, using modern pollen data from SW N America, demonstrates that the ratio method can give calibrations as skillful as the MAT when vegetation representation (and associated climate gradients) are characterized by two dominant pollen taxa, in this case pine and oak. Paleotemperature reconstructions using the ratio method also compare well to MAT reconstructions, showing very minor differences. [Ratio values are defined as pine/(pine + oak), so they vary between 0 and 1.] (2) Uncertainty analysis is carried out in independent steps, which are combined to give overall probabilistic confidence ranges. Monte Carlo (MC) analysis utilizing Poisson distributions to model the inherent variability of pollen representation in relation to climate (assuming defined temperature normals at the modern calibration sites) allows independent statistical estimation of this component of uncertainty, for both the modern calibration and fossil pollen data sets. In turn, MC analysis utilizing normal distributions allows independent estimation of the addition to overall uncertainty from climate variation itself. (3) Because the quality tests in (1) indicate the ratio method has the capacity to carry as much information as the multi-taxonomic MAT in well-selected regions, the ratio inverse model is used as the starting point for developing a temperature-pollen forward model. Such a forward model can be directly incorporated into a Bayesian hierarchical model (BHM) framework: the reason for seeking a taxonomically simplified pollen model is itself driven by computational considerations within the BHM. The BHM framework thus allows pollen- derived climate information to contribute multi-decadal to centennial information into a simultaneous reconstruction framework that also incorporates annually-resolved climate information from tree ring sources and centennial to multi-centennial information from borehole reconstructions. In the forward model, the pine and oak pollen counts are treated as binomial random variables with expected values (p) and (1-p), conditional on the calibration temperature values and the total counts (n) used to determine the ratios. Model parameters and (p) values have been estimated using a binomial logistic form of the generalized linear model. Comparison of the estimated conditional expected (p) values with the associated calibration ratio/temperature relationships shows good performance for the SW N American test case, with little (but non- 0) bias and good overall tracking (R2 = 0.79). (4) The results from (3) indicate that appropriate forward models can be developed for a suite of sites in North America with temporally dense, well-dated fossil pollen records over the past 2ka, and then incorporated into the BHM. Modeling results for the suite of sites will be presented.
PP51C-1507
Pseudo-Proxy Strategies for the PAGES/CLIVAR Paleoclimate Reconstruction (PR) Challenge
Using climate model output as a means for testing large scale reconstruction methods requires subsampling of various model climate fields. Starting from perfectly known climate data offers the distinct advantage that the individual role of a particular uncertainty out of a cascade of factors can systematically be tested. Therefore, the PR-Challenge will offer a suite of climate model-derived pseudo-proxy networks with increasing uncertainty to allow assessment of reconstruction methods from best-case all the way to worst- case situations (e.g., including non-stationarities, dating problems, as well as limited calibration potential). "Perfect proxy" networks will simply form the foundation on which one can evaluate the potential signal detection skills across the participating methods. As the model samples are perturbed with first simple and then increasing structural noise, the comparisons will become more more realistic. One goal of the PR- Challenge is to motivate the proxy communities to think more rigorously about and to come up with physically or biogeochemically-based process models that realistically reproduce the governing mechanisms of variations in their respective archives. The more advanced these methods become, and as they can be included for pseudo-proxy generation, the more realistic will be the test that the PR-Challenge platform offers.
PP51C-1508
Statistical Relationships Among Proxies of Climate, Productivity and the Carbon Cycle Across Climatic Regimes, Santa Barbara Basin, California
The ultra-high-resolution record from Santa Barbara Basin (SBB) provides an opportunity to study the relationships among climate, productivity and the carbon cycle across climatic regimes. We conducted a statistical analysis of multi-proxy data derived chiefly from splits of identical samples from Core MD02-2503, recovered from the central SBB near ODP Site 893. The record spans 33.5 to 4.5 ka (Cal. Yr bp) and includes Glacial, Pre Bolling Warming, Deglaciation, and Interglacial climatic intervals. Data sets of % CaCO3, % total organic carbon (TOC), C/N ratios, δ18O G. bulloides, % silt, and porosity (gamma ray attenuation) were examined by principal components analysis (PCA) to group the best- correlated variables into two components in order to better understand the sensitivity of the proxies to external forcing. Component one includes δ18O G. bulloides (negative loading) and TOC (positive loading), whereas component two includes CaCO3 (negative loading) and C/N (positive loading). PCA not only indicates which variables are related in their behavior, but also the relative sensitivity to forcing under different climatic conditions, thus suggesting the existence of threshold changes in response to forcing functions. Examined over time, PCA suggests during the glacial interval δ18O and C/N ratios are more sensitively responsive to forcing, whereas during the Holocene interglacial, variation in sensitivity of TOC and CaCO3 is emphasized. PCA also highlights distinctive behavior during shorter events within the glacial and Holocene, where one or both components deviated from the general trends. For example, during the Younger Dryas, neither component departs from zero and, during Dansgaard-Oeschger events, both components deviate in the same direction. Further investigation is needed, but linear regressions over the entire record show the highest overall negative correlation between δ18O and TOC, with warmer surface water correlating with higher TOC. The next best correlation is a negative one between C/N and CaCO3, reflecting the influence of marine productivity vs. terrestrial organic carbon contribution. This study lays the groundwork for statistical analysis and characterization of proxy behavior during earlier climatic cycles that will be studied from new cores from the SBB, hopefully ranging back to 1.2 Ma.
PP51C-1509
Evolution of the Pleistocene Climate Continuum
Orbital-climate theory provides a powerful deterministic framework for the analysis of Pleistocene climate change, and has become a cornerstone of modern Paleoclimatology. The stochastic component of Pleistocene climate has received considerably less attention, although some studies have argued for the dominance of stochastic climate processes. Undoubtedly, a complete understanding of the controls on Pleistocene climate change necessitates an assessment of both deterministic and stochastic processes, as well as potential linkages between the two. In this study, we investigate changes in the dominance of deterministic versus stochastic climate processes associated with evolution of the Pleistocene climate system. Achievement of this objective requires: (1) careful selection and analysis of paleoclimate data series, to isolate true climate noise from other proxy noise sources (e.g., proxy fidelity, time-scale distortion, diagenesis, analytical error), and (2) application of quantitative methods capable of separating deterministic periodic signals (the spectral "lines") from the stochastic component of climate (the spectral "continuum"). This study focuses on an analysis of published benthic foraminifera oxygen isotopic records using a number of techniques rooted in Thomson's multi-taper method, which is specifically designed to separate deterministic "line" energy from stochastic "continuum" energy. Our analysis indicates large, and sometimes abrupt, changes in the relative dominance of stochastic versus deterministic energy. These changes in energy distribution parallel the evolution of the Pleistocene climate system. This analysis also demonstrates linkages between stochastic and deterministic climate processes, and yields insight into the mechanisms of Pleistocene climate change.
PP51C-1510
Potential Biases in Inferring Holocene Temperature Trends from Long-Term Borehole Information
We use a set of global coupled ocean-atmosphere general circulation model (OAGCM) experiments for timeslices over the Holocene from 9K Yr BP onwards to examine differences in Northern Hemisphere Ground Surface Temperature (GST) and Surface Air Temperature (SAT) trends. The model results are driven mainly by the orbitally-forced changes in seasonal insolation, and in particular, the increase in northern hemisphere seasonality in the mid Holocene. We see a significant increase in early to mid Holocene average GST compared to SAT as a result of a competition between the effects of changing seasonal insolation, and the varying extent of insulating seasonal snow cover. The model shows a mid-Holocene peak in annual mean terrestrial Northern Hemisphere GST, but not in annual Surface Air Temperatures (SAT). The model reproduces qualitatively the known trends in NH temperatures, but may underestimate their magnitude. However, the difference between GST and SAT trends appears to be a robust result from the greatly varying seasonal changes. We conclude that the factors influencing long-term GST trends are potentially quite complex, and that considerable care must be taken in interpreting SAT changes from the GST evidence when there is the possibility of substantial variation in seasonal warmth and snow cover.
PP51C-1511
The Sensitivity of the Oxygen Isotopes of Ice Core Sulfate to Changes in Climate and Chemistry: Implications for the Quantitative Interpretation of the Oxidizing Capacity of the Paleo-atmosphere
Variability in methane concentrations in the ice core record has been attributed mainly to changes in terrestrial (wetlands) climatic conditions. The oxidative capacity of the atmosphere controls the lifetime of reactive greenhouse gases and pollutants, and may contribute significantly to the observed variability in methane concentrations. However, there is currently very little information about past oxidant concentrations, making it difficult to determine to what extent variability in past methane concentrations can be attributed to terrestrial climatic conditions. The Δ17O of sulfate (Δ17O = δ17O - 0.5 δ18O) is a promising new proxy for the oxidative capacity of the atmosphere, as the unique isotopic signature of different oxidants is passed to sulfate during oxidation. Antarctic ice core measurements of the Δ17O of sulfate show large variability on the glacial-interglacial timescale [Alexander et al., 2002], suggesting a qualitatively higher oxidative capacity during cold glacial conditions. In contrast, both Antarctic [Kunasek et al., 2008] and Greenland [Alexander et al., 2008] ice cores show little change in the Δ17O of sulfate over the preindustrial-industrial time period. Using a global chemistry- climate model and a photochemical box model, we investigate the sensitivity of the oxygen isotopes of sulfate to changes in climate and chemistry on preindustrial-industrial and glacial-interglacial time scales, and quantify the role oxidants play in the observed variability of Δ17O of sulfate.
PP51C-1512
The importance of ship log data: reconstructing North Atlantic, European and Mediterranean Sea Level Pressure back to 1750
Climate variability at local to regional scale is to a large extent driven by processes exerted by the atmospheric circulation. Knowledge of its past changes are therefore of great importance to understand past and current changes in temperature and precipitation on various temporal and spatial scales. However, instrumental station pressure series allowing the construction of gridded sea level pressure (SLP) reanalyses only became widely available in the early 19th century. To reconstruct SLP fields further back in time, indirect information from documentary evidences or natural proxies had hitherto to be used. However, these reconstructions usually share predictors with existing temperature and precipitation reconstructions leading to circular reasoning in dynamical studies. Recently, wind information derived from ship logbooks became available as a new, direct and marine source. After extensive pre-processing to cope with the high spatial and temporal variability in the logbook availability, we combined these series with a few very long instrumental station pressure series from Europe and the eastern United States to seasonally reconstruct larger North Atlantic and European SLP fields back to 1750. Multivariate principal component regression as well as regularized expectation maximization (RegEM) were used to highlight differences due to the reconstruction methodology applied. This new reconstruction, exclusively based on information which is direct linked to the large-scale atmospheric circulation, has clearly higher skill than existing SLP reconstructions during winter and over the southern North Atlantic. This allows a more adequate representation of the Azores High and with more ship log information and instrumental pressure series becoming available from the north it is expected that the strength and location of the Icelandic Low can also be better resolved. This SLP field reconstruction, completely independent to temperature and precipitation reconstructions can then be used for e.g. dynamical studies relating past and current climate changes in the North Atlantic European realm to the large-scale circulation.
PP51C-1513
Estimated Marine Residence Times for Drowned Barbadian Paleoreefs
Fossil corals are used to estimate past sea level and also to calibrate 14C ages with the aid of U-Th and U-Pa dating methods. These coral fossils have often been subaerially exposed and thus are affected by diagenesis during their initial interaction with fresh water. In an effort to understand when such disequilibria in fossil coral reefs occurred, we have quantified our 'dissolution-cum-adsorption' model (Mey, 2008) for the uranium series disequilibria using a geometrical construction, based on the evolution of the activities in a 230Th/238U versus 234U/238U diagram for closed versus open systems. The traditional age equations for the uranium-series with excess daughters have been used to construct a relationship between (i) the angles of the equal age lines in the 230Th/238U versus 234U/238U activity diagrams, and (ii) the quantified angles of the regressed lines of several uranium series disequilibria trends from Barbados. Our results indicate that the severity of the Barbados uranium series disequilibria is not only explained by 234U and 230Th addition, but may also reflect a loss of 238U through dissolution of coral skeletal structure. The net effect is 238U removal, whereas 234U and 230Th remain; thus, the disequilibria for the extant coral increase the excess daughters' ratio. Our results further indicate that the activity of 234U is reduced (compared to 230Th), as would be expected in regard to the lower mobility of trapped 230Th. It is proposed that the major dissolution that caused the uranium series disequilibria occurred during one relatively short-lived event when the paleoreefs experienced the very first freshwater exposure. During this event, the diagenetic potential was at its maximum for redistribution of the uranium series; this then caused the 234U and the 230Th to behave in a systematic way, resulting in linear trends. The linear trends in the open system uranium series were set early, as shown in the 230Th/238U versus 234U/238U activity diagrams. The timing of the first exposure of the freshwater in the reefs is calculated based on the results of our new model. From the relationship between, (i) dissolution, (ii) in-grown 230Th, and (iii) excess 234U, we derived that the 60,000 old Marine Isotope stage 3 (MIS 3) reef was exposed to freshwater 36-38,000 years after growth in the marine environment. We have calculated these 'marine residence times' for the MIS 3 5a, 5c, 5e, 6.0, 7a and 7c reefs; our results correspond with the duration of the sea level high stand in each of the stages. References: Mey, J. L., (2008) The Uranium Series Diagenesis and the Morphology of Drowned Barbadian Paleoreefs, PhD dissertation, 325pp: Graduate Center, City University of New York, New York.
PP51C-1514
Neural Network: a Statistical Treatment Well Suited to Coral Skeleton Proxies
It is commonly admitted that chemical measurements of coral skeleton provide the best climate archives of tropical zone. However, studies earlier published failed to quantitatively reconstruct climate conditions of the last centuries. All the biases introduced by using an empirical SST calibration based on the classical isotopic thermometer have been called "vital effect". Indeed, by comparing chemical and metabolic characteristics of coral cultured in controlled conditions, we highlighted the intimate relationships between coral skeleton and physiology of the coral and its symbionts. It drives that several environmental factors determine one proxy. For instance, temperature may impact oxygen isotopic ratio both by kinetic isotopic fractionation but also through photosynthesis. Due to various incorporation processes, each proxy of a unique sample will be affected differently by one environmental factor, through a single biologic filter. In addition, the fractionations involved non-linear mechanisms, which may differently affect annual average or seasonal variability of a chemical parameter. Therefore, neural networks applied on time data series of several proxies measured on each sample along a coral core, successively following the filtered inter-annual and sub-annual signal, appear as well suited method.
PP51C-1515
Peptides Enhance Mg Content of Calcite: Toward a Process-Based Understanding of Proxy Models
Investigations of modern organisms relating magnesium content of calcified skeletons to temperature often exhibit unexplained deviations from the signature expected for inorganically precipitated calcite. These 'vital effects' are believed to have kinetic and taxonomic origins but the mechanistic basis for measured offsets remains unclear. A complicating factor is that mineralization is isolated from the external environment within an organic-rich matrix whose roles in mineralization are implicated but not well understood. Here we show that a simple hydrophilic peptide, sharing the same acidic character as macromolecules isolated from sites of calcification, increases the magnesium content of calcite up to 3 mol%. Using in situ AFM, we demonstrate that (Asp3Gly)6Asp3 also enhances growth rate and step edge energy of calcite compared to inorganic controls. Kinetic and thermodynamic measurements indicate that biomolecules interact with calcite surfaces to lower the energy barrier to desolvating the more strongly hydrated magnesium ion, thereby increasing the probability of its incorporation relative to calcium. Comparisons to previous studies that correlate Mg content of carbonate minerals with temperature show this peptide-induced Mg- enhancement is equivalent to offsets of several degrees Centigrade. The findings suggest local macromolecule chemistry influences Mg signatures— a plausible origin of vital effects. Further, studies of nonskeletal carbonates have long-asked whether the natural marine humic and protein substances found in sedimentary environments may influence mineralization. These insights provide a physical basis for anecdotal evidence that organic chemistry modulates the mineralization of inorganic carbonates.
PP51C-1516
Establishing a Baseline for Influence of Salinity on Mg Contents in Calcite
Over the last 20 years, salinity was assumed to have little impact on elemental signatures in the skeletal structures of biominerals. Recently however, Ferguson et al. (EPSL, 2008) demonstrated that Mg/Ca ratios in foramanifera tests correlate more strongly with salinity than with temperature. Implications of this finding could be significant because 1) salinity oscillations are significant and 2) paleo temperatures are largely reconstructed from carbonate sediments that may have experienced substantial salinity fluctuations and 3) if salinity indeed enhances Mg/Ca, then temperatures during the Last Glacial Maximum (Adkins and Schrag, GRL, 2001) may have been overestimated. In a new study, we are conducting experiments to measure Mg/Ca signatures in calcites grown in solutions of variable salinities and constant chemical driving force, or constant Mg and Ca activities. By this approach, we test the hypothesis that high salinity environments promote uptake of Mg and if true, we will assess the origin of this behavior. In previous work we have demonstrated that 1) calcite takes up more Mg as growth rate is accelerated, particularly in the presence of rate-enhancing hydrophilic peptides and 2) biochemical environments that favor cation desolvation also promote Mg uptake. From these insights, we predict that changes in Mg/Ca signature will arise due to one of these two effects: the kinetic effect, or reductions in the barrier to Mg desolvation relative to Ca. Indeed, kinetic measurements of growth rate using in situ atomic force microscopy show that calcite growth rate increases significantly with ionic strength. Mg content in the overgrowths was analyzed by Secondary Ion Mass Spectrometry (SIMS) using a Cameca IMS-7f Geo magnetic sector instrument. In conducting a carefully designed array of experiments, we expect to establish a baseline to determine the extent by which salinity affects Mg/Ca signatures in a system where biological influences are absent.
PP51C-1517
Oxygen Isotope Calibration Study of Modern Diatoms in Freshwater Environments: Implications for Biogenic Silica as a Paleoclimate Proxy
Oxygen isotope values of biogenic silica from diatom frustules are a commonly used proxy in freshwater and marine environments, and provide a valuable archive of paleoclimatic information such as temperature and water cycle processes. Advances in analytical techniques have made oxygen isotope measurements of diatom silica more robust; however, to date, there are multiple published fractionation factors for biogenic silica, with no general consensus on which is 'correct.' Previous studies (e.g. Moschen et al, 2005) demonstrated that there is no difference in SiO2-H2O fractionation between different size fractions of diatoms and, therefore, no species-dependent effects. The SiO2-H2O fractionation factors observed in laboratory grown diatoms analyzed by Brandriss et al. (1998) and modern lacustrine diatoms (Moschen et al., 2005) are in close agreement (τ = -0.2‰/°C) and are defined by the equations 1000lnα SiO2-H2O = 15.56 (103 T-1) - 20.92 and 1000lnα SiO2-H2O = 20.5 (103 T-1) - 36.2, respectively. However, these studies are not in agreement with other published SiO2-H2O fractionation factors for biogenic silica in marine and freshwater environments. In order to effectively utilize diatom δ18O values as a climate proxy, it is necessary to understand how oxygen isotopes are fractionated during silica frustule formation and identify potential errors in δ18O values obtained through different analytic/purification processes. Here we present oxygen isotope data from modern diatom species collected from a wide variety of natural riverine and lacustrine environments in northern New Mexico, USA. Temperatures at collection sites ranged from 5.5°C to 37.8°C. Preliminary isotope data indicate a SiO2-H2O fractionation factor identical to Brandriss et al. (1998). Additional experiments were undertaken to examine the effect of differing chemical purification techniques (i.e. HNO3, H2O2, and NaOH) on modern diatoms to see if processing techniques might affect the δ18O values of modern samples. Visual inspection of diatom frustules with a scanning electron microscope before and after treatment with HNO3 indicates no physical alteration of the frustule structure. To discount the possibility of oxygen exchange between diatom SiO2 and HNO3, samples were treated with an 18O-enriched nitric acid (1000‰), and the resulting δ18O values were essentially unchanged. Organic content following treatment with HNO3 was measured with an elemental analyzer and diatoms were considered to be pure SiO2 once weight percent carbon dropped below 0.01%. When diatoms were treated with H2O2 alone, significant organic material (>5 weight percent carbon) remained. Oxygen isotope values were obtained using a laser-extraction, stepwise fluorination technique that provides an additional visual confirmation of diatom purity. When pure F2 was introduced to the laser chamber during prefluorination, any sample with greater than 0.5 weight % carbon reacted violently to produce CF4 and O2 gas, and resulted in anomalous δ18O diatom values.
PP51C-1518
Morphological variation of Emiliania huxleyi a new paleosalinity proxy: Evidence from plankton analysis
We analysed the morphology of Emiliania huxleyi from globally distributed plankton samples and demonstrate that the size of E. huxleyi placoliths is highly correlated with in-situ sea surface water salinity. A standard step-wise multiple linear regression analysis was used to link morphological parameters of E. huxleyi to in-situ salinity and in-situ temperature and the best multiple regression model yielded an R2 = 0.87 with a standard residual error of 0.59 for in-situ salinity over a gradient from 32.6 to 38.8, whereas no significant correlation exists with temperature. Our analysis confirms the results recently published from sediment core-top samples (Bollmann and Herrle, 2007) and suggests that the morphology of E. huxleyi placoliths of recent and ancient sediments provide the potential to reconstruct sea surface salinities. However, the plankton-derived multiple regression model for in-situ salinities is different from that reported from Holocene sediments for annual mean salinities most probably due to taphonomic processes or biogeographically biased data sets (open ocean versus near shore).
PP51C-1519
Effects of pH and Temperature on Calcification of the Planktic Foraminifer O. universa: Insights from Culture Experiments.
The effect of future ocean acidification on marine calcifying organisms is uncertain. Because these calcifiers are intimately tied to ocean chemistry, nutrient cycles, and carbon export to the ocean floor, their sensitivity to pH could impact marine environments and global climate. Here, we present the results of culture experiments that explore the impacts of pH and temperature on the shell morphology and microstructure of the planktic foraminifer Orbulina universa between 18-26°C and pH 7.7-8.8. The largest and heaviest shells are formed at pH 8.4 and 22°C. Average shell diameter and wall thickness decrease ~17% and ~ 35% between pH 8.4 and 7.7. Extrapolation of these data to predicted surface ocean pH over the next century (~ pH 7.9) would imply a decrease in O. universa diameter and thickness of ~ 5% and 12%, respectively, during this period. Two classes of shell pores are observed in O. universa, which are distinctly different in diameter. Large pores are used for cytoplasm flow, food transport and symbiont-streaming in and out of the shell, whereas small pores are likely used for gas and ion exchange. Analysis of pore diameter and pore density (number of pores per 100 μm2 area) show that the large pore parameters do not change with either pH or temperature. In contrast, small pores show a significant increase in pore diameter and decrease in pore density across a pH range of 7.7-8.8. No relationship is found between temperature and porosity, which contradicts early suggestions that porosity can be used as a temperature proxy (Be et al, 1973). The observed negative correlation between pore density in and pH O. universa suggests that this property could serve as a qualitative pH proxy.
PP51C-1520
Should we care about diurnal temperatures when calculating the precipitation isotope thermometer?
Long records of the concentrations of stable isotopes of precipitation (SIPs) have long been used as proxies for regional and global climates for periods when meteorological measurements were not made. SIPs' longstanding correlation to local surface temperatures (in many locations) and molecular thermal dynamics have lead to many interpretations of variability in SIPs to be changes in local temperatures. In order to create accurate temperature-SIP transfer functions one needs to link modern SIP concentrations to temperatures of when precipitation happened. A well-sited example of complexities in the temperature-SIP relationships - For simplicity one may assume that annual precipitation occurred at the same time of year throughout a long SIP archive, however, it is possible that the timing of precipitation actually shifted from summer to winter months. If the temperature difference between the seasons is large the SIP archive could be wrongly interpreted as a several degree cooling in average annual temperatures. Temperature changes similar in magnitude to seasonal fluctuations are also observed throughout a given day. What would happen if precipitation shifted from mid-afternoon to nighttime events? This line of thinking implies that diurnal effects plausibly should be considered when calculating SIP-transfer functions. This is particularly convincing when precipitation for a region is powered by middle of the day (summer) heat causing convective precipitation or evening cooling increasing relative humidities near the land's surface. This study examines both theoretical and observed (5 locations within North America) surface temperatures at the time of precipitation throughout a day and estimates diurnal effects on SIP-transfer functions. Ultimately, one must ask, how high does condensation form, and what are daily temperature patterns at those heights?
PP51C-1521
High Resolution Time Series Cave Ventilation Processes and the Effects on Cave Air Chemistry and Drip Waters: Speleoclimatology and Proxy Calibration
Continuous high resolution (sub-hourly), long-term (Nov 2007-present) monitoring of cave air chemistry (Temperature, Relative Humidity, Barometric Pressure, Radon-222, CO2, Air flow, Wind speed and direction) in a shallow subtropical cave (Hollow Ridge) in N Florida reveals two major ventilation mechanisms: 1) ventilation driven by winds across the cave entrances, and 2) ventilation driven by density differences between atmospheric and cave air. The degree and type of ventilation strongly influence the 222Rn and CO2 of cave air, which in turn affects the timing and extent of calcite deposition in speleothems. The degree of ventilation is estimated using a cave air CO2-δ13CO2 Keeling Plot, or a simple radon deficiency model. Results show cave air has an atmospheric component ranging from 10-90%. During fall and winter, average CO2 (700 ppmv) and 222Rn (50-100 dpm/L) are lower than in spring and summer (CO2 = 1200 ppmv; 222Rn = 1000 dpm/L) due to increased winter ventilation. Decreased ventilation during the summer allows CO2 and 222Rn levels to rise. Winter daily ventilation is primarily a function of density gradients between cave air and atmospheric air, while summer daily ventilation is primarily a function of late morning NW-NE winds above the cave. Stable isotope analyses of drip water (fracture drip and pore flow drip) and aquifer water from Hollow Ridge agree with previous isotope studies of drip water at Florida Caverns State Park, 2 km to the NE. During summer, isotopic composition of pore flow drip water (δ18O -3.8 to -4.0 per mil; δD -17.3 to -20.2 per mil VSMOW) and aquifer water (δ18O -4.0 per mil; δD -18.0 to -21.1 per mil) are similar to average annual weighted isotopic composition of precipitation (δ18O -3.6 per mil) while fracture drip waters (δ18O -3 to -3.4 per mil; δD -11.9 to -14.3 per mil) likely reflect the isotopic composition of individual precipitation events. Pore flow drip waters δ18O are weakly correlated with drip rates (enriched δ18O during periods of higher drip rates) but show no correlation to precipitation amount. Knowledge of the type of drip flow is important when considering stalagmites for paleoclimate studies. A significant decrease in drip rate was observed from June (1034 drips/hour) through August 2008 (34 drips/hour). Higher water demands during summer months with increased evapotranspiration may be responsible for this decrease. A semi-diurnal drip rate cycle, negatively correlated with barometric pressure, is also observed throughout the period. This strong negative correlation is hypothesized to be controlled by atmospheric tidal oscillations. Observations into the fall and winter seasons should reveal seasonality, if any, and if there is an evapotranspiration effect present in the water cycle. High resolution studies of cave air chemistry and ventilation processes will enhance knowledge of the timing, extent, and isotopic and chemical composition of calcite deposition. When combined with drip water and precipitation isotope analyses, these studies will improve the understanding and interpretation of high- resolution (sub-annual) speleothem paleoclimate records.
PP51C-1522
Chain-Length Distribution and Hydrogen Isotopic Fraction of n-alkyl Lipids in Aquatic and Terrestrial Plants: Implications for Paleoclimate Reconstructions
Recent studies demonstrate that in lacustrine sediments, aquatic plant lipids (e.g., C22-fatty acid) record lake water D/H ratio variation, while long-chain fatty acids (C26-C32, major components of terrestrial plant leaf waxes), record D/H ratios of precipitation (especially in arid regions). However, there are insufficient literature data for the distribution and hydrogen isotopic fractionation of n-alkyl lipids in aquatic and terrestrial plants. In this study, we determined the chain-length distributions and D/H ratios of n-alkyl lipids from 17 aquatic plant species (9 emergent, 4 floating and 4 submerge species) and 13 terrestrial plant species (7 grasses and 6 trees) from Blood Pond, Massachusetts. Our results are consistent with previous studies and provide a solid basis for the paleoclimatic reconstruction using D/H ratios of aquatic and terrestrial plant biomarkers. In addition, systematic hydrogen isotopic analyses on leaf waxes, leaf, stem and soil waters from trees and grasses significantly advance our understanding of our previously observed large D/H ratio difference between tree and grass leaf waxes. Our data indicate that the observed difference is not due to differences in leaf water D/H ratios. In comparison with grasses, trees use greater proportion of D-enriched residual or stored carbohydrates (as opposed to current photosynthetic carbohydrates) for leaf wax biosynthesis, resulting in higher leaf wax D/H ratios. The residual carbohydrates are enriched in deuterium because of the preferential consumption of light-hydrogen substrates during plant metabolism.
PP51C-1523
Climate, Hydrology, and Lake Sediment
Components of lake sediment such as endogenic minerals, organic compounds, and fossils as well as chemistry and isotopes of those components are sources of information about past continental climate. No matter the size of the lake, the components of their sedimentary records reflect some integration of catchment processes operating on the landscape, within the ground water, and in the lake. Climate change influences and often drives change among these processes, yet the integration of the processes commonly yields unique records in lakes residing under the same climate. An estimate of the actual climate history from lake records as opposed to the apparent climate history from a given lake depends on how well we understand the modern system, or in the absence of a modern system, making use of known modern systems. We single out the effect of hydrology on lake records to illustrate the importance of understanding the modern system. The Waubay Lakes Chain in NE South Dakota is a series of lakes at different elevations and spill points that may be hydrologically connected, even coalescing, or may be isolated, though some remain hydrologically open (Niehus et al. 1999). Pickerel is an "upland" open freshwater lake whose level remains nearly constant, while Waubay is a shallow lake at intermediate elevation having more changes in level and salinity. Bitter, at the lowest elevation, receives spillover from Waubay during high water periods, exhibits wide ranges of salinity and elevation including going dry. Pickerel has a short residence time with lower δ18O and TDS than Waubay and Bitter (08/1991: -4.1 permil, -1.5 permil, and -1.8 permil; 204 mg/L, 6,420 mg/L, and 13,128mg/L respectively). The three lakes see the same climate, but their hydrologies lead to these different characteristics that interpreted in climate terms would suggest Pickerel resides under a wet climate while Bitter resides under a dry climate. The sediment records from these lakes though incompletely known appear to contain proxies that reflect the modern characteristics. Fossil ostracode species in Pickerel sediment (Schwalb & Dean, 1998) indicate that euryhaline species Fabaeformiscandona rawsoni dominanted during the mid-Holocene (8.4 to 6ka), while Candona ohioensis, a freshwater species was common at other times. In contrast, in the last 1000 years, F. rawsoni is common at Waubay and C. ohioensis only appears during high lake periods (Shapley et al., 2005). The ostracode fossil record from Bitter isn't known, but the modern assemblage is made up of Limnocythere staplini, a saline water ostracode, along with F. rawsoni, and other taxa. The modern day ostracode fauna in Bitter then implies a more saline environment than the entire Holocene record from Pickerel. True climate of the region is wetter than indicated by modern day Bitter and perhaps Waubay and drier than indicated by the modern Pickerel limnology and ostracode assemblages. Hydrology thus modifies the actual climate signal creating an apparent climate signal. Recognition of an approximation of an actual climate signal requires an understanding of how a given hydrology modifies the actual climate signal and if possible examination of records from several lakes each having a different hydrology. In the absence of an extant lake, ostracode assemblages offer insight into the hydrologic behavior of the past lake and so how its hydrology may have modified the climate signal. Niehus et al. (1999) USGS Water-Resources Investigations Report 99-4122, 166p. Schwalb & Dean (1998) Quaternary Science Reviews 21:1541-1554. Shapley et al. (2005) The Holocene 15:29-41
PP51C-1524
Ecology of lacustrine Crenarchaeota in Lake Superior: Implications for the application of the TEX86 temperature proxy
The TEX86 paleotemperature proxy, based on tetraether membrane lipids derived from aquatic Crenarchaeota has been applied in a variety of marine and lacustrine systems. A recent study analyzing a suite of 50 globally distributed lakes for TEX86 discovered that this proxy does not appear to work in all lake systems and that the TEX86 correlates well with both annual and winter lake surface water temperature in those systems where it does appear to work. Besides this observed empirical relationship between TEX86 values and lake surface temperatures, very little is known about the ecology of the crenarchaeota in lakes. We combined both biogeochemical and molecular techniques in a multiyear study of Lake Superior using both sediment trap collection of settling particulate matter over the annual cycle and filtration of suspended particulate matter from lake water to create vertical profiles of crenarchaeotal cell numbers and lipid concentrations to investigate the spatial and temporal ecology of the lacustrine Crenarchaeota. Initial results show that the flux of the tetraether lipids is highly seasonal and mainly occurs during two time periods in winter and spring. The flux-weighted TEX86-derived temperatures from the sediment trap material agrees with the TEX86 temperature from a sediment core top from the sampling site and mixed water temperatures during the two periods of highest flux within the error of the method. Spatially, lipids used in TEX86 are found throughout the water column when the Lake Superior is isothermal, but mainly in the hypolimnion when the lake is stratified. During stratification tetraether lipids in the eplimnion appear to reflect a surface water temperature, while the more abundant tetraether lipids in the hypolimnion reflect a deep water temperature. These data suggest that the TEX86 in sediments of Lake Superior mainly reflect the water temperatures of times of highest lipid flux, mixed with a smaller portion of lipids that are mainly produced in the hypolimnion.
PP51C-1525
Relating Phylogeny to Alkenone Distributions in Lacustrine Alkenone-Producing Haptophytes: Implications for Continental Paleotemperature Reconstructions
The alkenone unsaturation index (Uk'37) is a widely used surface water paleotemperature proxy in marine settings, but has seen limited use in lacustrine environments. On-going discovery of alkenone- containing lake sediments worldwide expand our ability to reconstruct continental paleotemperatures. However, disparate alkenone profiles among these lakes suggest a diversity of alkenone-producing organisms. The utility of the paleotemperature proxy is constrained by the accurate calibration of the Uk'37 against temperature for individual lakes. In this study, we report the findings from an 18S rRNA gene-based phylogenetic survey of globally distributed lakes containing alkenone-producing haptophyte algae to infer which haptophyte lineages likely possess common alkenone production pathways. Our phylogenetic analyses reveal that monophyletic groups of haptophytes possess similar lipid profiles. This correlation indicates the potential for applying a minimal number of calibrations to a multitude of diverse geographic settings. Furthermore, the expanded dataset of alkenone-containing lake sediments and robust phylogenetic analyses reflect the evolution of alkenone-producing haptophytes, and provide insights into the last common ancestor that was capable of alkenone production in the Cretaceous. The physiology, behavior, and culture conditions of a newly-isolated alkenone-producing haptophyte from polar waters possessing anomalous alkenone-concentrations, will also be discussed.
PP51C-1526
Environmental and Climatic Control on the Occurrence and Abundance of Long Chain Alkenones in Lakes of the Interior United States
Long chain alkenones are a class of temperature sensitive lipids with great potential for quantitative paleoclimatic reconstructions from continental locations. We have surveyed 57 lacustrine core-top sediments in the in the interior US for alkenones in order to constrain the environmental conditions that control the occurrence of alkenone-producing haptophytes in lakes and determine the possibility of using lacustrine alkenones to reconstruct temperature. Thirteen out of the 57 sites surveyed contain alkenones in surface sediments. Three of the 13 lakes contain abundant C37:4 alkenones as commonly found in lake sediments, whereas the rest of the lakes show a surprising absence of C37:4 alkenone. 18S Ribosomal DNA sequences amplified from the C37:4 containing lakes suggests the alkenone-producing haptophyte falls amongst the Isochrysis spp. and showcases a strong similarity to a haptophyte derived from Ace Lake, Antarctica, whereas the other lakes probably contain a different species. We also find cold, oligosaline conditions with high concentrations of sodium and sulfate promote high concentrations of alkenones. Salinity serves as a threshold control on the C37:4 alkenone such that C37:4 is only found in lakes with salinity above 2.74 g / L. Sites with the C37:4 alkenone plot along a previously published calibration. We also create a new alkenone temperature calibration using water column filters from Lake George spanning a temperature range of 8 to 24 degrees C. In addition, we obtain an alkenone- inferred temperature record from Lake George, ND for the instrumental period and a high-resolution (~20 year resolution) Holocene length record from Brush Lake, MT, in order to further assess the applicability of lacustrine alkenones for continental paleo-temperature reconstructions.