On the Application of Uk’37 In Narragansett Bay (Rhode Island, U.S.A.) by Jeffrey M. Salacup B.S., University of Massachusetts, Amherst, MA, 01003 M.Sc., University of Massachusetts, Amherst, MA, 01003 A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Geological Sciences at Brown University PROVIDENCE, RHODE ISLAND MAY 2014 © Copyright 2014 by Jeff Salacup This dissertation by Jeffrey M. Salacup is accepted in its present form by the Department of Geological Sciences as satisfying the dissertation requirement for the degree of Doctor of Philosophy. Date ______________ __________________________________ Timothy D. Herbert, Adviser Recommended to the Graduate Council Date ______________ __________________________________ Warren L. Prell, Co-adviser Date ______________ __________________________________ Yongsong Huang, Reader Date ______________ __________________________________ Meredith Hastings, Reader Date ______________ __________________________________ Steven Parman, Reader Date ______________ __________________________________ Thomas Cronin, Reader Approved by the Graduate Council Date ______________ __________________________________ Peter M. Weber, Dean of the Graduate School iii JEFFREY M. SALACUP DATE OF BIRTH: MARCH 18, 1978 PLACE OF BIRTH: W INCHESTER, MA, USA Earth System History Group Department of Geological Sciences Brown University 324 Brook St. Box. 1846 Providence, RI, 02912 EDUCATION PH.D. (expected completion Fall 2013) Geology, Brown University, Providence, RI, USA, On the Application of the Uk’37 Sea Surface Temperature Proxy in Marginal Settings: Narragansett Bay (Rhode Island, U.S.A.) Thesis Advisor: Drs. Timothy Herbert and Warren Prell M.S. (2008) Geology, University of Massachusetts-Amherst, MA, USA, The Effects of Sea Level on the Molecular and Isotopic Composition of Sediments in the Cretaceous Western Interior Seaway: Oceanic Anoxic Event 3, Mesa Verde, CO, USA Thesis Advisor: Dr. Steven Petsch B.S. (2006) Earth Systems, Cum Laude University of Massachusetts-Amherst, MA, USA, The biogeochemistry of a sub-surface fermentative/methanogenic consortia: Impact of biodegradation on hydrocarbon geochemistry of the Antrim Shale, Michigan Basin, U.S.A. Thesis advisor: Dr. Steven Petsch iv Courses: Sedimentary Geochemistry Physical Oceanography Organic and Biogeochemistry Aqueous Env. Geochemistry Organic & Inorganic Chemistry Environmental Evolution Climate Change Paleoclimatology Paleoceanography Climatology/Meteorology Isotope Geochemistry Oceanic Biogeochemistry Terrestrial Biogeochemistry Numerical Climate Change Geomicrobiology Estuarine Oceanography RESEARCH INTERESTS  Application of organic biomarkers and stable isotope ratios in the reconstruction of past environments, biota, and the evolution of the coupled Earth system processes  Elucidation of the biosynthesis and function of important organic geochemical paleo- environmental proxies  Oceanography of the northwest Atlantic Ocean during the Holocene  Carbon cycle dynamics and feedbacks associated with atmospheric/oceanic chemistry, climate stabilization, and the anthropogenic influence  Microbial activity and imprints on the geochemistry of anoxic environments v PROFESSIONAL EXPERIENCE Environmental Protection Agency STAR Fellow, 2010-2013 Department of Geological Sciences, Brown University Advisor: Drs. Timothy Herbert and Warren Prell -A new approach to assessing the anthropogenic impact on an urbanized estuary: sediment record of pre-historical and historical environmental change in Narragansett Bay, RI, USA Graduate Research Assistant, 2008 Department of Geological Sciences, Brown University Advisor: Dr. Timothy Herbert - method development toward the analytical isolation of alkenones with differing numbers of double-bonds Graduate Research Assistant, 2007 Department of Geosciences, University of Massachusetts-Amherst Advisor: Dr. Steven Petsch - extraction and separation of lipids from the Mancos Shale, Mesa Verde, CO (Cretaceous Western Interior Seaway) RCOM Summer Research Fellow, 2006 Research Center for Ocean Margins (RCOM), University of Bremen, Germany, Advisor: Dr. Kai-Uwe Hinrichs - Isolation of diagnostic lipid biomarkers from sediments recovered from the Fish Clay (K/T) at Stevns Klint, Denmark to investigate ecological and environmental change over the Cretaceous/Tertiary boundary. Molecular signatures of non-photosynthetic primary production during the collapse of a photosynthesis-driven “biological pump”. vi Undergraduate Research Assistant, 2005-2006 Department of Geosciences, University of Massachusetts-Amherst Advisors: Dr. Steven Petsch and Dr. Michael Formolo - Biogeochemistry of sub-surface fermentative/methanogenic consortia to analyze the impact of biodegradation on hydrocarbon geochemistry of the Antrim shale of the Michigan Basin. Investigating the influence of meteoric water recharge during the Last Glacial Maximum and implications for atmospheric concentrations of methane during termination. Tactical Aircraft Maintenance Craftsman, 1997-2002 United States Air Force, Rank – Technical Sergeant - Inspection, repair, and management of A10-A (Warthog) aircraft. - Deployed three times in support of war-time missions. RESEARCH SKILLS AND EXPERIENCE  Gas Chromatography- Mass Spectrometry-Flame Ionizing Detection o identification and quantification of biomarkers in complex natural samples o troubleshooting and maintenance of HP 6890 gas-chromatograph - Flame Ionizing Detector o data analysis using Chemstation and Xcalibur software  Compound Specific Stable Isotopes o Carbon and Hydrogen – compound isolation and analysis on HP 6890 GC coupled to a Thermo Delta V Advantage isotope-ratio mass spectrometer o data analysis using Isodat software vii  Elemental Stable Isotope Ratios (C, N, O) o organic carbon and total nitrogen - sample preparation and analysis on a Costech Elemental Analyzer coupled to a Thermo Delta V Advantage isotope-ratio mass spectrometer o inorganic carbon – sample preparation and analysis on a Kiel Device coupled to a Thermo Delta V Advantage isotope-ratio mass spectrometer o data analysis using Isodat software  Extraction Techniques o soxhlet apparatus o accelerated solvent extraction (ASE) o microwave-assisted extraction o sonication  Sample Preparation o column chromatography (Si-gel, NH4-Si-gel, AgNO3-Si-gel) for the isolation of discrete classes of organic compounds o saponification of fatty acid methyl esters (FAMEs) o Urea Adduction for the purification of alkenones from heavily impacted marginal sediments  Elemental Analysis o carbon (organic and inorganic) and total nitrogen on a Costech Elemental Analyzer o maintenance of Costech Elemental Analyzer o analysis on EAS software viii MANUSCRIPTS Salacup, J. Farmer, J. Herbert, T.D., and Prell, W.L. (in prep), Alkenone production with open ocean characteristics in a near-shore setting: Narragansett Bay, R.I. Salacup, J., Herbert, T.D., and Prell, W.L. (in prep), Alkenones in Narragansett Bay (Rhode Island, U.S.A.) Reveal a High-Resolution Record of North Atlantic Common Era Climate. Salacup, J., Herbert, T.D., and Prell, W.L. (in prep), A four year long record of water column alkenone ratios and concentrations from Narragansett Bay: seasonal variability and mean annual sea surface temperature. Salacup, J., Herbert, T.D., and Prell, W.L. (in prep), Landscape disturbance and nutrient mobilization in response to Colonization, Industrialization and Urbanization of the Narragansett Bay watershed. Salacup, J., Petsch, S.T., and Leckie, R.M. (in prep), The geochemical imprint of water column denitrification during OAE 3 (Coniancian-Santonian) in the Western Interior Seaway. Formolo, M.J., Salacup, J., Petsch, S.T., Martini, A.M., Nüsslein, K. (2008), Geology, v.36, no. 2, p.139-142, A new model linking atmospheric methane sources to Pleistocene glaciation via methanogenesis in sedimentary basins. POSTERS Salacup, J.M., Herbert, T.D., and Prell, W.L. (2013) The missing ocean – Generation of high resolution records of sea surface temperature for the Common Era, PAGES Open Science Meeting, Goa, India ix Salacup, J., Herbert, T.D., and Prell, W.L. (2012) But what is the ocean doing? Generation of high resolution records of sea surface temperature for the Common Era, Gordon Conference on Organic Geochemistry, Holderness, NH Salacup, J., Herbert, T.D., and Prell, W.L. (2012) A coupled biomarker-genetic approach to understanding the Uk’37 SST proxy in estuaries, Goldschmidt Geochemistry Conference, Montreal, QC Salacup, J.M., Gonclaves, M., Herbert, T.D., Prell, W.L. (2012) A new tool to track the environmental evolution of the Gulf of Maine: The Uk’37 sea surface temperature proxy, Workshop on the Gulf of Maine in a Changing Climate, Bowdoin University, ME Salacup, J. Herbert, T.D., and Prell, W. (2011) Time-series of water column alkenone and 18S rRNA confirm that Uk’37 is a viable SST proxy in Narragansett Bay, RI, 2011 AGU Fall Meeting, San Francisco, CA Salacup, J. (2011) A 450-Year-Long Record of Climate change from Narragansett Bay RI, USA, 2011 EPA STAR Graduate Fellowship Conference, Washington D.C. Salacup, J.M., Herbert, T.D., and Prell, W. (2010) Pushing open-ocean organic paleo- environmental proxies to the margin: Narragansett Bay, RI, 2009 AGU Fall Meeting, San Francisco, CA Salacup, J.M., Petsch, S.T., Leckie, R.M., (2007) Organic matter production and preservation during the Niobrara Cyclothem, OAE 3 in the Cretaceous Western Interior Seaway, 2007 AGU Fall Meeting, San Francisco, CA, USA Petsch, S.T., Formolo, M.J., Martini, A.M., Salacup, J.M., Nusslein, K., (2007) Hydrocarbon biodegradation in sedimentary rocks linked to atmospheric methane variations during continental deglaciation. 2007 EGU Spring Meeting, Vienna, Austria x Salacup, J.M., Formolo, M.J., Petsch, S.T., (2006) The Anaerobic Biodegradation of the Devonian Antrim Shale, presented at:  Five-Colleges Geosciences Undergraduate Research Symposium (2006) Amherst College, Amherst, MA  First annual University of Massachusetts-Amherst Geosciences Dept. research review (2006), UMASS-Amherst Geosciences, Amherst, MA Formolo, M.J., Petsch, S.T., Salacup, J.M., Waldron, P., Martini, A.M., Nusslein, K., (2006) Deep subsurface biodegradation of sedimentary organic matter in a methane-rich shale gas reservoir. 2006 AGU Fall Meeting, San Francisco, CA TALKS Salacup, J.M., Herbert, T.D., and Prell. W.L. (2012) But what is the ocean doing? Generation of high resolution records of sea surface temperature for the Common Era, Graduate Climate Conference, University of Washington’s Pack Forest Conference Center, WA Salacup, J., Herbert, T.D., and Prell, W. (2011) The Environmental History of Narragansett Bay: Insights from the Sediment Record, First Annual Rhode Island Sea Grant Research Symposium, University of Rhode Island, Kingston, RI Salacup, J.M., Farmer, J.R., Herbert, T.D., and Prell, W. (2009) Estuarine Alkenones: A High-Resolution Record of Sea-Surface Temperature from Narragansett Bay over the Past Millennium, 2009 AGU Fall Meeting, San Fransisco, CA Leckie, R.M., Salacup, J.M., Petsch, S.T., (2008) Oceanic Anoxic Event 3 (Coniacian- Santonian, Late Cretaceous) in the Western Interior Seaway, Invited Lecture, 2008 GSA Fall Meeting, Denver, CO xi TEACHING EXPERIENCE Earth: Evolution of a Habitable Planet (GEOL 0240) Teaching Assistant, Spring 2009, Brown University Stratigraphy and Sedimentation (GEOL 1240) Teaching Assistant, Fall 2008, Brown University Sedimentary Geochemistry (GEO-517) Teaching Assistant, Fall 2007, University of Massachusetts-Amherst Geological Mapping (GEO-331) Teaching Assistant, Fall 2007, University of Massachusetts-Amherst History of the Earth (GEO-201) Teaching Assistant, Spring 2007 & 2008, University of Massachusetts-Amherst, responsible for lecturing and instructing during two weekly lab sections Introductory Oceanography (GEO-103) Teaching Assistant, Fall 2006, University of Massachusetts-Amherst MENTORED STUDENTS Jesse Farmer, May 2008 – May 2009, Evaluating the Use of Organic Paleotemperature Proxies in Estuaries: Application of Uk’37 and TEX86 to Narragansett Bay, Rhode Island, Senior Research Thesis, Brown University Mariama Goncalves, Summer 2012, Holocene climate change recorded in the oxygen isotopes of forams, NSF GK12 High School Summer Student Internship. xii FIELD EXPERIENCE Narragansett Bay, 2009-2013 – surveys of Narragansett Bay for the collection of water samples for particulate organic matter extractions toward the better understanding of the biomarker-based UK37 and TEX86 sea-surface temperature proxies; collection of sediment cores via traditional piston push-coring, vibra- coring, and grab sampling Eastern Equatorial Pacific, R/V Knorr, 2009 – retrieval of deep sea sediments on a transect from the Galapagos Platform, along the Carnegie Ridge, and into the Bay of Guayaquil, Peru employing the R/V Knorr’s new long piston coring capability, gravity coring, and multi-coring Colorado Plateau, 2007 – investigation of sequence stratigraphy associated with the Greenhorn and Niobrara Cyclothems of the Cretaceous Western Interior Seaway Long Island Sound, 2006 – water and sediment sampling of two locations in Long Island Sound to include freeze and gravity coring, Thompson grabs, and pore- water extractions AWARDS Rhode Island Science and Technology and Advisory Council (STAC) Grant, Spring 2012-Spring 2013, ($200,000), “Understanding coastal environmental change, past, present, and future: A novel approach combining algal physiology, genetics and lipid biomarkers”, (A proposal I wrote to support personal research that was funded in Dr. Timothy Herbert’s name) Environmental Protection Agency - Science and Technology Achieves Results (STAR) Fellow, Fall 2010 – present (award through Aug, 2013), Brown University (~ $111,000) xiii Outstanding Teaching Assistant, Fall 2006-Spring 2007, University of Massachusetts- Amherst HTU Smith Memorial Geology Award, 2007, University of Massachusetts-Amherst (~ $1000) RCOM Summer Research Fellow, 2006, Research Center for Ocean Margins, University of Bremen, Germany xiv To my parents, my sister, and my wife, for everything. I love you. To Hannah Bee, my little angel xv ACKNOWLEDGEMENTS First, I would like to thank my advisers Tim Herbert and Warren Prell for their interest, trust, patience, advice, criticism, support, and knowledge. Without you this work would not have been possible. I would also like to thank the rest of the T.D.H Lab past and present (Sam, Rocio, Alexa, Caitlin, Emily, April, Dan, Angel, Shuo) for helping me wade through and hone my ideas through countless lab group meetings. Thank you to my (then) undergraduate assistant Jesse Farmer for helping me get this proxy working in Narragansett Bay by spending innumerable hours troubleshooting in the hood. To the GC, for hanging in there long enough for me to finish, despite all the shit I put through it, I’m sorry. Thank you to all my grad-school peeps who made lunches by the rocks, evenings at the GCB, and weekends in the mountains happen. Shannon, Bronwen, Jess, Chris, Nick, Elizabeth, Getz, Danielle, Aron, Alex, Scott, Peter, Kate, Willy D, Will L., Mark, Colin, Steph, Leah, and the GeoPathfinder group (Tim, Kat, Mary, Nick, Lauren, Sandra, and Charlie) made Providence habitable when I was here and the mountains greener when I was lucky enough to be in them. I would like to thank Steve Petsch, Mark Leckie, Mike Formolo, Carrie Petrik, Kenna Wilkie, Julio Sepulveda, and Elizabeth Gordon for being there to help me through my first days in the lab, for teaching me how to do research and revealing that I love it, and xvi for first showing me science could be a rewarding, supportive, exciting, and self-directed career option. I would like to thank the wonderful ladies in the front office without whom the department would likely collapse. I would like to thank my homeboys (Tom, Andy, and Joe) and U.S.A.F. brothers (Kevin and Nick) for being there as we turned from children into men. Anyone who knows me knows you’re among the most important people in my life. I love you guys and am honored to be your friend. I would like to thank my Mom, for getting me started off on the right foot, buying my first chemistry set, cheering my early successes and soothing my failures, and for always being there, then and now. I would like to thank my Dad, for teaching me about responsibility, honor, hard work, and family, and for making sure we never wanted for anything. I wish you could be here to see me now. I want to thank my Sister, for being my oldest friend, for growing up with me, for all of your support, and for introducing me to my wife. Lastly, I need to thank my wife, for her patience, support, love, kindness, understanding, trust, and general awesomeness. I finished babe, it’s done, I love you! xvii Table of Contents Introduction ………………………………………………..………………………………………………………………1 Chapter One:..……………………………………………………………………………………………………………9 Alkenone production with open ocean characteristics in a near-shore setting: Narragansett Bay (Rhode Island, U.S.A.) 1.1 Abstract ………………………………………………………………………..…………………………………………..10 1.2 Introduction …..………………………………………………………………………………………………………….10 1.2.1 Alkenones as a Proxy for Estuarine Sea Surface Temperature ..……………...12 1.2.2 Narragansett Bay …………………………………………….…………………………………………14 1.3 Methods ………………………………………………………………………………..…………………………………..15 1.3.1 Water-Column Survey ……………………………………………………………..…………………15 1.3.2 Sediment Cores ……………………………………………………..…………………………………...15 1.3.3 Alkenone Analyses …………………………………………………………………..…………………17 1.3.4 Development of a long instrumentally-derived SST record ……..………………..18 1.4 Results ………………………………………………………………………………………………….………………….20 1.4.1 Water-column Uk’37 ………………………………………………………………..……………………20 1.4.2 Sediment Uk’37……………………………………………………………………………….……………20 1.4.3 Alkenone Indices ………………………………………………………………………….…………….20 1.4.4 Age Control and Chemotratigraphy ……………………………….…………………………..21 1.5 Discussion ……………………………………………………………………………………………..…………………22 1.5.1 Evidence for in situ alkenone production………………………………..…………………..22 1.5.2 Characterizing NB Alkenone Production ……………………………..…………………….22 1.5.3 Uk’37 SST offsets and production seasonality ………………………..…………………...23 1.5.4 Response of Uk’37 to SST …………………………………………………..……………………….24 1.5.5 Estimation of Uncertainty and Reproducibility …………………..…….…………………26 1.6 Conclusions ………………………………………………………………………………..……………………………27 Chapter Two: ………………………………………………………………………………..………………………….47 A four year long record of water column alkenone ratios and concentrations from Narragansett Bay: salinity segregation, seasonal variability, and mean annual sea surface temperature. 2.1 Abstract ………………………………………………………………………………..…………………………………..48 2.2 Introduction ……………………………………………………………………………..……….………………………49 2.3 Methods …………………………………………………………………………………………..………………………..51 2.3.1 Environmental Data ……………………………………………………………………………...…….51 2.3.2 Sampling Scheme……………………………………………………………………………………….52 2.3.3 Alkenone Analysis ………………………………………………………………………………………53 xviii 2.4 Results ………………………………………………………………..…………………………………………………….54 2.5 Discussion ………………………………………………………………….………………………………….…………56 2.5.1 The annual cycle ……………………….…………………………………….…………….……………56 2.5.2 Salinity segregation of alkenone producers in NB …………………….………...……57 2.5.3 Deviations between Uk’37-inferred and instrumental SST……………………..……58 2.3.5.1 Warm Uk’37 offset during winter………………………………………………….59 2.3.5.2 Cool Uk’37 offset during summer…………………………………..…………….60 2.6 Conclusions ………………………………………………………..……………………………………………………63 Chapter Three: ………………………………………………………………………………..………………………76 Alkenones in Narragansett Bay (Rhode Island, U.S.A.) Reveal a High-Resolution Record of North Atlantic Common Era Climate 3.1 Abstract ……………………………………………………………..……………………………………………………..77 3.2 Introduction ………………………………………………………………………..…………………………………….77 3.3 Chronology ………………………………………………………………………..……………………………………..80 3.4 Temperature Reconstruction …………………………………………………………………………..….....81 3.5 Conclusions ……………………………………………………………………..………………………………………89 Chapter Four: ………………………………………………………………………………………………………101 Seven hundred year record of land clearance, nutrient mobilization, and productivity in Narragansett Bay, RI 4.1 Abstract …………………………………………………………………………………………………………………..102 4.2 Introduction ……………………………………………………………………………………………..…………….104 4.2.1 Geological Setting ……………………………………………………………………………..……..105 4.2.2 Pre-Contact (1300-1600) ……………………………………………………………..…………..106 4.2.3 Colonization (1600-1800) ………………………………………………………………..……….107 4.2.4 The Industrial Revolution and Urbanization (1800-present) ……………….…..108 4.2.5 Proxies of environmental change ……………………………………..………..…………….109 4.3 Methods …………………………………………………..……………………………………………………………...110 4.3.1 Water-Column Survey …………………………………………………………..………………….110 4.3.2 Nitrogen Isotopes ……………………………………………………………..………………………110 4.3.3 Benthic Foraminiferal Counts ……………………………………………………………..……111 4.3.4 Ambrosia Pollen …………………………………………………………..…………………………...111 4.3.5 GDGTs …………………………………………………………….……………………………………….111 4.3.6 Sediment Cores ………………………………………………………………..………………………112 xix 4.4 Results …………………………………………………………………..………………………………………………..113 4.4.1 Sediment Accumulation Rates …………………………………………………………….…..113 4.4.2 GDGTs ……………………………………………………………………..……………………………....113 4.4.3 Ambrosia Pollen………………………………………………………….…...……………………… 115 4.4.4 Magnetic Susceptibility …………………………………………….………………..…………….116 4.4.5 Nitrogen isotopes…………………………………………………………….…………..……………116 4.4.6 Benthic Foraminifera ………………………………………………….…………………..………. 116 4.5 Discussion ………………………………………………………………….………………………………………….117 4.5.1 Sources of GDGTs to Narragansett Bay …………………………………………………117 4.5.2 pre-Contact (1300-1600) Landscape Stability …………………………………...…. 119 4.5.3 The Colonial Disturbance (1600-1800) …………………………………..………..………120 4.5.4 Industrialization and Urbanization (1800-present)....……………………………… 122 4.6 Conclusions ……………………………………………………..…………………………………………………….123 xx List of Tables Table 1.1 Identification, location, and description of water quality monitoring buoys, NMFS cruise sampling stations, and cores used in this work..……………….……………………..29 Table 1.2 Dates, concentrations (ng/L), and locations of all water-column alkenone detections……...……………………………………………………………………………….…………………………………30 Table 1.3 Age-depth constraints used for NB cores……………………………………………………..31 Table 3.1 Age-depth constraints used for NB cores….…………………………………………………..90 Table 4.1 Results of correlation (r) and significance (p) tests between GDGTs and environmental variables……………………………………………………………………………………………...….124 xxi List of Figures Figure 1.1 Map of Narragansett Bay and points of interest (see Table 1.1): core locations (stars; GB = Greenwich Bay (NB12); PC = Potter’s Cove (NB25); FI = Fox Island (NB44v)), surface water POM sampling locations (numbered circles). GSO = Graduate School of Oceanography, TFG = TF Green International Airport, CP = Conimicut Point buoy station. Arrows represent major sources of fresh water. Stippled lines denote the ecofunctional boundaries between sections of the Bay [Costa-Pierce and Desbonnet, 2008]................................................................................................................................................................................32 Figure 1.2 Plots of the annual cycle of near surface (A) SST, (B) SSS, and (C) Chl for Conimicut Point (CP, green), Greenwich Bay (GB, red), Potter’s Cove (PC, blue), and the Graduate School of Oceanography (GSO, black, see Fig.1.1) provide important context for our water column and sedimentary Uk’37 results. A number of year’s data (CP = 2009 - 2012; GB = 2007 - 2010; PC = 1996 – 2010; GSO = 1996 - 2010) was averaged then treated with a low pass filter to produce a monthly smooth. The patterns depicted are resistant to the length of time averaged and/or the removal of any one year’s data…………………………………………………………………………………………….………………………….33 Figure 1.3 Development of a 115 year-long inferred NB SST record. (A) Providence monthly average air temperature at TF Green Airport (solid line; Fig. 1.1, TFG; Table 1.1) and Conimicut Point monthly average SST (dashed line; Fig. 1.1, CP; Table 1.1) from 1999 to 2008. (B) Regression between monthly average TFG air temperature and CP SST (n=108) used to construct a 115-year-long inferred SST record for Conimicut Point (SSTAIR-CP). (C, D, and E) Regressions between CP and GB, PC, and GSO (FI), respectively. Buoy data are from 2007 and 2008 (Table 1.1; n = 24). These regressions were used to transform SSTAIR-CP into 115-year-long SST records for each core site (SSTAIR-GB, PC, and FI). These results compare well with an instrumental SST record from Woods Hole, MA (Shearman and Lentz, 2010) confirming the utility of our regression approach…………………………………………………………………………….…………………………………………….34 Figure 1.4 Plot of POM-inferred surface water alkenone concentrations during the 2010 NMFS cruises (Fig. 1.2; Table 1.1). Upper Bay / Providence River concentrations are hash-marked, mid- and lower Bay concentrations are in white. Columns are an average of all alkenone detections in a given month. Error bars represent the maximum and minimum concentrations for that month. All data is provided in Table 1.3. The results show one bloom in the upper Bay and Providence River in the spring (Mar and Apr) and another in the lower Bay in the late summer / fall (Jul – Nov). A representative chromatogram showing a high contribution of the C37:4 moiety for the spring bloom suggests a brackish alkenone producing population inhabits the upper Bay at this time. The chromatogram for the late summer / fall bloom lacks C37:4, consistent with a marine producer. The C38 alkenone fingerprints also differ between the two blooms further suggesting different producers………………………………………………………………………………………………………………………….35 xxii Figure 1.5 Uk’37 depth profiles for each core showing age control points for 210Pb (solid bar), [Pb] (open squares), Ambrosia pollen (open triangles), and 14C (filled squares). All cores are tied to one another at 1850 via the [Pb] increase. Cores NB12 and NB25 are tied to each other at 1700 via increases in sedimentary Ambrosia…………………………………………………………………………………………….…………………………….36 Figure 1.6 Plots of alkenone indices used to classify NB alkenone fingerprints as ‘marine-like’. (A) plot of C37total vs. C38total for all sediment core data showing a strong positive relationship with a slope near 1 (Herbert, 2003). Our slope is higher than 1 for reasons discussed in the text. (B) A plot of Uk’37 vs. Uk’38me, another tracer of SST, for sediment and water column samples from this study compared with those of Herbert (2003) show good agreement of NB alkenones with the open-ocean trend………………………………………………………………………………………………….………………………………37 Figure 1.7 Key stratigraphic and age control results used to develop age models for Narragansett Bay cores. Plots of sedimentary profiles of [Pb] (squares), Ambrosia pollen (trinagles), 210Pb age estimates (open circles), and 14C ages (filled squares) for cores (A) NB12, (B) NB25, and (C) NB44v. Filled circles represent modern aged core top assumptions. Detection and removal of an instantaneous depositional event in core NB12 ((A), dashed lines, see text) resulted in the profiles depicted in (D). Arrows represent important age control points discussed in the text……………………………………………………………………………………………...…………………………………….38 Figure 1.8 Age models for (A) GB, (B) PC, and (C) FI based 210Pb excess (circles), [Pb] increases (open squares), Ambrosia pollen (triangles), and 14C (filled squares). Age- depth relationships were calculated using cubic splines (NB12 and NB25) or linear fits (NB44v) and assume core top reflects modern sedimentation (filled circles). Average sedimentation rates on the order of millimeters per year make sub-decadal sampling readily achievable……………………………………………………………………………………………………...……..39 Figure 1.9 Plot of the difference in instrumental SST between Greenwich Bay (GB) and Potter’s Cove (PC) defined as GB SST – PC SST. The Uk’37 offset between NB12 and NB25 suggests that the alkenones are produced during a part of the year when GB is warmer than PC (dashed lines). Vertical green bars represent the periods of alkenone production inferred by our surface water samples (Table 1.2, Fig. 1.4). Darker color represents generally higher alkenone concentrations…………………………….……………………………………………………………………………………40 Figure 1.10 Comparison of Uk’37 reconstructed SSTs at SSTAIR at each site using the calibration of Prahl et al. [1988]. The thick black line is a stack of all three records. Significant co-variance between Uk’37 and SSTAIR at two of our sites (see text) provides strong evidence that Uk’37 is responding directly to growth water temperature. Significant intra-core variability (see text) suggests in situ gradients of salinity, nutrients, and productivity do not compromise the interpretation of Uk’37 in NB. Instead, an in situ gradient in SST captured by our reconstructions implies Uk’37 in NB is a sensitive indicator of local conditions on decadal time scales. Between 1905 and 2005, the rates of warming reflected in both the Uk’37 (1.8°C/100yr ±0.4°C) and instrumentally inferred (1.7°C/100yr ±0.2°C) SST records are within error…………………………………………....……………41 xxiii Figure 2.1 . Map of Narragansett Bay and points of interest: Sampling stations (filled circles) GSO = University of Rhode Island’s Graduate School of Oceanography; RWU = Roger Williams University; Field’s Point. Arrows represent major sources of fresh water. Core and surface sediment locations from Salacup et al [in prep-a; in prep-b] (stars; GB = Greenwich Bay (NB12/45v); PC = Potter’s Cove (NB25/42v); FI = Fox Island (NB44v)). Stippled lines denote the ecofunctional boundaries between sections of the Bay [Costa- Pierce and Desbonnet, 2008]……………………………………………………………………..……………………64 Figure 2.2 . Plots of the annual cycle of near surface (A) SST, (B) salinity, and (C) Chl for Conimicut Point (CP, green), Greenwich Bay (GB, red), Potter’s Cove (PC, blue), and the Graduate School of Oceanography (GSO, black, see Fig. 1.1) provide important context for our Uk’37 results. A number of year’s data (CP = 2009 - 2012; GB = 2007 - 2010; PC = 1996 – 2010; GSO = 1996 - 2010) was averaged then treated with a low pass filter to produce a monthly smooth. The patterns depicted are resistant to the length of time averaged and/or the removal of any one year’s data………………………………….………65 Figure 2.3 Plots of Uk’37 inferred SST and alkenone concentration comparing these two variables between samples taken after 12-24 hours in a settling tank and pre-filtration (circles) and those taken ‘raw’ directly from NB (diamonds)………..………………………………….66 Figure 2.4 Plot of each years’ time series of alkenone ratios and concentrations onto a generic annual cycle and smoothed with a polynomial fit to highlight the 4 year average annual cycle. Alkenone ratios and concentrations are compared with instrumentally measured SST and chlorophyll measurements (GSO)………………………………………………….67 Figure 2.5 Plots comparing the time series of Uk’37-inferred and instrumental SST (A), alkenone concentrations (B), and nutrients (C). (A) Comparison of Uk’37-inferred SST (circles and diamonds) with instrumental SST (solid black line) highlights periods of both agreement and deviation. Periods in summer during which Uk’37 SST is cooler than instrumentally measured SST coincide with periods of (sometimes slightly) increased alkenone concentrations (B). (A and B) Good agreement in Uk’37 SST and alkenone concentration is noted between samples taken from settled prefiltered water (circles) and raw water (diamonds). Periods of cool summer Uk’37 SST may or may not coincide with times of nutrient nitrogen depletion (C, solid line, phosphate = stipled line).………………….68 Figure 2.6 A zoomed-in plot of C37:4-rich alkenone production at Fields Point in April of 2013 showing alkenone concentration and fluorescence (A), salinity (B) and Uk’37- inferred and instrumental SST………………………………………………………………………………...………69 xxiv Figure 3.1 (left) Major surface currents of the North Atlantic Ocean: GS = Gulf Stream, AC = Azores Current, NAC = North Atlantic Current, LC = Labrador Current, IC= Irminger Current, EG = East Greenland Current, WG = West Greenland Current, LC = Labrador Current, CSWS = Coupled Slope Water System. White circles and diamonds show the locations of records discussed in the text. The white square shows the location of the coarsely resolved (~60 year spacing) Sargasso Sea record of Keigwin [1996] included in the IPCC compilation but not discussed here. Together, the circles (excluding Greenland and Icelend) and square reflect the totality, and paucity, of SST records at least 1000 years long that are included in the IPCC Common Era reconstruction. (inset) Location of Narragansett Bay cores (white circles). (right) Coupled Slope Water System [Pickart et al., 1999]: ATSW = Atlantic Temperate Sea Water, LSSW = Labrador Sub- Arctic Sea Water. Dashed lines and arrows depict the movement of the boundary between LSSW and ATSW in relation to changes in the AO/NAO.……………..……………………………………………………………………………………………..……………..92 Figure 3.2 (top) Age models for cores NB25/42 (left), NB12/45 (middle), NB44 (right) were established using 14C of mollusks and foraminifera (filled squares), pollen horizons (triangles), [Pb] increases (open squares), and 210Pb (circles). Black and red lines are the mean and 2σ error age determinations, respectively. (bottom) The evolution of sedimentation rate (black) and 2σ age error (red) through time………………..…..………………93 Figure 3.3 Plots comparing our raw (top) SST reconstructions for cores NB25/42 (blue diamonds), NB12/45 (red squares), and NB 44v (green triangles). The individual NB SST records are offset by an amount consistent with in situ SST gradients. Sample spacing of the resulting SST stack averaged 4 years (dashed line).……………………………………………………….……………………………………………..……………94 Figure 3.4 Narragansett Bay SST reconstructions compared with other North Atlantic records. Arranged from north to south are: Voring Plateau August SST [Anderssen et al. 2003], coastal Iceland warm season SST [Sicre et al., 2011], Greenland air temperature [Kobashi et al., 2011], Narragansett Bay [this work], Chesapeake Bay warm season SST [Cronin et al., 2010], and Gulf of Mexico mean annual SST [Richey et al., 2007]. The horizontal red bars and vertical blue arrows point out times corresponding to the Medieval Climate Anomaly and the maximum Little Ice Age cooling, respectively. Recent solar minima are shown with open rectangles at the top of the plot.……………………………...95 Figure 3.5 (A) Comparison of instrumental Woods Hole SST [top; Shearman and Lentz, 2010] with the North Atlantic Oscillation [middle; NAO; provided by the Climate Analysis Section, NCAR, Boulder, USA], and instrumental Chesapeake Bay SST [lower; Cronin personal comm.]. A significant (p<0.01) and positive relationship between the NAO and Woods Hole SST is evident over the past 50 years, consistent with findings from the nearby Gulf of Maine [Pickart et al., 1999]. SST at Woods Hole has warmed ~0.47°C since 1950 while Chesapeake Bay SST has cooled 0.75°C. (B) Comparison of Uk’37- inferred Narragansett Bay SST with a tree ring based reconstruction of the Arctic Oscillation [D’Arrigo et al., 2003] shows a significant (p<0.0001) and positive relationship between the two over the past ~350 years, consistent with NAO-based findings and suggesting high latitude atmospheric dynamics like the AO and NAO impact regional oceanography…………………………………………………………….………………………..………………………….96 xxv Figure 4.1 Map of Narragansett Bay and points of interest: core locations (stars; Greenwich Bay (NB12/45v); Potter’s Cove (NB25/42v); Fox Island (NB44v)), surface water POM sampling locations (open circles); surface sediment samples (plus sign). Stippled lines denote the functional boundary between sections of the Bay [Costa-Pierce and Desbonnet, 2008]……….………………………………………………………………………………………..125 Figure 4.2 The north-south gradients of nutrient induced productivity and pollution is evident in two plots showing the concentrations of organic matter (a proxy for productivity) and Pb (an industrial pollutant) in the surface sediments of Narragansett Bay. Maps adapted from Murray et al. [2007].………………………………………………………………126 Figure 4.3 Structures and representative m/z ratios for brGDGTs analyzed in this study.………………………………………………………………………………………………………………………………127 Figure 4.4 Age models for cores NB25/42v (left), NB12/45v (middle), NB44v (right) were established using 14C of mollusks and foraminifera (filled squares), pollen horizons (triangles), [Pb] increases (open squares), and 210Pb (circles). Sediment accumulation rates are shown below the appropriate core. ………………..…………………………………………..128 Figure 4.5 Profiles of brGDGT concentrations and BIT Index in water column (left) and surface sediment (right) samples with latitude down Bay away from the major sources of terrestrial sediments at the Providence and Taunton Rivers.……………………………………….129 Figure 4.6 Plots of the relative contributions of the three different branched GDGT groups in surface water (cirlces) and surface sediment (triangels) with latitude..…….….130 Figure 4.7 Timeseries of brGDGT concentration, crenarchaeol concentration, and the BIT Index at Greenwich Bay (cirlces), Potter’s Cove (squares), and Fox Island (triangles)……….…………………………………………………………………………………………………………...131 Figure 4.8 Timeseries of the relative contributions of the three different branched GDGT groups.……………………………………………………………………………..…………………………..…….…………..132 Figure 4.9 Scatter plot of our PCA results. GDGTs are shown in black circles, environmental data in open circles.………………………………………………………………...…..…………133 Figure 4.10 Timeseries of (A) % Ambrosia pollen, (B) bulk sediment nitrogen isotopes (δ15N); (C) benthic foraminiferal abundance, and (D) magnetic susceptibility (MS). Nitrogen fluxes estimated by Nixon et al. [2008] are co-plotted in (C, dashed line). Dashed lines in (B) represent the average δ15N for the given interval of time…………...…134 xxvi Introduction ____________________________________________________ Today, understanding the evolution of the coupled Earth system (litho-, cryo-, atmos-, hydro-, and biospheres) is critical as we realize and begin to deal with the repercussions of our participation with an important part of the Earth system, the carbon cycle. Knowing how the Earth system will respond to the human release of greenhouse gases (GHG) requires knowing something about how it has responded to similar events, if they exist, in the past. To do this requires two things: 1) a robust spatial and temporal characterization of both natural and human induced climate variability (particularly during the last 2000 years, the Common Era) [eg. IPCC, 2007], and 2) an understanding of how past large changes in atmospheric GHG inventories have affected natural climatic and biotic variability [eg. Gingerich, 2006]. Knowledge of these will aid in the ultimate goal of predicting and mitigating the impacts of the ongoing release of GHGs to the atmosphere. The story of Earth’s climatic and biotic evolution is archived in its sedimentary basins and decades of paleoclimatologists have developed and applied a wide array of ever-more quantitative proxies (preserved physical remains representing one or more climate variables during the time from which they come) to interpret that evolution [Imbrie et al., 1992; Webb, 1986; Zachos et al., 2001]. Recent advances in analytical and organic geochemistry provide new biomarker-based proxies, such as the Uk’37 sea surface temperature proxy [Brassell et al., 1986], the TEX86 water temperature proxy [Schouten et al., 2002], and the MBT/CBT air temperature proxy [Weijers et al., 2007] shedding light on parts Earth’s history that were previously unreadable or poorly understood. Building on these advances, this thesis focuses on the development and application of the Uk’37 SST biomarker proxy for use in estuarine sediments. 1 The Uk’37 proxy is based on the ratio of two polyunsaturated long-chain alkyl ketones, called alkenones, produced by specific haptophyte algae [Conte et al., 1994; Volkman et al., 1995] and initially correlated with sediment core δ18O ratios from planktonic foraminifera, known to partially record SST [Brassell et al., 1986]. Culture [Prahl and Wakeham, 1987; Prahl et al., 1988] and core-top sediment [Müller et al., 1998] calibration studies led to the development of the Uk’37 Index as a quantitative SST proxy [Herbert, 2003]. Reconstructed Uk’37 temperatures correlate best with mean annual SST for a variety of climate and haptophyte production regimes in the global ocean [Conte et al., 2006] where alkenones are primarily produced by the coccolithophores Emiliania huxleyi and Gephyrocapsa oceanica [Conte et al., 1994; Volkman et al., 1995]. Whether they are the primary alkenone producers in estuarine and near-shore settings is unclear. Instead, the genera Isochrysis and Chrysotila, which have a different relationship to SST, may dominate [Versteegh et al., 2001; Volkman et al., 1995]. Importantly, the alkenones produced by Isochrysis and Chrysotila may contain high contributions of the C37:4 alkenone not normally detected in open marine environments, and which may signal a ‘brackish’ environment in which Uk’37 functions poorly [Mercer et al., 2005; Schulz, 2000]. Estuaries provide an attractive opportunity to reconstruct biomarker-based Common Era paleo-environments because they often contain thick sequences of Common Era sediment that have accumulated much faster than in the open ocean [Boothroyd, 2008] making sub-decadal sampling readily achievable. Estuaries are transition zones between marine and fresh waters, thus their sediments concurrently capture both terrestrial and marine responses to environmental change. As such, estuarine sediments are important to understanding future local-to-regional level ecosystem responses in the face of global climate change. The development and application of organic biomarker proxies hold promise in these dynamic, carbonate-deprived, mud-rich settings. 2 Narragansett Bay is a high-salinity, relatively well-mixed, north-south oriented estuary located in RI, USA. Fresh water and suspended sediment enters the system from the Providence-Seekonk and Taunton Rivers. Overall, NB displays cyclonic non-tidal surface flow. Gradients of sediment load, nutrients, primary productivity, and fresh water follow a north-south gradient of decreasing intensity [Oviatt, 2008]. NB also experiences typical estuarine flow whereby salty marine-sourced water enters the bay at depth, is diluted via fluvial and groundwater inputs, and is returned to the ocean at the surface. Narragansett Bay’s location, near the boundary of cool northern sourced Labrador Sea Water and warm southern sourced Gulf Stream water [Levitus, 1982], make it an ideal place to monitor both changes in North Atlantic climate and the effects of those changes [Borkman and Smayda, 2009]. In an effort to produce a well-constrained multi-proxy reconstruction of natural and human-induced change in the Narragansett Bay watershed I applied a diverse set of techniques including chemostratigraphy, stable isotope geochemistry, and organic geochemistry in concert with micro-paleontology, microbiology, and archaeology. Highlights include: The verification that the Uk’37 Index accurately reflects sea surface temperature in Narragansett Bay. In contrast to other estuarine settings [Mercer et al., 2005], characteristic alkenone ‘fingerprints’ lack the C37:4 isomer, suggesting that alkenones are likely produced by species closely related to open-ocean haptophyte algae. Sediments from three spatially- distributed core tops have Uk’37 inferred SSTs close to modern and comparison of our sediment core Uk’37 time series with a long (>100 years) regional instrumental SST record indicates Uk’37 in Narragansett Bay responds directly to historical increases in SST (p<.05). A near constant 1.3°C (±0.5°C) offset between two geographically separated cores over the past 300 years is consistent with modern in situ SST gradients and suggests the nominal ±1.4°C Uk’37 calibration error [Prahl et al., 1988] is pessimistic. Evidence suggests Uk’37 in Narragansett Bay reflects mean annual SST (with a small positive offset) and is reproducible to within ±0.3°C. 3 The characterization of water-column alkenone production over four annual cycles. Between 2009 and 2013, monthly to thrice-weekly filtrations of water column particulate matter for the analysis of alkenones revealed periods of both close agreement (≤ 1.5°C) and wide deviation (≥ 10°C) between Uk’37 inferred and instrumental SST. Uk’37 often underestimates actual SST during summer when alkenone concentrations are high and over estimates actual SST during winter when alkenone concentration is low. Based on comparison of sedimentary alkenone-to-aluminum ratios with water column ratios, we conclude warm Uk’37 offsets during winter likely result from the resupension of surface sediment during winter storms. The causes of cool Uk’37 offsets during summer are not as clear and include 1) changes in nutrient concentrations, 2) changes in alkenone producing flora, and 3) yet undescribed changes in physiological life stage, demand, or non-thermal environmental response. Despite this seasonal variability, the mass weighted average Uk’37 SST over the course of our study is 14.6°C, within 1.1°C of instrumental mean (12.7°C). The development of a 1500 year-long record of SST for Narragansett Bay. Based on the success of the both the 300 year-long Uk’37 SST reconstruction and the water column time series, we produced a Uk’37-inferred SST record, based on five cores from three locations, for Narragansett Bay spanning the past 1400 years with an average sample spacing of 15 years. The records contain SST structure consistent with the Medieval Climate Anomaly, the Little Ice Age, and high rate of modern warming (~1.1°C/100years). Decadal and centennial structure show evidence of a relationship with the North Atlantic and Arctic Oscillation Oscillations, respectively, in which relatively warm SSTs accompany NAO/AO positive periods, and vice versa. This positive relationship, and the timing of the Medieval Climate Anomaly and Little Ice Age in comparison to other North Atlantic records, highlights a north-south divide in eastern 4 North Atlantic SSTs that is persistent on millennial timescales. This work represents the first successful paleoclimate application of the Uk’37 SST proxy in an estuary. Reconstructing the effects of land-use change on sediment and nutrient delivery in Narragansett Bay. Using isotopes of bulk sedimentary nitrogen (δ15N), pollen of the invasive weed Ambrosia, magnetic susceptibility, and the abundance of benthic foraminifera we investigate the progression of landscape disturbance in the Bay’s watershed over the past 700 years. Our results suggest that modern north-south gradients in fresh water, sediment load, and nutrients have been a persistent, although changing, feature of the Bay over the past 700 years. The pre-Industrial balance between marine and terrestrial organic matter sources to the Bay was interrupted three times: first (1600s) by the addition of terrestrial nutrients after the extirpation of the native beaver population, second by the wide spread land clearance and rapid growth in the watershed’s domestic animal population by the Colonists (1700s) and, third by the massive ~30 fold increase in nutrient flux from the northern Bay after the Industrial Revolution. The nutrient N load added to the Bay concurrent with these events directly affected the Bay’s ecosystem leading to an increase in its benthic productivity. This work implies that humans began to more heavily influence the Bay’s ecosystem as soon as the early 1600s, but that the environmental degradation that occurred over the past 150 years is unprecedented in Narragansett Bay over at least that past 700 years. 5 Boothroyd, J. C. a. A., Peter, V. (2008), Geologic and Contemporary Landscapes of the Narragansett Bay Ecosystem, in Science for Ecosystem-Based Management, edited by A. D. a. B. A. Costa-Pierce, p. 570, Springer, New York, NY. Borkman, D. G., and T. Smayda (2009), Multidecadal (1959-1997) changes in Skeletonema abundance and seasonal bloom patterns in Narragansett Bay, Rhode Island, USA, J. Sea Res., 61(1-2), 84-94. Brassell, S. C., G. Eglinton, I. T. Marlowe, U. Pflaumann, and M. Sarnthein (1986), Molecular Stratigraphy - a New Tool for Climatic Assessment, Nature, 320(6058), 129-133. Conte, M. H., A. Thompson, and G. Eglinton (1994), Primary production of lipid biomarker compounds by Emiliania huxleyi: results from an experimental mesocosm study in fjords of southern Norway, Sarsia, 79, 319-332. Conte, M. H., M. A. Sicre, C. Ruhlemann, J. C. Weber, S. Schulte, D. Schulz-Bull, and T. Blanz (2006), Global temperature calibration of the alkenone unsaturation index (U-37(K ')) in surface waters and comparison with surface sediments, Geochemistry Geophysics Geosystems, 7, -. Gingerich, P. D. (2006), Environment and evolution through the Paleocene-Eocene thermal maximum, Trends in Ecology & Evolution, 21(5), 246-253. Herbert, T. D. (2003), Alkenone paleotemperature determinations, in Treatise in Marine Geochemistry, edited by H. Elderfield, pp. 391-432, Elsevier, Amsterdam. Imbrie, J., et al. (1992), ON THE STRUCTURE AND ORIGIN OF MAJOR GLACIATION CYCLES 1. LINEAR RESPONSES TO MILANKOVITCH FORCING, Paleoceanography, 7(6), 701-738. 6 IPCC (2007), Summary for Policymakers, in Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, edited by S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Levitus, S. (1982), Climatological Atlas of the World Ocean, 173 pp., U.S. Government Printing Office, Washington D.C. Mercer, J., M. Zhao, and S. Colman (2005), Seasonal variations of alkenones and Uk37 in the Chesapeake Bay water column, Estuarine, Coastal and Shelf Science, 63(4), 675-682. Müller, P. J., G. Kirst, G. Ruhland, I. von Storch, and A. Rosell-Melé (1998), Calibration of the alkenone paleotemperature index U37K′ based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S), Geochimica et Cosmochimica Acta, 62(10), 1757-1772. Oviatt, C. A. (2008), Impacts of Nutrients on Narragansett Bay Productivity: A Gradient Approach, in Science for Ecosystem-based Management, Narragansett Bay in the 21st Century, edited by A. Desbonnet, pp. 523-543, Springer, New York. Prahl, F. G., and S. G. Wakeham (1987), Calibration of Unsaturation Patterns in Long-Chain Ketone Compositions for Paleotemperature Assessment, Nature, 330(6146), 367-369. Prahl, F. G., L. A. Muehlhausen, and D. L. Zahnle (1988), Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions, Geochimica et Cosmochimica Acta, 52(9), 2303-2310. 7 Schouten, S., E. C. Hopmans, E. Schefuss, and J. S. S. Damste (2002), Distributional variations in marine crenarchaeotal membrane lipids: a new tool for reconstructing ancient sea water temperatures?, Earth and Planetary Science Letters, 204(1-2), 265-274. Schulz, H.-H., Schoener, A., and Emeis, K. C. (2000), Long-chain alkenone patterns in the Baltic Sea an ocean-freshwater transition, Geochimica et Cosmochimica Acta, 64, 469-477. Versteegh, G. J. M., R. Riegman, J. W. de Leeuw, and J. H. F. Jansen (2001), U(37)(K ')values for Isochrysis galbana as a function of culture temperature, light intensity and nutrient concentrations, Organic Geochemistry, 32(6), 785-794. Volkman, J. K., S. M. Barrett, S. I. Blackburn, and E. L. Sikes (1995), Alkenones in Gephyrocapsa-Oceanica - Implications for Studies of Paleoclimate, Geochimica et Cosmochimica Acta, 59(3), 513-520. Webb, T. (1986), IS VEGETATION IN EQUILIBRIUM WITH CLIMATE - HOW TO INTERPRET LATE-QUATERNARY POLLEN DATA, Vegetatio, 67(2), 75-91. Weijers, J. W. H., S. Schouten, O. C. Spaargaren, and J. S. S. Damste (2006), Occurrence and distribution of tetraether membrane lipids in soils: Implications for the use of the TEX86 proxy and the BIT index, Organic Geochemistry, 37(12), 1680-1693. Zachos, J., M. Pagani, L. Sloan, E. Thomas, and K. Billups (2001), Trends, rhythms, and aberrations in global climate 65 Ma to present, Science, 292(5517), 686-693. 8 CHAPTER ONE ____________________________________________________________ ALKENONE PRODUCTION WITH OPEN OCEAN CHARACTERISTICS IN A NEAR-SHORE SETTING: NARRAGANSETT BAY (RHODE ISLAND, U.S.A.) Salacup, J.M.1, Farmer, J.R.2, Herbert, T.D.1, and Prell, W.L.1 1 Brown University, Department of Geosciences 2 Columbia University, Lamont Doherty Earth Observatory 9 1.1 Abstract Our understanding of decadal to centennial-scale climate change during the Common Era (last 2ky) is heavily biased towards land-based records, primarily from tree rings. Given the the importance of ocean sea surface temperature (SST) in the global energy and hydrological cycles, reliable, high resolution, quantitative, records of past SST are needed to help balance this bias. If such proxies can be developed, the high sedimentation rates encountered in estuaries may afford a new opportunity to develop these valuable records. Here we demonstrate through a water-column and sediment-based approach that the Uk’37 alkenone SST proxy, widely applied for temperature reconstruction in the open ocean, may be applied successfully in the novel setting of Narragansett Bay (Rhode Island, U.S.A.). Characteristic Narragansett Bay alkenone profiles lack the C37:4 moiety, suggesting that alkenones are likely produced by species closely related to open-ocean haptophyte algae. SSTs inferred from converting water-column particulate Uk’37 to temperature using the well-validated open ocean relationship gives values very close to instrumental SST. Additionally, sediments from three spatially-distributed core tops have Uk’37 inferred SSTs close to modern. Comparison of our sediment core Uk’37 time series with a long (>100 years) regional instrumental SST record indicates Uk’37 in Narragansett Bay responds directly to historical increases in SST (p<.05). These results should prompt the investigation of Uk’37 in other near shore environments of normal marine salinity in an effort to develop a database of high resolution marine climate variability during the Common Era. 1.2 Introduction We cannot fully characterize anthropogenic global warming within the context of natural climate variability using instrumental temperature records alone. The compilation of instrumental temperature records did not begin until after the system was perturbed (~mid-1800s); hence it fails to separate baseline natural variability from anthropogenic. However, reliable paleotemperature reconstructions that span the last millennia can be used to quantify the 10 current anthropogenic perturbation within this broader framework [ex. Jansen, 2007; Mann et al., 2008; Moberg et al., 2005]. The twelve high resolution compilations of paleotemperature composited to produce the IPCCs most current understanding of global temperature for the last two thousand years (the Common Era) [Jansen, 2007] contains records from nearly two thousand trees. Unfortunately, this same compilation only incorporates ten records of sea surface temperature (SST) variability [Anderson, D M et al., 2002; De Leeuw et al., 1980; Dunbar et al., 1994; Gupta et al., 2003; Heiss, 1994; Keigwin, 1996; Lough, personal communication; Nyberg et al., 2002; Quinn et al., 1993]. Our understanding of global temperature over this period therefore largely depends on terrestrial data, and disproportionately on one proxy, tree rings. Given the importance of the ocean temperatures to both local and global climate, inclusion of more SST records into the IPCC compilations over the Common Era should be a primary target of future paleoclimate investigations. Estuaries provide an opportunity to reconstruct paleo-environments during the Common Era because they often contain thick sequences of Common Era sediment that have accumulated much faster than in the open ocean. Estuarine sedimentation rates on the order of mm/yr make sub-decadal sampling readily achievable. Additionally, estuaries are transition zones between marine and fresh waters, thus their sediments capture both terrestrial and marine responses to environmental change within the same samples. As such, estuarine sediments are important to understanding future local-to-regional level ecosystem responses in the face of global climate change. Unfortunately, traditional inorganic paleotemperature proxies, such as the oxygen isotope [e.g. Wanamaker et al., 2011] or magnesium-to-calcium ratios in biogenic carbonates [e.g. De Leeuw et al., 1980] often do not work well in estuarine settings, due to the lack of adequate sedimentary carbonate and/or the confounding effects of salinity [e.g. Dissard et al., 2010] . However, the development and application of organic biomarker SST proxies hold promise in these dynamic, carbonate-deprived, mud-rich settings. 11 1.2.1 Alkenones as a Proxy for Estuarine Sea Surface Temperature The most well-established and widely-applied of the open-ocean biomarker SST proxies is Uk’37 [Eq. 1) Uk’37 = (37:2) / (37:2+37:3); see Herbert, 2003 for a review]. This proxy is based on the ratio of two polyunsaturated long-chain alkyl ketones, called alkenones, produced by specific haptophyte algae [Conte et al., 1994; Volkman et al., 1995]. Initially, this ratio was correlated with sediment core δ18O ratios from planktonic foraminifera, known to partially record SST, and clearly identified glacial-interglacial temperature cycles [Brassell et al., 1986]. Culture [Prahl and Wakeham, 1987; Prahl et al., 1988] and core-top sediment [Müller et al., 1998] calibration studies led to the development of the Uk’37 Index as a quantitative SST proxy ((Eq. 2) Uk’37 = 0.034(SST) + 0.039 [Prahl et al., 1988], ±1.4°C from 0 to 28°C). Reconstructed Uk’37 temperatures correlate best with mean annual SST for a variety of climate and haptophyte production regimes in the global ocean [Conte et al., 2006]. Alkenones have been detected in marine sediment cores of early Eocene to modern age [Marlowe et al., 1990], and in exposed outcrops of uplifted marine sediment [Cleaveland and Herbert, 2009]. Uk’37 has been used to document paleo-SST changes on decadal [e.g. Sicre et al., 2008] to orbital [e.g. Herbert et al., 2010] timescales. In the open ocean, alkenones are primarily produced by the coccolithophores Emiliania huxleyi and Gephyrocapsa oceanica [Conte et al., 1994; Volkman et al., 1995]. Whether they are the primary alkenone producers in estuarine and near-shore settings is unclear. Instead, the genera Isochrysis and Chrysotila may dominate [Versteegh et al., 2001]. This is an important distinction because Isochrysis and Chrysotila have different Uk’37-SST calibrations than their marine counterparts [Versteegh et al., 2001; Volkman et al., 1989]. The alkenones produced by Isochrysis and Chrysotila may contain high contributions of the C37:4 alkenone not normally detected in open marine environments. For example, previous investigations in the low salinity Baltic Sea (salinity = 3-20psu) [Schulz, 2000] and Chesapeake Bay (salinity = 0-30) [Mercer et al., 2005; Schwab and Sachs, 2011] detected high concentrations of the C37:4 moiety and also 12 failed to identify a robust relationship between Uk’37 and SST. By extension, an alkenone ‘fingerprint’ with a high contribution of C37:4 may signal a brackish environment, hosting Isochrysis and/or Chrysotila, in which application of the Uk’37 SST proxy may not be appropriate. Thus, the absence of 37:4 is often used to determine the utility of Uk’37 in a given setting. In fact, a growing body of evidence suggests high C37:4 concentrations may be directly linked to low salinities and therefore may be useful instead as a salinity proxy [Bendle et al., 2005; Harada et al., 2003; Rosell-Mele, 1998; Sicre et al., 2002]. Another possible challenge encountered in estuaries is the seasonality of peak alkenone production. Nutrients, salinity, SST, primary production, and ecology are far more dynamic over the annual cycle in estuaries than in the open ocean. Primary production in estuaries often peaks in the spring; when snow and groundwater, and its resident nutrient load, melt and are released into the watershed. Another 'bloom' may occur in the fall, when the thermal stratification of the water column breaks down, mixing nutrients back to the surface. In many situations, nutrient dynamics may be further complicated by the release of waste water to the system. This can lead to spatial gradients in productivity [e.g. Oviatt, 2008] that when coupled with seasonal dynamics, may strongly affect the integrated Uk’37 ratio preserved in the sediments and later interpreted downcore as SST. Here, we seek to reconstruct Narragansett Bay SST (NB; Fig. 1.1) for the last three hundred years using Uk’37. Our goals include:1) verifying the alkenones are being produced in NB as opposed to being advected in from the open ocean, 2) attributing NB alkenone production to well-calibrated open-ocean producers like E. huxleyi and G. oceanica, 3) confirming that the open ocean calibration of Prahl [1988] is appropriate, 4) addressing seasonality with respect to Uk’37-inferred SSTs in Narragansett Bay sediment, 5) determining that down core Uk’37 ratios reflect mean annual SST via comparison with instrumental SST, and 6) showing that these results are reproducible. Our strategy includes a water column and sediment sampling scheme that spans the major gradients of salt, nutrients, and productivity in 13 NB in an effort to capture the breadth of possible alkenone producing communities, from the intermittently brackish Providence River to the nearly marine salinity of the lower Bay. 1.2.2 Narragansett Bay NB (Fig. 1.1) provides an appropriate setting to investigate the use of Uk’37 in an estuarine environment. Its mean volume-weighted salinity is high [29.96‰; Prell personal communication], suggesting the potential for alkenone production by well-calibrated open ocean species such as E. huxleyi and G. oceanica. However, the Providence River is seasonally diluted by runoff, providing a suitable habitat for brackish alkenone producers, such as Isochrysis and Chrysotila, if present. NB also has an array of water quality monitoring buoys (Table 1.1) that have been in place for over a decade [e.g. Codiga et al., 2009]. These monitoring data (Fig. 1.2), including well resolved near surface SST, salinity, and chlorophyll (Chl) measurements, provide important context for our alkenone results with detail not typically possible in the open ocean [Rhode Island Department of Environmental Management, 2012]. NB SST gradients grow during the spring and summer months (Fig. 1.2 top; March- August) as shallower Providence River and upper Bay sites warm more rapidly than mid- or lower Bay sites. This differential heating results in a nearly 5°C SST difference between these locations during the summer (July and August). The reverse occurs in the fall and winter. These records provide spatio-temporal context in which we will deduce the impacts of seasonality on the Uk’37 signal incorporated into NB sediments. Sea surface salinity (Fig. 1.2 middle) also varies along the Bay as the proximity of the major fresh water sources results in lower, more variable surface salinity in the upper Bay and Providence River (ex. CP) and higher and less variable surface salinity in the lower Bay. A bay- wide spring freshening results from the melting and runoff of winter snow and ice and low evapotranspiration, with the lowest surface salinity values occurring in the upper Bay and Providence River near the primary source of fresh water. Surface salinity peaks during the late 14 summer due to low precipitation and high rates of evapotranspiration. At this time, surface salinity varies from ~31psu in the lower Bay to ~27psu in the upper Bay and Providence River. Near surface Chl concentrations also fluctuate across space and time (Fig. 1.2 bottom). Spatially, higher Chl is detected in regions that are more heavily impacted by the delivery of nutrients [Oviatt, 2008] with Chl concentration decreasing with distance down Bay. Seasonal peaks in Chl define the spring (February-March) and summer / fall (May – September) blooms. The lowest concentrations of the year are detected in November. The water quality monitoring buoys only cover the past fifteen years (shorter in most cases) and a much longer record is required to test our sediment core alkenone SST estimates. Fortunately, T.F. Green airport in Providence, RI has a one hundred and fifteen year-long record of air temperature that can be used to develop a correlation between SST and local air temperature. We can then extrapolate this relationship to produce a long instrumentally derived estimate of SST (SSTAIR) for comparison with our Uk’37 reconstructions. 1.3 Methods 1.3.1 Water-Column Survey To identify water column alkenone concentrations and Uk’37 ratios, surface water samples for particulate organic matter (POM) analyses were obtained at eight locations (Fig. 1.1, numbered circles; Table 1.1) on monthly National Marine Fisheries Service (NMFS) cruises (Jun 2009-Nov 2010). One 20L carboy was filtered through a 1µm glass fiber filter (Pall Corp.) and then frozen for biomarker extraction. 1.3.2 Sediment Cores Three sediment cores (Fig. 1.1; numbered stars, Table 1.1), sampled at ≤ 2cm resolution, were used to establish an alkenone based paleo-SST reconstruction. Cores NB12 (1.75m), from Greenwich Bay (GB), and NB25 (1.25m), from Potter’s Cove (PC), were obtained in 2008 using piston push coring. Core NB44v (1.18m), from near Fox Island (FI), was obtained in 2010 using a vibra corer. All cores were taken in proximity of a water quality monitoring buoy. 15 Age control (Table 1.2) is based on unsupported 210Pb activities (usually helpful over the past 100 years), the characteristic increase in relative Pb concentrations ([Pb]) deposited in NB sediments circa 1850 related to industrialization of the watershed [Bricker, 1993; Corbin, 1989], Ambrosia pollen horizons diagnostic of landscape disturbance circa 1700 [Parshall et al., 2003], and 14C analyses of benthic foraminifera. Sedimentary [Pb] were acquired every ≤2 cm using an INNOV-X 4000 hand held XRF. A rapid increase in [Pb], usually expressed as a near-stepwise increase within ~10cm, reflects industrialization circa 1850 [Bricker, 1993; Corbin, 1989; this study]. The actual depth of onset was defined as the last sample depth before this increase. Samples for 210Pb analyses were submitted to Flett Research Ltd. for measurement and initial age models. Samples for pollen-based age control from NB12 and NB25 were cleaned using the standard pollen processing techniques of Faegri and Iversen [1989]. We selected sediment samples (~1 cm3) for analysis based on the preliminary [Pb]-based age-model. They were treated with KOH (to remove organics), 10% HCl (to detect, and then if present, remove carbonates), HF (to remove silicates), and lastly, acetolysis (to remove organics and stain pollen residue). The samples were then transferred to vials, and amended with silicon oil. Pollen age control was based on the well-established onset of Ambrosia at about 1700, associated with major land clearance [Hubeny et al., 2008; Parshall et al., 2003]. Because pollen was not analyzed on NB44v, we used radiocarbon analyses to provide pre-1850 age control. Radiocarbon analyses for NB44v were performed at the National Ocean Sciences Accelerator Mass Spectrometry facility (NOSAMS) on benthic foraminifera (Elphidium sp.) after cleaning in Nanopure water and low temperature drying (<50°C). Reported radiocarbon ages were corrected for the standard marine reservoir effect (~400 years) minus a local reservoir correction [~120 years; McNeely et al., 2006] and calibrated using OxCal. 16 1.3.3 Alkenone Analyses Alkenones were extracted from sediments and whole water column filters, and isolated from co-eluting compounds, before quantification of the Uk’37 ratio. Sediments (~1g) were freeze-dried, powdered, and extracted with dichloromethane on an Accelerated Solvent Extractor (Dionex, ASE200) at 150°C and 1500 psi to produce a total lipid extract (TLE). TLEs of filtered samples were saponified in 2N KOH (in 5% H2O in MeOH) and heated for 2.5 hours before addition of salt (5% in H2O) and HCl (6N to pH 2). This mixture was then extracted three times with hexane to remove the alkenone bearing fraction. This hexane, and the TLEs of down- core samples, was separated on silica gel using hexane, dichloromethane, and methanol to yield hydrocarbon, ketone (alkenone), and polar fractions, respectively. The preliminary investigation of alkenones in surface sediments highlighted an extremely complex matrix of co-eluting compounds that more typical derivitization, saponification and/or, silica gel clean-up procedures would not resolve. Therefore, samples deposited after ~1850 ([Pb] onset) were urea adducted to remove this complex matrix [Murphy, 1969]. Briefly, the dichloromethane (alkenone) fractions were dried gently under N2 then re-suspended in 1.5mL DCM:Hex (2:1). To this 1.5mL urea in MeOH (50mg/mL for NB12 and NB25, 100mg/mL for NB44v) was added. Solvents were removed, and urea crystals formed, by gentle drying under N2 on a hotplate set to ~30°C. Urea crystals were rinsed three times with hexane to remove the non-adducting material. The washed crystals were then dissolved in Nanopure water which was then extracted three times with hexane to remove the alkenone bearing fraction. In order to make sure this procedure did not affect Uk’37 ratios, we performed multiple applications of urea adduction on our laboratory Uk’37 standard and on replicate Narragansett Bay sediment samples. The average Uk’37 difference between adducted and non-adducted sediment samples was 0.046 ± 0.010 (0.20 ± 0.31°C, n=18, using the calibration of Prahl et al, [1988]). The average recovery of urea adduction was 94 ± 14% (n=18). We therefore conclude that urea adduction does not introduce a systematic bias to alkenone unsaturation estimates, and results 17 in uncertainties comparable to routine determinations of Uk’37 by GC-FID in samples with simpler organic matrices. Two quantification standards (n-C36 and n-C37 alkanes) were added to all samples before being injected from an autosampler into a 112°C CIS-PTV (cooled injection system- programmed temperature vaporizer) inlet operated in solvent vent mode. After the initial vent, the inlet was ramped at 12°C/min to 240°C, held isothermally for 5 minutes, ramped again at 12°C to 320°C, and held isothermally for 2 minutes before cryogenic cooling. A 60m, 0.32mm ID, 0.10um film DB-1 with a 5m fused guard column (DB-1 duraguard) was used. The oven temperature began at 90°C for 2 minutes, was ramped at 40°C/min to 255°C, at 1°C to 302°C, and at 10°C to 325°C where it was held isothermally for 20 minutes. Often, an additional ramp at 10°C/min to 340° and an isothermal hold for 10 minutes was required to remove high-boiling- point compounds from the column. Hydrogen was used as a carrier gas. Analytical accuracy, tracked via the injection of a laboratory Uk’37 sediment standard, was ±0.042 Uk’37 units (±0.1°C). Reproducibility of replicate sample injections averaged ±0.045 Uk’37 units (0.17°C, n=78). Reproducibility of replicate sample extractions averaged ±0.046 Uk’37 units (0.22°C; n = 19). Estimates of error were calculated using the calibration of Prahl et al., [1988]. Reproducibility of alkenone concentrations (C37total) was within 8 and 13% for replicate injections (n=78) and extractions (n=19), respectively. Individual alkenones were identified in samples suspected of containing the C37:4 alkenone via GC-MS and comparison with accepted mass spectra. 1.3.4 Development of a long instrumentally-derived SST record In order to compare down-core alkenone estimates of SST to the historical record, we used the longest available air surface temperature record to develop an estimate for NB temperatures over the last 115 years. On the monthly timescale, mean air and sea surface temperatures are tightly coupled in Narragansett Bay (Fig. 1.3A). A linear regression (Fig. 1.3B) between monthly average air temperature for Providence [NASA-GISS] and monthly average 18 SST from Conimicut Point from 2000 to 2008 [Rhode Island Department of Environmental Management, 2012] gives the equation: Eq.3) SSTAIR-CP = 0.867 x (air temp) + 3.176 (Fig. 1.3B; r2 = 0.94; n = 108) On the basis of this relationship, we use the 115 year-long record of Providence monthly air temperature to produce a monthly resolved SST reconstruction for Conimicut Point (SSTAIR- CP). Similarly, SSTs at Conimicut Point are closely coupled to SSTs at our core locations (Fig. 1.3C and D). Regression of monthly SST between Conimicut Point and Greenwich Bay for 2007 and 2008 gives the equation: Eq.4) SSTAIR-GB = 1.09 x SSTCP – 0.82 (Fig. 1.3C; r2 = 0.99; n = 24) Regression of monthly SST between Conimicut Point and Potter’s Cove for 2007 and 2008 gives: Eq.5) SSTAIR-PC = 0.96 x SSTCP + 0.27 (Fig. 1.3D; r2 = 0.96; n = 24) Regression of monthly SST between Conimicut Point and GSO (closest long instrumental SST record to FI) for 2007 and 2008 gives: Eq.6) SSTAIR-FI = 1.13 x SSTCP – 0.72 (Fig. 1.3E, r2 = 0.98; n = 22) Thus, we use Eqs. 4, 5, and 6 to transform the 115 year Conimicut Point SST estimate generated by Eq.3 into Greenwich Bay, Potter’s Cove, and Fox Island SST records, respectively (Fig. 1.3F). These records agree well, in both slope (rate of warming) and timing and amplitude of annual-to-decadal scale structure, with an instrumental SST record from Woods Hole [Shearman and Lentz, 2010] verifying the utility of our regression approach (Fig. 1.3F). We use these long locally-tuned records of SST as a comparison for our historical Uk’37-inferred SST reconstructions. 19 1.4 Results 1.4.1 Water-column Uk’37 C37:4-free alkenone fingerprints were detected in the surface waters of Narragansett Bay during 6 of the 9 sampled months, for a total of 13 detections (Fig. 1.4; Table 1.2). High C37:4 concentrations were detected in three samples; two in the Providence River (Sta.30; Fig. 1.1), and in smaller concentrations in one sample from the mid Bay (Sta. 39). These samples were taken in March and April of 2010, near the time of peak discharge of fresh water into the Bay (Fig. 1.4, Table 1.2). Maximum monthly average alkenone concentrations were detected in the upper Bay and Providence River (~55ng/L;C37:4-rich) and mid- and lower Bay (~7ng/L; C37:4- free) in April and July, respectively. The concentration weighted average Uk’37 during the sampling period (27 Jan, 2010 – 16 Nov, 2010), excluding samples with contributions of the C37:4 alkenone, was 0.495 (13.4°C using the Prahl et al [1988] calibration). The average instrumental SST at GSO over this same period was 13.0°C [Rhode Island Department of Environmental Management, 2012]. 1.4.2 Sediment Uk’37 Down core values of Uk’37 in all cores vary between 0.44 and 0.57 (Fig. 1.5). Rapid increases in Uk’37 are detected above 35 and 37cm for cores 12 and 25, respectively. No rapid increase is detected in core 44. The average core top Uk’37 is 0.502. There were no C37:4 alkenones detected in our sediment samples. All samples contained alkenones in concentrations between 0.06 and 0.59 µg/g (avg = 0.34µg/g). 1.4.3 Alkenone Indices Alkenone indices, determined by comparing the concentrations of different alkenone moieties, can be helpful in comparing alkenone synthesis in NB to that of open ocean species. For example, one index compares the total concentration of C37 and C38 alkenones. In the open ocean, this ratio is approximately 1:1 [Herbert, 2003]. A plot of C37total vs. C38total for all alkenones detected in NB sediment cores is shown in Figure 1.6 (A). The ratio in our cores is 20 0.7, a bit lower than the global open-ocean average. The C37total/C38total ratio of water-column samples containing C37:4 is ~1.5. The relationship between Uk’37 and a similar index built using the C38 methyl alkenones (Uk’38me) also varies directly and linearly with a slope very near 1 in the open ocean [Herbert, 2003]. Our sediment and water-column data fall along this pre- established line (Fig. 1.6B). 1.4.4 Age Control and Chemotratigraphy Sediment concentrations of [Pb], Ambrosia pollen, and unsupported 210Pb, as well as 14C age estimates, provide important age control (Fig. 1.7, Table 1.3). In NB25, [Pb] shows an abrupt increase at 43cm and Ambrosia pollen increases at ~63cm. In this core, unsupported 210 Pb activities provided 11 age estimates and age was modeled using a cubic spline (Fig. 1.8). In core 44v, [Pb] shows an abrupt increase at 31cm and unsupported 210Pb concentrations provided 3 age estimates. Radiocarbon dates of benthic foraminifera tests from NB44v at 51 and 111cm gave calibrated ages of 539 (± 25) and 1039 (± 25) years before present. Age was modeled in NB44v using linear interpolation of control points (Fig. 1.8). In NB12, [Pb] exceeds baseline values several times (between 61 and 31cm) before the irreversible increase in [Pb], diagnostic for industrial Pb addition to NB, at 38cm. Ambrosia pollen increases at ~59cm (between 56 and 62cm), although we note concentrations are variable below this. Unsupported 210Pb activities provided 7 age estimates. The increase then decrease of sedimentary [Pb] detected between 61 and 31cm is not consistent with previously published profiles of NB [Pb] [Bricker, 1993; Corbin, 1989]. Based on this, we believe this section represents an event of rapid (geologically-instantaneous) sediment deposition. If we thus collapse this section, the [Pb] and Ambrosia profiles look more like previously published profiles. Based on this adjustment, we define our [Pb] and Ambrosia horizons at 38 and 77 (between 74 and 80) cm, respectively (Fig. 1.7). Age for NB12 was modeled using a cubic spline (Fig. 1.8). Our age models assume that core top sediment reflects the year when the core was taken. This cannot be verified and should be considered a possible weakness in our age 21 models. Lastly, all cores are anchored to one another at 1850 via [Pb] horizons. Cores NB12 and NB25 are anchored at 1700 via pollen horizons. 1.5 Discussion 1.5.1 Evidence for in situ alkenone production Greenwich Bay Uk’37 is consistently higher than at Potter’s Cove (avg = 1.3°C using the calibration of Prahl et al. [1988]). This difference is consistent with a seasonal SST gradient between the two locations that is documented by instrumental buoy data (Fig. 1.2) suggesting alkenones are being produced in Narragansett Bay, as opposed to being produced in Rhode Island Sound and being advected in, as this would result in identical Uk’37 across the Bay. 1.5.2 Characterizing NB Alkenone Production The absence of the C37:4 alkenone in most water column samples and in the three sediment locations of NB is promising and stands in contrast to results from the only other study of Uk’37 in estuaries [Mercer et al., 2005]. Values of the C37total/C38total ratio near 1 (0.7) suggest production by open ocean producers in contrast with values from the low salinity producer I. galbana near 4 [3.7; Salacup, unpublished data]. We note that the slope of the C37total vs. C38total relationship (1.4) is slightly higher than that typical for the open ocean (~1). This is most likely due to the complex nature of these estuarine samples and the co-elution of unidentified compound peaks with our C38 alkenones. This results in artificially large C38total concentrations and a higher slope than for open ocean sediments, which typically have much simpler chromatograms This co-elution may also explain the extra scatter displayed in plot of Uk’37 vs. Uk’38me (Fig. 1.8). We think low salinities in the Providence River and mid-Bay during the spring of 2010 favored the growth of one or more “brackish” producers that do not produce alkenones with the same ‘fingerprint’ as the open ocean producers E. huxleyi and G. oceanica. Salinities in the Providence River on 9 March near 25psu decreased to as low as ~5psu by 8 April due to an historically wet March that resulted in the worst local flooding in over 100 years [Narragansett 22 Bay Project]. This lowered salinity may have allowed a senescent population of Isochrysis or Chrysotila to flourish. However, at no other time and in no other place did we detect C37:4 alkenones in the water column, or sediments, of NB. Instead, NB appears to be dominated by alkenone production that bears all the hallmarks of production by open ocean species like E. huxleyi and G. oceanica. This finding is consistent with preliminary work genetically identifying these species in NB [Salacup et al., in prep-c] and is likely a function of its high salinity. We therefore move forward by applying the open ocean calibration of Prahl et al. [1988] to our Uk’37 results. 1.5.3 Uk’37 SST offsets and production seasonality The effects of production seasonality on Uk’37 has been discussed extensively [Herbert, 2003 and references there in] and may be more important in shallow basins, such as Narragansett Bay, than in deep oceans due to the effects of settling depth on exposure time [ex. Broerse et al., 2000]. Longer exposure times can lead to signal attenuation (reduced seasonality) via grazing, and vertical mixing (both affecting signal integration) [ex. Broerse et al., 2000]. Thus, longer exposure times can lead to the ‘blurring’ of seasonally high surface alkenone fluxes by the time they reach the deep sea. Conversely, in the case of NB, shorter exposure times may lead to better transmission of alkenones synthesized during peak production to the sediments, and, potentially, an enhanced seasonal bias. Seasonal NB SST gradients may therefore help in the determination of alkenone production seasonality. For example, sedimentary alkenone-inferred SSTs in GB are positively offset from those in PC, suggesting that more of the Uk’37 signal in Narragansett Bay is produced during a part of the year when GB is warmer than PC (i.e. early Mar through late Oct; Fig. 1.9, dashed line). Comparison of Uk’37-inferred SST and SSTAIR (Fig. 1.10) indicates Uk’37 SST is 1.8°C warmer than SSTAIR at each core site. This may also be due to the seasonality of alkenone production. If we assume the Uk’37 SST recorded in the sediments reflects exactly that produced in the water column, we can use the monthly resolved TFG air temperature record, Eqs. 3, 4, 5, 23 and 6, and the 1.8°C offset to hypothesize during what months the SST records were produced. For the period between 1895 and 2009, the average Uk’37 SST in GB and PC is 14.6 and 13.6°C, respectively. Conversion of the air temperatures to SSTAIR (see Methods) for the second week in Aug through the end of Nov from 1895 to 2009 yields instrumentally-inferred average SSTs of 14.7 and 14.0°C for Greenwich Bay and Potter’s Cove, respectively. These results are close to the Uk’37-inferred means and are consistent with our results based on in situ SST gradients (Fig. 1.9) also suggesting alkenone production in NB occurs, at least in part, during late summer and fall. However, we note that multiple non-unique combinations of months could yield the observed 1.8°C offset and therefore conclude that any bias is best explained by contributions from late-spring/summer and/or late-summer/fall production, both of which are very much like mean annual. Thus, we move forward interpreting Uk’37-inferred SSTs as a mean annual signal with a small (+1.8°C) absolute offset. 1.5.4 Response of Uk’37 to SST Sediment Uk’37-inferred SSTs (Fig. 1.10) appear coherent between cores. SSTs cooled from ~1700 to ~1850 with maximum cooling between 1820 and 1882 (depending on core). Temperatures begin to warm again between 1863 and 1931. The missing SST structure between 1820 and 1930 in core NB25 may be due to an undetected hiatus. We note the cooling seen in the very top of NB12 and 44v could be due to incomplete capture or compression of the most modern sediments, local effects due to sediment dredging, or shortcomings of the alkenone proxy. Nonetheless, if we interpolate the records on a common timescale with 8 year sample spacing (the average through all 3 cores) we detect a significant positive correlation between core NB12 and NB25 (r2 = 0.72, p < .0001). The relationships between cores NB12 and NB44v and NB25 and NB44v were not significant at the 95% confidence interval. To further validate the use of Uk’37 as a paleothermometer in the sediment record of NB, we compare the Uk’37 inferred SSTs over the past century with SSTAIR at our core sites (Fig. 1.7). Comparison of sediment core Uk’37 SST with SSTAIR on the same common age axis as 24 described above (8 year sample spacing) also produces significant positive correlations for NB12 (r2 = 0.27, p <.05) and NB25 (r2 = .58, p <.002). There was no relationship between NB44 and SSTAIR. These estimates may be pessimistic as they ignore error in our age control (likely the largest source of error in this estimate). Regional SSTs have warmed during the last ~100 years [Nixon, S X et al., 2004; Shearman and Lentz, 2010]. The Uk’37 inferred rate of warming for the 100 year period between 1905 and 2005 is 1.5°C/100 yr (±0.4°C) for Greenwich Bay. The rate at Potter’s Cove for this same period is 2.0°C/100yr (±0.3°C). These rates are within error of the observed rates of Providence air temperature and by extension SSTAIR at each site (1.7°C/100yr ±0.2°C). The rate of warming for NB44v is -0.05°C/100 yr (±0.1°C) and is not consistent with instrumental records over this period. The rate of 20th century warming implied by the NB12 and 25 reconstructions (avg. = 1.8°C) is more than three times that of the proxy-based Northern Hemisphere mean anomaly (based primarily on tree-rings) reported by Mann et al. [2008; ~0.5°C], and more than twice the instrumental global mean [0.74°C; IPCC, 2007]. Our NB estimates are closer to, but still exceed, rates of warming implied by a 132-year long regional instrumental record for the Gulf of Maine (1.0°C/100 yr) and the Mid-Atlantic Bight (0.7°C/100 yr) [Shearman and Lentz, 2010]. However, the rate of warming implied by our records is similar to the rate obtained from instrumental SST records taken at Woods Hole, MA [1.6°C/100yr from 1900-2000, Nixon, S X et al., 2004; 1.2°C/100yr from 1905-2005, Shearman and Lentz, 2010] confirming a regional pattern of SST change that is warming faster than the global mean, most likely due to the polar amplification of global warming [Shearman and Lentz, 2010]. These findings highlight the importance of regional, long, high-resolution, proxy-based SST reconstructions to understanding the heterogeneity of global ocean climate dynamics on decadal to centennial time scales. 25 1.5.5 Estimation of Uncertainty and Reproducibility As mentioned earlier, the nominal calibration error associated with Uk’37 is ±1.4°C [Prahl et al., 1988], nearly the same amplitude of reconstructed temperature variability over the Common Era (signal-to-noise ~1) [Jansen, 2007]. This calibration error assumes core top sediments used in the calibration have zero age and that the ocean atlases used accurately reflect instrumental mean annual SST. This implicitly assumes that core-top Uk’37 ratios everywhere actually reflect mean annual SST, which is not always the case [Prahl et al., 2010; this work]. Therefore, the nominal error (±1.4°C) is a worst case scenario estimate in globally distributed surface sediments, of uncertain age, underlying waters hosting a wide range of alkenone producing populations, and thus may be unrealistic in a geologic context when many of the above assumptions are avoided. To empirically determine the actual precision and reproducibility of Uk’37 in NB, we compared the Uk’37 SST variability from core NB25 and NB12. Each record was again linearly interpolated on a common age axis with 8 year sample spacing. They were normalized to each other by subtracting their means. The root mean squared error between the normalized records, a better statistical indicator of our relative reconstructed Uk’37 error (or noise), is ±0.3°C. This error is pessimistic, as it includes not only error associated error in the Uk’37 measurement but also error associated with our age models (the major source of error in the above analysis). The resulting minimum S/N, calculated by dividing the range in SST for each record by the noise is 9.5 and 8.8 for Greenwich Bay and Potter’s Cove, respectively. Furthermore, the near constant 1.3°C (±0.5°C) offset between Uk’37 SST in NB12 and NB25 over the past 300 years in inconsistent with a random error of ±1.4°C. So, although the absolute value of our SST reconstructions may only be correct to within ± 1.4°C, we have confidence in our ability to quantify SST variability quite accurately. 26 1.6 Conclusions Alkenones were detected in all Narragansett Bay sediment samples analyzed. Offsets of Uk’37 between the individual NB cores site are consistent with in situ SST gradients indicating the alkenones are being produced in NB, with enhanced production from late spring through fall. Water column and sediment-based evidence suggests NB alkenone production is dominated by species similar to the open ocean species E. huxleyi and G.oceanica. The agreement of our sediment core Uk’37 SSTs, calculated using the open-ocean calibration of Prahl [1988], with SSTAIR provides strong evidence that alkenone synthesis responds directly to water temperature, and is therefore a valid recorder of mean annual SST in NB, with a small positive offset. Significant decadal coherence of alkenone-reconstructed SST between cores suggests that in situ gradients of salinity, nutrients, and productivity do not compromise the interpretation of Uk’37 in NB. Instead, an in situ gradient in SST captured by our reconstructions implies Uk’37 in NB is a sensitive indicator of local conditions. The rate of 20th century warming implied by our records is greater than those suggested by more ‘global’ approaches and instead shows important regional evidence of the distal effects of the polar amplification of global warming. Despite the 1.4°C nominal calibration error of the Uk’37 proxy, acquisition of multiple sediment cores has the ability to drastically improve this pessimistic estimate of error, in our case SST variability was reproducible to within ±0.3°C If the occurrence of open-ocean alkenone production in near shore settings similar to NB can be confirmed in future work, estuarine alkenone paleothermometry will take its place with other high-resolution archives such as tree rings, ice cores, corals, and varved basins. In comparison with the atmosphere, the ocean has slow modes of variability with broad spatial scale [Deser et al., 2010] meaning fewer records will be needed to provide a robust marine complement to terrestrial paleotemperature records. Near shore Uk’37 may also provide a new opportunity to reconstruct high resolution temperatures in regions where tree rings do not work due to lack of annual banding or the confounding effects of precipitation, such as in southern New England 27 [Cook, E.R. and Mayes, 1985]. The high temporal resolution and expanded geographic coverage afforded by this new capability should lead to a better understanding of the spatial and temporal dynamics of internal modes of climate variability (ex. El Nino Southern Oscillation, North Atlantic Oscillation, changes in Atlantic Merdional Overturning Circulation) and the evolution of prolonged climate states (ex. Medieval Climate Anomaly, Little Ice Age). 28 1.7 Tables Table 1.1 Identification, location, and description of water quality monitoring buoys, NMFS cruise sampling stations, and cores used in this work ID Latitude Longitude Water Depth Reference (abrev.) 41 N (deg.) 71 E (deg.) (m) Instrumental Data Conimicut Point (CP) 0.724 0.340 " Rhode Island Dept. of Greenwich Bay (GB) 0.673 0.425 " Environmental Potter's Cove (PC) 0.643 0.340 " Management URI's Graduate School of Oceanography 0.450 0.430 " (GSO) Providence Air Temperature 0.727 0.437 NASA-GISS (TF Green Intl' Airport) Cruise Sampling Locations Ft. Wetherill (11) 0.477 0.351 surface Gould Island (13C) 0.545 0.334 " Hog Island (14) 0.619 0.287 " Fall River (18B) 0.704 0.183 " Field's Point (30) 0.780 0.380 " West North Prudence (39) 0.636 0.377 " Fox Island (1A) 0.548 0.403 " Dutch Island (4) 0.480 0.408 " Cores Length (m) Greenwich Bay (12) 0.673 0.425 4.5 1.76 Fox Island (44v) 0.593 0.389 7.9 1.18 Potter's Cove (25) 0.643 0.340 3.8 1.25 29 Table 1.2 Dates, concentrations (ng/L), and locations of all water-column alkenone detections West Passage East Passage North South North South 30 39 1A 4 18B 14 13C 11 1/27/2010 n.d. n.d n.d n.d n.d n.d n.d 2.1 2/9/2010 n.d. n.d n.d n.d n.d n.d n.d n.d 3/9/2010 6.8 n.d n.d n.d n.d n.d n.d 1.5 4/8/2010 55.2 5.5 n.d n.d n.d n.d n.d n.d 6/16/2010 n.d. n.d n.d n.d n.d n.d n.d n.d 7/30/2010 n.d. n.d 2.2 5.5 n.d n.d 3.9 15.3 8/27/2010 n.d. n.d n.d 5.2 n.d 2.4 1.7 3.3 10/14/2010 n.d. n.d n.d 2.4 n.d n.d n.d 2.7 11/16/2010 n.d. n.d n.d n.d n.d n.d n.d 2.0 30 Table 1.3. Age-depth constraints used for NB cores Depth (cm) Age (cal. bp) Year (AD) Method NB 12 0 -58 2008 core top 3 -52 2002 Pb-210 8 -41 1991 Pb-210 13 -27 1977 Pb-210 18 -1 1951 Pb-210 23 21 1929 Pb-210 28 33 1883 Pb-210 35.5 75 1850 Pb-210 35 75 1850 Industrial [Pb] 55 167 (±19) 1783 14 C 70 180 (±41) 1770 14 C 77 250 1700 pollen NB25 0 -58 2008 core top 2 -53 2003 Pb-210 5 -46 1996 Pb-210 8 -38 1988 Pb-210 11 -30 1980 Pb-210 14 -23 1973 Pb-210 17 -17 1967 Pb-210 22 -13 1963 Pb-210 27 -9 1959 Pb-210 30 -6 1956 Pb-210 33 9 1941 Pb-210 36 14 1936 Pb-210 43 75 1850 Industrial [Pb] 63 250 1700 pollen NB44v 0 5 1945 core-top 2 13 1937 Pb-210 4 26 1924 Pb-210 7 44 1906 Pb-210 31 75 1875 Industrial [Pb] 14 51 539 (±25) 1411 C 14 111 1039 (±25) 911 C 31 1.8 Figures Figure 1.1. Map of Narragansett Bay and points of interest (see Table 1.1): core locations (stars; GB = Greenwich Bay (NB12); PC = Potter’s Cove (NB25); FI = Fox Island (NB44v)), surface water POM sampling locations (numbered circles). GSO = Graduate School of Oceanography, TFG = TF Green International Airport, CP = Conimicut Point buoy station. Arrows represent major sources of fresh water. Stippled lines denote the ecofunctional boundaries between sections of the Bay [Costa-Pierce and Desbonnet, 2008]. 32 Figure 1.2. Plots of the annual cycle of near surface (A) SST, (B) SSS, and (C) Chl for Conimicut Point (CP, green), Greenwich Bay (GB, red), Potter’s Cove (PC, blue), and the Graduate School of Oceanography (GSO, black, see Fig. 1.1) provide important context for our water column and sedimentary Uk’37 results. A number of year’s data (CP = 2009 - 2012; GB = 2007 - 2010; PC = 1996 – 2010; GSO = 1996 - 2010) was averaged then treated with a low pass filter to produce a monthly smooth. The patterns depicted are resistant to the length of time averaged and/or the removal of any one year’s data. 33 Figure 1.3. Development of a 115 year-long inferred NB SST record. (A) Providence monthly average air temperature at TF Green Airport (solid line; Fig. 1.1, TFG; Table 1.1) and Conimicut Point monthly average SST (dashed line; Fig. 1.1, CP; Table 1.1) from 1999 to 2008. (B) Regression between monthly average TFG air temperature and CP SST (n=108) used to construct a 115-year-long inferred SST record for Conimicut Point (SSTAIR-CP). (C, D, and E) Regressions between CP and GB, PC, and GSO (FI), respectively. Buoy data are from 2007 and 2008 (Table 1.1; n = 24). These regressions were used to transform SSTAIR-CP into 115- year-long SST records for each core site (SSTAIR-GB, PC, and FI). These results compare well with an instrumental SST record from Woods Hole, MA (Shearman and Lentz, 2010) confirming the utility of our regression approach. 34 Month 2010 Figure 1.4. Plot of POM-inferred surface water alkenone concentrations during the 2010 NMFS cruises (Fig. 1.2; Table 1.1). Upper Bay / Providence River concentrations are hash-marked, mid- and lower Bay concentrations are in white. Columns are an average of all alkenone detections in a given month. Error bars represent the maximum and minimum concentrations for that month. All data is provided in Table 1.3. The results show one bloom in the upper Bay and Providence River in the spring (Mar and Apr) and another in the lower Bay in the late summer / fall (Jul – Nov). A representative chromatogram showing a high contribution of the C37:4 moiety for the spring bloom suggests a brackish alkenone producing population inhabits the upper Bay at this time. The chromatogram for the late summer / fall bloom lacks C37:4, consistent with a marine producer. The C38 alkenone fingerprints also differ between the two blooms further suggesting different producers. 35 Depth (cm) Depth (cm) Figure 1.5. Uk’37 depth profiles for each core showing age control points for 210Pb (solid bar), [Pb] (open squares), Ambrosia pollen (open triangles), and 14C (filled squares). All cores are tied to one another at 1850 via the [Pb] increase. Cores NB12 and NB25 are tied to each other at 1700 via increases in sedimentary Ambrosia. 36 Figure 1.6. Plots of alkenone indices used to classify NB alkenone fingerprints as ‘marine-like’. (A) plot of C37total vs. C38total for all sediment core data showing a strong positive relationship with a slope near 1 (Herbert, 2003). Our slope is higher than 1 for reasons discussed in the text. (B) A plot of Uk’37 vs. Uk’38me, another tracer of SST, for sediment and water column samples from this study compared with those of Herbert (2003) show good agreement of NB alkenones with the open-ocean trend. 37 Depth (cm) Depth (cm) Figure 1.7. Key stratigraphic and age control results used to develop age models for Narragansett Bay cores. Plots of sedimentary profiles of [Pb] (squares), Ambrosia pollen (trinagles), 210Pb age estimates (open circles), and 14C ages (filled squares) for cores (A) NB12, (B) NB25, and (C) NB44v. Filled circles represent modern aged core top assumptions. Detection and removal of an instantaneous depositional event in core NB12 ((A), dashed lines, see text) resulted in the profiles depicted in (D). Arrows represent important age control points discussed in the text. 38 Figure 1.8. Age models for (A) GB, (B) PC, and (C) FI based 210Pb excess (circles), [Pb] increases (open squares), Ambrosia pollen (triangles), and 14C (filled squares). Age-depth relationships were calculated using cubic splines (NB12 and NB25) or linear fits (NB44v) and assume core top reflects modern sedimentation (filled circles). Average sedimentation rates on the order of millimeters per year make sub-decadal sampling readily achievable. 39 Figure 1.9. Plot of the difference in instrumental SST between Greenwich Bay (GB) and Potter’s Cove (PC) defined as GB SST – PC SST. The Uk’37 offset between NB12 and NB25 suggests that the alkenones are produced during a part of the year when GB is warmer than PC (dashed lines). Vertical green bars represent the periods of alkenone production inferred by our surface water samples (Table 1.2, Fig. 1.4). Darker color represents generally higher alkenone concentrations. 40 Figure 1.10. Comparison of Uk’37 reconstructed SSTs at SSTAIR at each site using the calibration of Prahl et al. [1988]. The thick black line is a stack of all three records. Significant co-variance between Uk’37 and SSTAIR at two of our sites (see text) provides strong evidence that Uk’37 is responding directly to growth water temperature. Significant intra-core variability (see text) suggests in situ gradients of salinity, nutrients, and productivity do not compromise the interpretation of Uk’37 in NB. Instead, an in situ gradient in SST captured by our reconstructions implies Uk’37 in NB is a sensitive indicator of local conditions on decadal time scales. Between 1905 and 2005, the rates of warming reflected in both the Uk’37 (1.8°C/100yr ±0.4°C) and instrumentally inferred (1.7°C/100yr ±0.2°C) SST rcords are within error. 1.9 References Anderson, D. M., J. T. Overpeck, and A. K. Gupta (2002), Increase in the Asian southwest monsoon during the past four centuries, Science, 297(5581), 596-599. Bendle, J., A. Rosell-Mele, and P. Ziveri (2005), Variability of unusual distributions of alkenones in the surface waters of the Nordic seas, Paleoceanography, 20(2). 41 Brassell, S. C., G. Eglinton, I. T. Marlowe, U. Pflaumann, and M. Sarnthein (1986), Molecular Stratigraphy - a New Tool for Climatic Assessment, Nature, 320(6058), 129-133. Bricker, S. B. (1993), The History of Cu, Pb, and Zn Inputs to Narragansett Bay, Rhode-Island as Recorded by Salt-Marsh Sediments, Estuaries, 16(3B), 589-607. Broerse, A., P. Ziveri, and S. Honjo (2000), Coccolithophore (CaCO3) flux in the Sea of Okhotsk: seasonality, settling and alteration processes, Mar Micropaleontol, 39, 179-200. Cleaveland, L. C., and T. D. Herbert (2009), Preservation of the alkenone paleotemperature proxy in uplifted marine sequences: A test from the Vrica outcrop, Crotone, Italy, Geology, 37(2), 179-182. Codiga, D. L., H. E. Stoffel, C. F. Deacutis, S. Kiernan, and C. A. Oviatt (2009), Narragansett Bay Hypoxic Event Characteristics Based on Fixed-Site Monitoring Network Time Series: Intermittency, Geographic Distribution, Spatial Synchronicity, and Interannual Variablity, Estuaries and Coasts, 32, 621-641. Conte, M. H., A. Thompson, and G. Eglinton (1994), Primary production of lipid biomarker compounds by Emiliania huxleyi: results from an experimental mesocosm study in fjords of southern Norway, Sarsia, 79, 319-332. Conte, M. H., M. A. Sicre, C. Ruhlemann, J. C. Weber, S. Schulte, D. Schulz-Bull, and T. Blanz (2006), Global temperature calibration of the alkenone unsaturation index (U-37(K ')) in surface waters and comparison with surface sediments, Geochemistry Geophysics Geosystems, 7, -. Cook, E. R., and P. Mayes (1985), Decadal-scale patters of climatic change over eastern North America inferred from tree-rings, in Abrupt Climatic Change: Evidence and Implications, edited by W. H. Berger and L. D. Labeyrie, pp. 61-66, NATO ASI Series. Corbin, J. (1989), Recent and historical accumulation of trace metal contaminants in Narragansett Bay sediments, 295 pp, University of Rhode Island, Kingston, RI. Costa-Pierce, B. A., and A. Desbonnet (2008), An "Ecofunctional" Approach to Ecosystem- based Management for Narragansett Bay, in Science for Ecosystem-based Management, edited by A. Desbonnet and B. A. Costa-Pierce, pp. 545-562, Springer, New York, NY. De Leeuw, J. W., F. W. V. D. Meer, W. I. C. Rijpstra, and P. A. Schnencck (1980), On the occurrence and structural identification of long chain unsaturated ketones and hydrocarbons in sediments, in Advances in Organic Geochemistry, 1979, edited by E. D. Douglas and J. R. Maxwell, pp. 211-217, Pergamon, Oxford. Deser, C., M. A. Alexander, S. P. Xie, and A. S. Phillips (2010), Sea Surface Temperature Variability: Patterns and Mechanisms, in Annual Review of Marine Science, edited, pp. 115-143. Dissard, D., G. Nehrke, G. J. Reichart, and J. Bijma (2010), The impact of salinity on the Mg/Ca and Sr/Ca ratio in the benthic foraminifera Ammonia tepida: Results from culture experiments, Geochimica et Cosmochimica Acta, 74(3), 928-940. 42 Dunbar, R. B., G. M. Wellington, M. W. Colgan, and P. W. Glynn (1994), EASTERN PACIFIC SEA-SURFACE TEMPERATURE SINCE 1600-AD - THE DELTA-O-18 RECORD OF CLIMATE VARIABILITY IN GALAPAGOS CORALS, Paleoceanography, 9(2), 291-315. Faegri, K., and J. Iverson (Eds.) (1989), Textbook of Pollen Analysis, Fourth Edition ed., 328 pp., John Wiley and Sons, Chichester. Gupta, A. K., D. M. Anderson, and J. T. Overpeck (2003), Abrupt changes in the Asian southwest monsoon during the Holocene and their links to the North Atlantic Ocean, Nature, 421(6921), 354-357. Harada, N., K. H. Shin, A. Murata, M. Uchida, and T. Nakatani (2003), Characteristics of alkenones synthesized by a bloom of Emiliania huxleyi in the Bering Sea, Geochimica et Cosmochimica Acta, 67(8), 1507-1519. Heiss, G. A. (1994), Coral reefs in the Red Sea: Growth, production and stable isotopes, GEOMAR Rep., 1-143 pp. Herbert, T. D. (2003), Alkenone paleotemperature determinations, in Treatise in Marine Geochemistry, edited by H. Elderfield, pp. 391-432, Elsevier, Amsterdam. Herbert, T. D., L. C. Peterson, K. T. Lawrence, and Z. H. Liu (2010), Tropical Ocean Temperatures Over the Past 3.5 Million Years, Science, 328(5985), 1530-1534. Hubeny, J. B., J. W. King, and M. Cantwell (2008), Anthropogenic influences on estuarine sedimentation and ecology: examples from the varved sediments of the Pettaquamscutt River Estuary, Rhode Island, Journal of Paleolimnology, 41(2), 297-314. IPCC (2007), Summary for Policymakers, in Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, edited by S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Jansen, E., J. Overpeck, K.R. Briffa, J.-C. Duplessy, F. Joos, V. Masson-Delmotte, D. Olago, B. Otto-Bliesner, W.R. Peltier, S. Rahmstorf, R. Ramesh, D. Raynaud, D. Rind, O. Solomina, R. Villalba and D. Zhang (2007), Palaeoclimate, in Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, edited by S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Keigwin, L. D. (1996), The Little Ice Age and Medieval warm period in the Sargasso Sea, Science, 274(5292), 1504-1508. Lough, J. M. (personal communication), edited by M. Mann. Mann, M. E., Z. H. Zhang, M. K. Hughes, R. S. Bradley, S. K. Miller, S. Rutherford, and F. B. Ni (2008), Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia, Proc Natl Acad Sci U S A, 105(36), 13252-13257. 43 Marlowe, I. T., S. C. Brassell, G. Eglinton, and J. C. Green (1990), LONG-CHAIN ALKENONES AND ALKYL ALKENOATES AND THE FOSSIL COCCOLITH RECORD OF MARINE- SEDIMENTS, Chem Geol, 88(3-4), 349-375. McNeely, R., A. S. Dyke, and J. R. Southon (2006), Canadian marine reservoir ages, preliminary data assessment, Open File 5049, pp. 3. Mercer, J., M. Zhao, and S. Colman (2005), Seasonal variations of alkenones and Uk37 in the Chesapeake Bay water column, Estuarine, Coastal and Shelf Science, 63(4), 675-682. Moberg, A., D. M. Sonechkin, K. Holmgren, N. M. Datsenko, and W. Karlen (2005), Highly variable Northern Hemisphere temperatures reconstructed from low- and high-resolution proxy data, Nature, 433(7026), 613-617. Müller, P. J., G. Kirst, G. Ruhland, I. von Storch, and A. Rosell-Melé (1998), Calibration of the alkenone paleotemperature index U37K′ based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S), Geochimica et Cosmochimica Acta, 62(10), 1757-1772. Murphy, M. T. J. (1969), Analytical Methods, in Organic Geochemistry, edited by G. Eglinton and M. T. J. Murphy, pp. 74-88, Springer, Berlin. Narragansett Bay Project Nu-Shuttle Transect Data, edited. NASA-GISS Surface Temperature Analysis, edited. Nixon, S. X., S. Granger, B. A. Buckley, M. Lamont, and B. Rowell (2004), A one hundred and seventeen year coastal water temperature record from Woods Hole, Massachusetts, Estuaries, 27(3), 397-404. Nyberg, J., B. A. Malmgren, A. Kuijpers, and A. Winter (2002), A centennial-scale variability of tropical North Atlantic surface hydrography during the late Holocene, Palaeogeogr Palaeocl, 183(1-2), 25-41. Oviatt, C. A. (2008), Impacts of Nutrients on Narragansett Bay Productivity: A Gradient Approach, in Science for Ecosystem-based Management, Narragansett Bay in the 21st Century, edited by A. Desbonnet, pp. 523-543, Springer, New York. Parshall, T., D. R. Foster, E. Faison, D. MacDonald, and B. C. S. Hansen (2003), Long-term history of vegetation and fire in pitch pine-oak forests on Cape Cod, Massachusetts, Ecology, 84(3), 736-748. Prahl, F. G., and S. G. Wakeham (1987), Calibration of Unsaturation Patterns in Long-Chain Ketone Compositions for Paleotemperature Assessment, Nature, 330(6146), 367-369. Prahl, F. G., L. A. Muehlhausen, and D. L. Zahnle (1988), Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions, Geochimica et Cosmochimica Acta, 52(9), 2303-2310. Prahl, F. G., J. F. Rontani, N. Zabeti, S. E. Walinsky, and M. A. Sparrow (2010), Systematic pattern in U37K′ – Temperature residuals for surface sediments from high latitude and other oceanographic settings, Geochimica et Cosmochimica Acta, 74(1), 131-143. 44 Quinn, T. M., F. W. Taylor, and T. J. Crowley (1993), A 173 YEAR STABLE-ISOTOPE RECORD FROM A TROPICAL SOUTH-PACIFIC CORAL, Quaternary Science Reviews, 12(6), 407-418. Rhode Island Department of Environmental Management (2012), Narragansett Bay Fixed-Site Monitoring Network (NBFSMN), edited, p. Data available at www.dem.ri.gov/bart, Office of Water Resources. Rosell-Mele, A. (1998), Interhemispheric appraisal of the value of alkenone indices as temperature and salinity proxies in high-latitude locations, Paleoceanography, 13(6), 694-703. Salacup, J., S. Theroux, T. D. Herbert, and W. L. Prell (in prep), A four-year long time-series of water column alkenones and 18S rRNA suggests Uk’37 is a viable proxy in Narragansett Bay, RI. Schulz, H.-H., Schoener, A., and Emeis, K. C. (2000), Long-chain alkenone patterns in the Baltic Sea an ocean-freshwater transition, Geochimica et Cosmochimica Acta, 64, 469-477. Schwab, V. F., and J. P. Sachs (2011), Hydrogen isotopes in individual alkenones from the Chesapeake Bay estuary, Geochimica et Cosmochimica Acta, 75(23), 7552-7565. Shearman, R. K., and S. J. Lentz (2010), Long-Term Sea Surface Temperature Variability along the U.S. East Coast, Journal of Physical Oceanography, 40(5), 1004-1017. Sicre, M. A., E. Bard, U. Ezat, and F. Rostek (2002), Alkenone distributions in the North Atlantic and Nordic sea surface waters, Geochem. Geophys. Geosyst., 3. Sicre, M. A., J. Jacob, U. Ezat, S. Rousse, C. Kissel, P. Yiou, J. Eiriksson, K. L. Knudsen, E. Jansen, and J. L. Turon (2008), Decadal variability of sea surface temperatures off North Iceland over the last 2000 years, Earth and Planetary Science Letters, 268(1-2), 137-142. Versteegh, G. J. M., R. Riegman, J. W. de Leeuw, and J. H. F. Jansen (2001), U(37)(K ')values for Isochrysis galbana as a function of culture temperature, light intensity and nutrient concentrations, Organic Geochemistry, 32(6), 785-794. Volkman, J. K., S. M. Barrett, S. I. Blackburn, and E. L. Sikes (1995), Alkenones in Gephyrocapsa-Oceanica - Implications for Studies of Paleoclimate, Geochimica et Cosmochimica Acta, 59(3), 513-520. Volkman, J. K., S. W. Jeffrey, P. D. Nichols, G. I. Rogers, and C. D. Garland (1989), Fatty-Acid and Lipid-Composition of 10 Species of Microalgae Used in Mariculture, J Exp Mar Biol Ecol, 128(3), 219-240. Wanamaker, A. D., K. J. Kreutz, B. R. Schone, and D. S. Introne (2011), Gulf of Maine shells reveal changes in seawater temperature seasonality during the Medieval Climate Anomaly and the Little Ice Age, Palaeogeogr Palaeocl, 302(1-2), 43-51. 45 46 CHAPTER TWO _______________________________________________________ A FOUR YEAR LONG RECORD OF WATER COLUMN ALKENONE RATIOS AND CONCENTRATIONS FROM NARRAGANSETT BAY: SALINITY SEGREGATION, SEASONAL VARIABILITY, AND MEAN ANNUAL SEA SURFACE TEMPERATURE Salacup, J.M.1, Herbert, T.D.1, and Prell, W.L.1 1 Brown University, Department of Geosciences 47 2.1 Abstract The alkenone Uk’37 ratio, one the most widely used and well-established sea surface temperature (SST) proxies, was developed for use in open ocean settings. Recent detections of alkenones in near-shore and lacustrine settings led to the exploration of Uk’37 in these ‘untraditional’ systems. However, in many cases, the C37 and C38 “fingerprints” deviate from open-ocean relationships, and the calibration of alkenone unsaturation to temperature is unclear. We present here the results of water column alkenone detections, collected from monthly-to-thrice weekly particulate filtrations, over four years (2009-2013) in the high salinity estuary Narragansett Bay. Alkenones were present in detectable quantities in 82% of the samples analyzed. In contrast to studies from other estuaries, the “brackish” C37:4 alkenone was present in less than 5% of our samples. When it was detected, it was confined to the lower salinity upper Bay, and in one year was detected together with a novel C37:3 isomer previously only detected in Arctic lakes. In the higher salinity lower Bay, Uk’37-inferred SST, calculated using an open ocean calibration, tracks the seasonal cycle with periods of both close agreement (≤ 1.5°C) and wide deviation (≥ 10°C). Uk’37 often underestimates actual SST during summer when alkenone concentrations are high and over estimates actual SST during winter when alkenone concentration is low. Based on comparison of sedimentary and water-column alkenone-to-aluminum ratios, we conclude warm Uk’37 offsets during winter most likely result from the resupension of surface sediment during winter storms. The causes of cool Uk’37 offsets during summer are not as clear and include 1) changes in nutrient concentrations, 2) changes in alkenone producing flora, and 3) yet undescribed changes in physiological life stage, demand, or environmental response. Despite this seasonal variability, mass weighted average Uk’37-inferred SST over the course of our study was 14.2°C, within 1.1°C of instrumental mean SST (13.1°C), and within 0.4°C of surface sediment Uk’37 determinations (13.8°C). 48 2.2 Introduction Certain organic compounds (biomarker) are sensitive recorders of important environmental and climatic information experienced during the life of the producing organisms. For example, in sediment records spanning the major ice ages, the ratio of two long chain alkyl ketones, one di-unsaturated (C37:2) one tri-unsaturated (C37:3), varied in concert with the isotopes of oxygen in foraminiferal tests, known to partially record the sea surface temperature (SST) at the time they lived [Brassell et al., 1986]. Further work showed these ketones, so called alkenone, are nearly ubiquitous in surface waters [Conte et al., 2006] and in marine sediments [Müller et al., 1998]. Most open ocean alkenone production comes from Group III haptophyte algae Emiliania huxleyi and Geophyrocapsa oceanica [Conte et al., 1994; Volkman et al., 1995]. Culture [Prahl and Wakeham, 1987; Prahl et al., 1988] and surface sediment [Müller et al., 1998] calibrations of Uk’37 against growth temperature led to the development of this ratio as a quantitative open ocean SST proxy (Uk’37 = 0.034 x SST + 0.039 between 0-28°C and a nominal calibration error of ±1.4°C [Prahl et al., 1988], revised slightly by Muller et al., [1998] to Uk’37 = 0.033 x SST + 0.044). The Uk’37 index has since found widespread use on ocean cores and outcropping marine sediments deposited over the past several million years [Cleaveland and Herbert, 2009; Herbert, 2003; Marlowe et al., 1990]. Because of the wide-spread success of Uk’37, and the ubiquity of alkenones, their application expanded to include environments in which Group III marine alkenone producers may not be present, including lakes [Chu et al., 2005; D'Andrea and Huang, 2005; Toney, J. L. et al., 2010; Zink et al., 2001] and shallow coastal waters [Mercer, 2002; Mercer et al., 2005; Salacup et al., in prep-a; Salacup et al., in prep-b; Schulz, 2000; Sicre et al., 2008]. The alkenone producing Group II haptophyes, Isochrysis and Chrysotila, which have different Uk’37-to-SST calibrations than E. huxleyi and G. oceanica [Nakamura et al., 2014; Versteegh et al., 2001], may dominate in these settings. The 49 distributions of C37 and C38 alkenones in these settings are often different from what is seen in open ocean samples [Mercer et al., 2005; Schulz, 2000; Schwab and Sachs, 2011; Toney, J.L. et al., 2012], and often include elevated concentrations of the tetra- unsaturated (C37:4) alkenone, and higher ratios of C37 to C38 alkenones (Chrysotila avg = 5.6, [Prahl et al., 1988]; Isochrysis avg = 10.9, Marlowe, 1984 than E. huxleyi (avg = 1.16, [Conte et al., 1994]) and G. oceanica (avg = 0.7, [Volkman et al., 1995]). For example, Chesapeake Bay hosts the alkenone producers E.huxleyi, G.oceanica, and I.galbana [Marshall, 1994] and Mercer et al [2005] detected high concentrations of the tetra-unsaturated alkenone in its waters and sediments. This is likely because lowered salinity shifts the alkenone producing populations in Chesapeake Bay away from the open ocean species toward brackish/lacustrine species that produce C37:4. Mercer et al [2005] suggested Uk’37 producers in Chesapeake Bay may adjust their Uk’37 ratio in response nitrate concentration as well as to SST, and that the slope of the response of unsaturation to SST was lower than that of the open ocean calibration. An in situ Uk’37-to- SST calibration applied to a local sediment core predicted SSTs as high as 55°C for this sub-tropical estuary [Mercer, 2002], while the widely-applied open ocean calibration of Prahl et al [1988] predicted unreasonably large SST variability (up to 15˚C over 200 years) leading the authors to conclude Uk’37 was not a viable SST proxy in Chesapeake Bay [Mercer, 2002]. More recently, Salacup et al [in prep-b] found that alkenones extracted from sediments in the high salinity Narragansett Bay (41.60˚N, 71.35˚W; Fig. 2.1) lacked the tell-tale contributions of the C37:4 alkenone, and instead displayed C37 and C38 alkenone patterns consistent with production by open ocean Group III producers E.huxleyi and G.oceanica. As part of this work, the authors also performed a survey of the Bay’s surface water. They found that while C37:4-free alkenone fingerprints dominated surface water samples from the high salinity mid and lower Bay, C37:4-rich fingerprints were produced in the 50 lower salinity Providence River and upper Bay. The discovery of alkenones in Narragansett represented the first detection of alkenone producing species in Narragansett Bay, thus the production systematics and responsible producers remain completely uncharacterized. Narragansett Bay’s mean volume-weighted salinity is high [29.96psu; Prell personal communication], suggesting the potential for alkenone production by well-calibrated open ocean Group III species such as E. huxleyi and G. oceanica. However, the seasonal dilution of the Providence River (Fig. 2.1) may provide a suitable habitat for Group II brackish alkenone producers, such as Isochrysis and Chrysotila, if present. Does salinity segregate alkenone producing populations in Narragansett Bay? In order to better understand what controls the spatial and temporal distribution of alkenone producers, alkenone fingerprints, alkenone concentrations, and Uk’37 ratios we have compared these variables, determined from samples taken across the Bay’s gradients of salt, nutrients, and productivity to instrumental measurements of salinity, temperature, and chlorophyll taken at each sampling station. 2.3 Methods 2.3.2 Environmental Data The Rhode Island Department of Environmental Management [2012] maintains a network of water quality monitoring buoys which provides context for our measurements. On August 14, 2012, we deployed a Seabird Conductivity, Temperature, and Depth Sensor (CTD), measuring SST, salinity, density, and fluorescence (a proxy for chlorophyll and thus water column productivity), on the dock owned by Save the Bay on Field’s Point (Fig. 2.1) in order to better interpret results obtained in the upper-Bay. Narragansett Bay SST gradients grow during the spring and summer months (Fig. 2.2 top; March-Aug) as shallower Providence River and upper Bay sites warm more rapidly than mid- or lower Bay sites. The reverse occurs in the fall and winter. Sea surface salinity (Fig. 2.2 middle) also varies along the Bay as the proximity of the major fresh 51 water sources results in lower, more variable surface salinity in the upper Bay and Providence River (Field’s Point) and higher and less variable surface salinity in the lower Bay (GSO). A bay-wide spring freshening results between March and May from the melting and runoff of winter snow and ice and low evapotranspiration, with the lowest surface salinity values ( as low as 0psu) occurring in the upper Bay and Providence River near the primary source of fresh water. Narragansett Bay experiences two broadly defined productivity blooms as expressed by measurements of water column chlorophyll (Fig. 2.2, bottom). Chlorophyll concentrations are generally greater near the source of waste-water supplied nutrients (Greenwich Bay and Conimicut Point) than in lower nutrient waters (GSO). Lastly, Narragansett Bay experiences annual sediment resuspension events associated with winter storms [Oviatt and Nixon, 1974]. 2.3.1 Sampling Scheme Two sea water facilities, one at the University of Rhode Island’s Graduate School of Oceanography (Fig. 2.1; GSO, mean sampling depth 3m) and one at Roger Williams University (Fig. 2.1; RWU, mean sampling depth 5m), were used to sample large volumes of water (500-1000L) from the Bay in short time (~1h) making high-resolution sampling readily achievable. We sampled at RWU between 28 October, 2009 and 4 February, 2011 and at GSO between 22 May, 2009 and 27 August, 2013. On 8 August, 2012, we added a sampling station at Save the Bay (StB) on Field’s Point in the Providence River (Fig. 2.1), in an attempt to capture C37:4 alkenone production there during low salinity events as in [Salacup et al., in prep-b]. For the period between 22 May, 2009 and 31 August, 2012, sea water at GSO spent 12-24 hours in a settling tank (average residence time) and was passed through a coarse sand filter before sampling. For the period between 27 June, 2012 and 27 August, 2013, we switched to a ‘raw’ line, the water for which came directly from the Bay without passing through settling tanks or sand filters. A period of overlap between 27 June and 31 August, 2012, during which we 52 sampled from both lines, allow us to detect and correct for line-dependent effects. Sea water at RWU flowed from nearby Mount Hope Bay into holding tanks then into the sea water system from which we sampled. Residence times in the holding tanks were similar to those at GSO (12-24 hours). Sea water at Fields Point came directly from the Bay’s surface (avg depth <1m) through a high-flow peristaltic pump with zero time lag. At each site, water column alkenones were filtered through a 1µm glass fiber filter (293mm diameter; Pall Corp.) at an approximate rate of 8L/min (avg total volume ~ 500L). We froze filters immediately for later biomarker extraction. In order to estimate the contribution of water column alkenones associated with sediment resuspension during winter storms, we compare the ratio of alkenones-to- aluminum on water column sample filters to that in surface sediments. Aluminium concentrations were determined using inductively coupled plasma mass spectrometry after digestion of a sub-sample in HF and HCl. Error on replicate sample analyses was within 2ppm/L (6%). Surface sediment metal concentrations are taken from Murray et al. [2007]. Surface sediment alkenone ratios and concentrations are taken from Salacup et al. [in prep-b]. 2.3.3 Alkenone Analysis We extracted alkenones from whole filters, and isolated them from co-eluting compounds, before quantification of the Uk’37 ratio. Filters were freeze-dried and extracted with dichloromethane on an Accelerated Solvent Extractor (Dionex, ASE200) at 150°C and 1500 psi to produce a total lipid extract (TLE). TLEs were then saponified in 2N KOH (in 5% H2O in MeOH) and heated for 2.5 hours at 65°C before addition of salt (5% in H2O) and HCl (6N to pH 2). We then extracted this mixture three times with hexane to remove the alkenone bearing fraction. This hexane was separated on silica gel using hexane, dichloromethane, and methanol to yield hydrocarbon, ketone (alkenone), and polar fractions, respectively. 53 We added two quantification standards (n-C36 and n-C37 alkanes) to all samples before injection from an autosampler into a 112°C CIS-PTV (cooled injection system- programmed temperature vaporizer) inlet operated in solvent vent mode, interfaced to an Agilent 6890 gas chromatograph. After the initial vent, the inlet temperature was ramped at 12°C/min to 240°C, held isothermally for 5 minutes, ramped again at 12°C to 320°C, and held isothermally for 2 minutes before cryogenic cooling. We used a 60m, 0.32mm ID, 0.10µm film DB-1 with a 5m fused guard column (DB-1 duraguard) for routine alkenone detection and quantification. A set of samples were later run on a 60m, 0.32mm ID, 0.10µm film DB-200 column to determine the presence or absence of a novel C37:3 isomer that has been shown to characterize some lacustrine/brackish alkenone producers in Arctic Lakes [Longo et al., 2013]. The oven temperature began at 90°C for 2 minutes, was ramped at 40°C/min to 255°C, at 1°C to 302°C, and at 10°C to 325°C where it was held isothermally for 20 minutes. We used hydrogen as a carrier gas and converted Uk’37 to SST using the calibration of Prahl et al. [1988]. Analytical accuracy, tracked via the injection of a laboratory Uk’37 sediment standard, was ±0.042 Uk’37 units (±0.1°C). Reproducibility of replicate sample injections averaged ±0.043 Uk’37 units (0.1°C) and within 8% for alkenone concentrations ([Alkenone]) (n=37 pairs). The estimated detection limit was 0.04ng/L for summed C37 alkenones. We identified individual alkenones in samples suspected of containing the C37:4 alkenone via GC-MS and comparison with accepted mass spectra. 2.4 Results Alkenones were detected in >99% of the lower Bay (GSO) samples analyzed (196 of 197 samples). Alkenone detections in the upper mid-Bay (RWU; 47%; 8 of 17) and Providence River (StB; 16%; 8 of 43) were less frequent. Tetra-unsaturated alkenones were detected in only 8 samples (3%), and were spatially and temporally constrained to the Providence River, upper, and upper mid-Bay during spring of 2010 54 and 2013. A novel C37:3 isomer [Longo et al., 2013] was detected in the upper-Bay in the spring of 2013 (Fig. 2.3). Samples from the upper-Bay spring bloom of 2010 were not analyzed for the isomer. The average sample spacing was 8 (1 – 114) days. We analyzed GSO samples from both the sand filtered and raw sea line between 27 June and 31 August, 2012 (Fig. 2.4) to check for sampling artifacts between the two different methods. Uk’37 in the filtered line ranged between 0.499 (12.0°C) and 0.775 (21.7°C) with an average Uk’37 of 0.610 (16.7°C). Uk’37 in the raw line was similar, between 0.433 (11.6°C) and 0.764 (21.3°C) with an average of 0.580 (15.9°C). The average alkenone concentration was 0.64ng/L (0.25-1.5ng/L) in the filtered line and 0.83ng/L (0.13-1.8ng/L) in the raw line, suggesting a ~25% loss of alkenones in the filtered line. The time series of Uk’37 inferred SST is very similar in both the raw and filtered line (Fig. 2.4) with the exception of samples taken during periods of rapid SST change, such as on 11 and 13 July and 22-29 August. Samples taken from the raw line on 11 and 13 July are ~4 and 2˚C warmer than in the filtered line, while samples taken from the raw line between 22 and 29 August are, on average, 3˚C cooler than the filtered line. The average Uk’37 difference between the raw and filtered lines is 0.07 (~0.8°C) (raw = 15.9, filtered = 16.7°C) suggesting the 25% loss of alkenones in the filtered line preferentially removed C37:3 alkenones. We therefore adjust the mean of the filtered line Uk’37-inferred SST time series down 0.07 Uk’37 units to correct for the effect. Water column alkenone concentrations and ratios varied throughout our study. The Uk’37 ratio varied between 0.18 (5.1°C) and 0.76 (21.3°C) (Avg = 0.47; 13.5°C) and concentrations varied between 0.04 and 8.2ng/L (avg = 0.86ng/L) (Fig. 2.5, 2.6). For samples containing the C37:4 alkenone, Uk’37 varied between 0.14 (2.8°C) and 0.30 (7.5°C) with an average of 0.21 (5.1°C) and concentrations were between 0.38 and 55.2ng/L (avg=8.9ng/L) (Fig. 2.6). Alkenone concentrations in samples from Narragansett Bay that lack the C37:4 moiety are on the low side of concentrations 55 detected in surface waters of the North Pacific [2-70ng/L; Harada et al., 2006], the Equatorial Pacific [0.5-15.5ng/L; Prahl et al., 2005], and Chesapeake Bay [10-360ng/L; Schwab and Sachs, 2011] and [10-4600ng/L; Mercer et al., 2005]. The mass weighted average Uk’37 inferred SST of all 37:4-free samples, regardless of sampling station, was 14.2°C. Comparison of the Uk’37 time series with the instrumental SST in the lower Bay (GSO) [Fig. 2.5, 2.6; Rhode Island Department of Environmental Management, 2012] highlights periods when Uk’37 closely tracks surface temperature (eg. March-July, 2010) and periods where it does not (eg. July-October, 2010). Generally speaking, Uk’37- inferred SST is warmer than instrumental SST (by up to 12°C) from December to March or April while Uk’37-inferred SST is cooler than instrumental (by up to 8°C) between April and October (Fig. 2.6). So, while Uk’37 does show a seasonal cycle, the amplitude of the signal is attenuated (Fig. 2.5). The cool Uk’37 offsets are coincident with (sometimes small) increases in water column alkenone concentrations (Fig. 2.5, 2.6). We also note a period between May and July 2012 when Uk’37 inferred SST appears to lag instrumental SST by approximately a month (Fig. 2.6). During this time Uk’37 also increases and decreases rapidly while following a general warming trend. 2.5 Discussion 2.5.1 The annual cycle Alkenone inferred SST in Narragansett Bay displays a direct but attenuated response to seasonal fluctuations in in situ water temperature (Fig. 2.5). While instrumental SST (2004-2010 average) varies between 3 and 21°C the generalized annual Uk’37 inferred SST (represented by a polynomial fit of the 2009 to 2013 alkenone data) only varies between 11°C (late-March) and 16°C (mid-August) (Fig. 2.5). Nonetheless, the mass weighted average Uk’37 inferred SST between 22 May, 2009 and 23 August, 2013 was 14.2°C, 1.1°C warmer than instrumental SST at GSO over the 56 same period (13.1°C), and only 0.4°C warmer than surface sediment Uk’37 measurements [13.8°C; Salacup et al., in prep-b] consistent with other work suggesting sediment Uk’37 records mean annual SST in Narragansett Bay with a small warm bias [Salacup et al., in prep-b], and implying our time series accurately captures the Uk’37 variability integrated in the sediments. Long (> 1 month) alkenone concentration-inferred blooms did not occur every year in Narragansett Bay. This may represent real inter-annual variability in the productivity of Narragansett Bay alkenone producers or the occurrence of very short-lived and therefore, unsampled, blooms. Sample spacing throughout the entire record (avg = 8 years) is quite tight compared to commonly reported bloom durations [2 – 3 months; Broerse et al., 2000], although blooms lasting on the order of days have been reported in macroaggregates of E. huxleyi [Cadee, 1985]. However, these macroaggregates, or any other form of E. huxleyi for that matter, have never been detected in Narragansett Bay, making it unlikely that such an event occurred in Narragansett Bay during our study without being detected. This annual heterogeneity in inferred alkenone blooms suggests water column alkenone studies that are limited in duration may miss important variability in alkenone concentrations and ratios that are important to making conclusions on the viability of the Uk’37 paleothermometer. For example, if we only analyzed alkenones from 2012, we would determine a mass weighted Uk’37 -inferred SST of 15.5°C. This is 3.5°C warmer than instrument measured SST (at GSO) in 2012 and would suggest that Uk’37 -inferred SST is actually quite a bit warmer than mean annual SST. Further, if we only sampled during a particular season, or for some part of the year, our conclusions on the viability of Uk’37 in Narragansett Bay would be very different. Blooms in alkenone production (Fig. 2.5) show a direct relationship with well- known spring and summer-fall increases in Bay productivity mainly driven by numerous 57 diatom species [Smayda and Borkman, 2008]. Average spring water column alkenone concentrations reach their maximum on 1 February, an average of 20 days before maximum average springtime chlorophyll maxima (20 February; Fig. 2.5), and average summer-fall alkenone concentrations reach their maximum on 20 August, an average of 30 days after maximum average summer-fall chlorophyll concentrations (20 July). The phase relationship between our seasonal alkenone results and the seasonal cycle of chlorophyll implies that summer-fall alkenone producers may prefer a warm low-nutrient late-succession niche, while springtime alkenone producers may prefer a cool high- nutrient early-succession niche, consistent with diatom-based succession patterns in Narragansett Bay which are paced by changes in SST [Karentz and Smayda, 1984]. The different SST tolerance of springtime and fall-summer alkenone producers implies different alkenone producing populations dominate during these two different alkenone bloom periods. 2.5.2 Salinity segregation of alkenone producers in Narragansett Bay Despite monthly sampling of the upper Bay between June 2009 and November 2010, and weekly sampling between July 2012 and May 2013 (total of 59 samples), we only detected the C37:4 isomer on 8 days, in the Providence River (at Fields Point) and upper Bay in March and April of 2010 and in the Providence River (at Fields Point) in April of 2013 (Fig. 2.6, 2.7). Each bloom last approximately one month (Fig. 2.6, 2.7) and was coincident with C37:4 free production in the mid and lower Bay consistent with two distinctly different alkenone producing populations (Fig 2.3) . Detection of the C37:4 alkenone often signals the failure of sedimentary Uk’37 to accurately record mean annual SST [Mercer, 2002; Mercer et al., 2005; Schulz, 2000]. The C37:4-rich bloom in the spring of 2010 coincided with historic rainfall and flooding in the watershed and salinities as low as 5psu in the Providence River near Fields Point. The C37:4-rich bloom detected in 2013 occurred while salinities were around 25psu (Fig. 58 2.7). These results are consistent with those detecting high contribution of C37:4 alkenones in other low salinity systems [Blanz et al., 2005; Mercer et al., 2005; Schulz, 2000; Schwab and Sachs, 2011]. In this light, our 2013 results suggest a cutoff may exist near a salinity of 25psu, below which Uk’37 is not viable. Samples from the upper-Bay spring bloom of 2013 also contained a novel C37:3 isomer, previously detected only in fresh Arctic Lakes dominated by Group I “Greenland” haptophytes [Longo et al., 2013]. This represents the first detection of this compound, thought to be a biomarker for fresh water Group I haptophytes, in a non-Arctic saline (~25psu) environment, greatly expanding the possible range of these alkenone producers. The detection of C37:4 alkenones and this Group I isomer in the Providence River and upper Bay, with C37:4 free production in the mid and lower Bay (Fig. 2.3), is again consistent with the existence of two separate alkenone producing populations, and implies the limited advection and/or mixing of alkenones in the Bay, at least during the season of production. The inference of only minor lateral advection of alkenones is also consistent with the presence of consistent Uk’37 offsets, reflecting in situ SST gradients, between geographically separated sediment cores in the Bay [Salacup et al., in prep-b]. In contrast to the Providence River and upper Bay, samples taken from the mid and lower Bay bared all the hallmarks of production by high salinity, open ocean, Group III haptophytes. The complete absence of C37:4 alkenones in these samples is consistent with production by E. huxleyi and G. oceanica in high salinity systems in which Uk’37 reflects mean annual SST [Herbert, 2003]. The ratio of C37 to C38 alkenones in these samples averaged 0.8 (0.26 – 1.82), again consistent with open ocean Group III production [Herbert, 2003], limited inputs of Group II or ‘brackish’ production, and low rates of alkenone advection. 59 2.5.3 Deviations between Uk’37-inferred and instrumental SST Surface sediment Uk’37 reflects the annually integrated Uk’37 signal produced in surface waters and exported to depth. In the open ocean, the integrated Uk’37-inferred SST is close to the mean annual SST of the overlying surface water [Conte et al., 2006]. However, in some settings, particularly high-latitude, upwelling, or coastal locations the Uk’37 ratio may reflect an SST other than annual mean [Prahl et al., 2010]. Several open ocean sediment traps studies report annually integrated Uk’37 ratios which reflect mean annual SST while seasonal Uk’37-inferred SSTs are too warm in the winter and too cool in the summer [Harada et al., 2006; Lee et al., 2011; Prahl et al., 2001; Seki et al., 2007; Sikes et al., 2005; Yamamoto et al., 2007]. Cool summer offsets are often coincident with maxima in alkenone flux [Lee et al., 2011; Sikes et al., 2005; Yamamoto et al., 2007]. This ‘warm winter / cool summer’ problem is explained in different ways. Some suggest cool Uk’37 inferred SSTs in summer result from a change in the depth habitat of the producing organisms from the warmer surface to the cooler thermocline [Prahl et al., 2001] and/or nutrient depletion [Sikes et al., 2005]. Others suggest the warm offsets during winter result from light limitation induced by the collapse of the seasonal thermocline and thus the greater time spent by alkenone producing species at greater water depths [Harada et al., 2006; Sikes et al., 2005]. Warm offsets during winter have also been explained by invoking the advection of allocthonous material produced in warmer water masses [Harada et al., 2006], and/or sediment resuspension [Yamamoto et al., 2007]. One attempt at explaining both sides of the ‘warm winter / cool summer’ problem invokes a seasonally variable time lag (1 month summer, 3 month winter) between alkenone production and alkenone arrival at a 1000m sediment trap [Lee et al., 2011]. It is important to note however that despite these interesting seasonal intricacies, all of these sediment trap studies reported annually integrated Uk’37 SSTs that were very close to mean annual. We analyzed our time series within an environmental context 60 provided by a network of water quality monitoring buoys monitored by the Rhode Island Department of Environmental Management [2012]. 2.3.5.1 Warm Uk’37 offset during winter We detect this ‘warm winter / cool summer’ problem repeatedly in our time series (Fig. 2.6). The shallow nature of Narragansett Bay (avg = 6.4m [Boothroyd, 2008]) eliminates long time lags between surface production and sampling depth, making the hypothesis of Lee et al [2011] implausible in this setting. It is possible that alkenone producers are light limited during the winter [Harada et al., 2006; Sikes et al., 2005], particularly during and following winter storms which can resuspend surface sediment [Oviatt and Nixon, 1974]. However, measurements of secchi depth are on average relatively high during winter months suggesting waters are clear and light permeable (Fig. 2.6). The average Uk’37 inferred SST during the warm winter offsets (defined as Uk’37 inferred SST >1.5°C warmer than instrumental SST) is approximately 12.9°C. Water of this temperature is present on the mid-Atlantic Bight and along the north wall of the Gulf Stream during the winter months [Levitus, 1982] and so we cannot rule out advective contributions of alkenones from these sources, although, as discussed above, evidence does suggest advection of alkenones in the Bay is limited. The shallow depth of Narragansett Bay does lead to sediment resuspension, especially during winter because of strong regional storms [Oviatt and Nixon, 1974]. The ratio of alkenones to aluminium (A proxy for lithic material) on our sample filters can inform how much, if any, of the alkenones present in the waters of the Bay during winter are associated with sediment resuspension. During summer, the ratio of alkenones to aluminum (Al) on our water filters is ~13 x 10-5. During winter, the ratio of alkenones to Al on our filters is one fifteenth of the summer value (~0.84 x 10-5), consistent with decreased alkenone production, and is very close to the value detected in the Bay’s surface sediments (1.4 x 61 10-5) suggesting resuspended alkenones may dominate the Uk’37 signal sampled at this time. 2.3.5.2 Cool Uk’37 offset during summer Cool Uk’37 offsets have been explained by invoking 1) a change in the depth habitat of the alkenone producing organisms from warm surface waters to the cooler thermocline [Prahl et al., 2010], and/or 2) nutrient depletion [Sikes et al., 2005]. The day on which Uk’37-inferred SST starts cooling relative to instrument inferred SST is nearly the same in each of the four years this phenomenon was detected: 7 July 2010, 13 July 2011, 11 July 2012, and 19 July 2013 (Fig. 2.6). Bottom water in Narragansett Bay is generally only 1-3°C cooler than surface water this time of year [Rhode Island Department of Environmental Management, 2012] and so sub-surface alkenone production in ~12°C water (such as that inferred by Uk’37 in the summer of 2010) is not plausible. Water of this temperature is only found as far away as the Labrador Sea this time of year, thus the likelihood of advection of alkenones produced in a cooler water mass is low, particularly given evidence suggesting alkenone advection is not an important process in the Bay. Nutrient depletion, and its ability to lower the growth rate of alkenone producers, may cause lower than expected Uk’37 ratios in E. huxleyi cultures, while also causing an increase in the inter-cellular concentration of alkenones [e.g. Prahl et al., 2003 and references therein]. When concentrations of nitrate dropped below 4µM (and 0.2µM phosphate), a 0.11 Uk’37 unit decrease, and a tripling of inter-cellular alkenone concentrations, was detected in culture Uk’37 at constant temperature [Prahl et al., 2003]. Using the calibration of Prahl et al [1988] this results in Uk’37 inferred SST being 3.2°C cooler than the growth temperature. In contrast, Epstein et al. [1998], found that nutrient limitation in culture experiments lead to warmer Uk’37 ratios (up to 5.5°C using the calibration of Prahl et al[1988]) and 15 a fold increase in inter-cellular alkenone concentrations. The cool Uk’37 offset in 2010 coincides with an increase in water column 62 alkenones from ~0.5ng/L to almost 7ng/L (13 fold increase; Fig. 2.6). Water column nitrate+nitrite concentrations (Fig. 2.6) measured in the lower Bay (GSO) are below 4µM at the time of the detected cool offset [Krumholz, 2012]. However even lower concentrations (~0.4µM), immediately preceding the cool offset in 2010 and 2011, coincide with periods during which Uk’37-inferred SST agrees well with instrumental SST, suggesting nitrogen concentrations are not driving cool Uk’37 offsets in Narragansett Bay. Alternatively, a change in the alkenone producing flora may be responsible for the cool offset in summer Uk’37. Culture [Conte et al., 1998; Ono et al., 2012; Versteegh et al., 2001] and water column [Conte et al., 2001; Ternois et al., 1997] studies have previously demonstrated the dependence of the Uk’37-to-SST calibration on the alkenone producing species, group of species, or species strain. Thus, an environment hosting several alkenone producing species, each having different ecological and physiological cues controlling when and how it produces alkenones, may instead require a ‘set’ of Uk’37-to-SST calibrations, one for each producing species or group of species. Each calibration in the set would accurately predict the in situ SST during the time of year when the matching species or species group was dominant in the system. However, applying the calibrations of Isochrysis (Uk’37 = 0.016T – 0.0607) and/or Chrysotila (Uk’37 = 0.0377T – 0.5992) [Schwab and Sachs, 2011 and references therein ]to Uk’37 ratios detected during these cool offsets does not change the direction of temperature change from cooling to warming since the slopes of the regressions are still positive, and results in unrealistically high inferred SSTs (31 and 27°C for Isochrysis and Chrysotila, respectively, for the period of deviation in 2010). Some researchers suggest different alkenone producing species may produce alkenones with a different characteristic distribution of the individual alkenone and alkenoate isomers [e.g. Volkman et al., 1995]. For example, Isochrysis galbana, often associated with low salinity coastal zones and lakes [Versteegh et al., 2001], produces 63 alkenone fingerprints rich in C37:4 and lacking C38 methyl alkenone isomers [Ono et al., 2012]. Inspection of alkenone chromatography (to include presence/absence of C37:4, presence/absence of C38 methyl alkenones [Ono et al., 2012], the ratio of C37 to C38 alkenones [Herbert, 2003], and the ratio of Uk’37 to Uk38me [Herbert, 2003]) before, during, and after cool summer offsets suggests an unchanging alkenone producing population, as the alkenone “fingerprints” are analytically identical, and lack the C37:4 alkenone associated with brackish Group II production. If a different species is producing alkenones in Narragansett Bay during summer months it has a remarkably similar alkenone distribution as the species producing alkenones the rest of the year, but a different calibration between Uk’37 and SST, or there are undocumented for non-thermal effects on the Uk’37 ratio. Resolution on this matter awaits more detailed investigation coupling biomarker analyses to genetic identification of possible alkenone producers. Lastly, it may be that some as-yet-undescribed physiological life stage, demand, or response is affecting Uk’37 ratios synthesized during summer in the Bay. The day on which Uk’37-inferred SST starts cooling with respect to instrument measured SST is nearly the same in each of the three years this phenomenon was detected. Perhaps the alkenone producing population responds to the increased day length (degree days) or the effects of increased photosynthetically active radiation (PAR). The alkenone producer E. Huxleyi is amazingly tolerant of high PAR conditions and likely exploits this capability to outcompete other algae (and bloom) under highly stratified, low nutrient conditions [Nanninga and Tyrrell, 1996]. However, a search of the literature yields no information on the effects of increased PAR on Uk’37 ratios outside of PAR’s effects in concert with nutrient depletion [Prahl et al., 2003 and references therein]. 64 2.6 Conclusions The four year-long water-column alkenone time series presented here sheds light on the behavior of this important SST proxy in a coastal setting. We detected alkenones displaying C37:4 free fingerprints in 82% of the samples analyzed suggesting open ocean- like alkenone production by Group III haptophytes may be abundant in coastal settings despite absence of previous detection. Tetra-unsaturated alkenones were detected in only 3% of the samples, and were spatially and temporally constrained to low salinity parts of the Bay during spring, suggesting the Bay is home to two distinctly different alkenone producing populations that are segregated by salinity. The mass weighted average Uk’37-inferred SST during our study was 14.2°C, 1.1°C warmer than instrumental SST over the same period (13.1°C), and only 0.4°C warmer than surface sediment Uk’37 measurements suggesting integrated water column alkenone production reflects mean annual SST. Comparison of the Uk’37 time series with instrumental SST highlights periods when Uk’37 closely tracks instrumental and periods where it does not. The seasonal cycle does support in situ production of alkenones, but also suggests a complex relationship between Uk’37 and SST and high temporal resolutions. Sediment and water column ratios of alkenones-to-aluminium concentration suggest warm Uk’37 offsets during winter are likely caused by resuspension of sedimentary alkenones during winter storms. Cool Uk’37 offsets during summer may be caused by a combination of: 1) nutrient limitation, 2) changes in the alkenone producing flora, and/or 3) yet undiscovered physiological responses to non-thermal environmental variables. Heterogeneity in alkenone concentrations and ratios between years suggests timeseries that are limited to one or even two years may miss important variability in these parameters that may lead one to infer a seasonal bias to Uk’37. 65 2.7 Figures Figure 2.1. Map of Narragansett Bay and points of interest: Sampling stations (filled circles) GSO = University of Rhode Island’s Graduate School of Oceanography; RWU = Roger Williams University; Field’s Point. Arrows represent major sources of fresh water. Core and surface sediment locations from Salacup et al [in prep-a; in prep-b] (stars; GB = Greenwich Bay (NB12/45v); PC = Potter’s Cove (NB25/42v); FI = Fox Island (NB44v)). Stippled lines denote the ecofunctional boundaries between sections of the Bay [Costa- Pierce and Desbonnet, 2008]. 66 Figure 2.2. Plots of the annual cycle of near surface (A) SST, (B) salinity, and (C) Chl for Conimicut Point (CP, green), Greenwich Bay (GB, red), Potter’s Cove (PC, blue), and the Graduate School of Oceanography (GSO, black, see Fig. 2.1) provide important context for our Uk’37 results. A number of year’s data (CP = 2009 - 2012; GB = 2007 - 2010; PC = 1996 – 2010; GSO = 1996 - 2010) was averaged then treated with a low pass filter to produce a monthly smooth. The patterns depicted are resistant to the length of time averaged and/or the removal of any one year’s data. 67 Figure 2.3 Representative chromatograms from the high salinity mid and lower Bay (top) and low salinity upper Bay and Providence River (bottom). The novel C37:3 isomer of Longo et al. [2013] resolves as a broad right hand shoulder on the typical C37:3 alkenone. 68 Figure 2.4. Plots of Uk’37 inferred SST and alkenone concentration comparing these two variables between samples taken after 12-24 hours in a settling tank and pre-filtration (circles) and those taken ‘raw’ directly from NB (diamonds). 69 Month Month Figure 2.5. Plot of each years’ time series of alkenone ratios and concentrations onto a generic annual cycle and smoothed with a polynomial fit to highlight the 4 year average annual cycle. Alkenone ratios and concentrations are compared with instrumentally measured SST and chlorophyll measurements (GSO). 70 Month Figure 2.6. Plots comparing the time series of Uk’37-inferred and instrumental SST (A), alkenone concentrations (B), and nutrients (C). (A) Comparison of Uk’37-inferred SST (circles and diamonds) with instrumental SST (solid black line) highlights periods of both agreement and deviation. Periods in summer during which Uk’37 SST is cooler than instrumentally measured SST coincide with periods of (sometimes slightly) increased alkenone concentrations (B). (A and B) Good agreement in Uk’37 SST and alkenone concentration is noted between samples taken from settled prefiltered water (circles) and raw water (diamonds). Periods of cool summer Uk’37 SST may or may not coincide with times of nutrient nitrogen depletion (C, solid line, phosphate = stipled line). 71 Figure 2.7. A zoomed-in plot of C37:4-rich alkenone production at Fields Point in April of 2013 showing alkenone concentration and fluorescence (A), salinity (B) and Uk’37 inferred and instrumental SST. 72 2.8 References Bendle, J., A. Rosell-Mele, and P. Ziveri (2005), Variability of unusual distributions of alkenones in the surface waters of the Nordic seas, Paleoceanography, 20(2). Blanz, T., K. C. Emeis, and H. Siegel (2005), Controls on alkenone unsaturation ratios along the salinity gradient between the open ocean and the Baltic Sea, Geochimica et Cosmochimica Acta, 69(14), 3589-3600. Boothroyd, J. C. a. A., Peter, V. (2008), Geologic and Contemporary Landscapes of the Narragansett Bay Ecosystem, in Science for Ecosystem-Based Management, edited by A. D. a. B. A. Costa-Pierce, p. 570, Springer, New York, NY. Brassell, S. C., G. Eglinton, I. T. Marlowe, U. Pflaumann, and M. Sarnthein (1986), Molecular Stratigraphy - a New Tool for Climatic Assessment, Nature, 320(6058), 129- 133. Cleaveland, L. C., and T. D. Herbert (2009), Preservation of the alkenone paleotemperature proxy in uplifted marine sequences: A test from the Vrica outcrop, Crotone, Italy, Geology, 37(2), 179-182. Conte, M. H., A. Thompson, and G. Eglinton (1994), Primary production of lipid biomarker compounds by Emiliania huxleyi: results from an experimental mesocosm study in fjords of southern Norway, Sarsia, 79, 319-332. Conte, M. H., A. Thompson, D. Lesley, and R. P. Harris (1998), Genetic and physiological influences on the alkenone/alkenoate versus growth temperature relationship in Emiliania huxleyi and Gephyrocapsa oceanica, Geochimica et Cosmochimica Acta, 62(1), 51-68. 73 Conte, M. H., J. C. Weber, L. L. King, and S. G. Wakeham (2001), The alkenone temperature signal in western North Atlantic surface waters, Geochimica et Cosmochimica Acta, 65(23), 4275-4287. Conte, M. H., M. A. Sicre, C. Ruhlemann, J. C. Weber, S. Schulte, D. Schulz-Bull, and T. Blanz (2006), Global temperature calibration of the alkenone unsaturation index (U-37(K ')) in surface waters and comparison with surface sediments, Geochemistry Geophysics Geosystems, 7, -. Costa-Pierce, B. A., and A. Desbonnet (2008), An "Ecofunctional" Approach to Ecosystem-based Management for Narragansett Bay, in Science for Ecosystem-based Management, edited by A. Desbonnet and B. A. Costa-Pierce, pp. 545-562, Springer, New York, NY. Epstein, B. L., S. D'Hondt, J. G. Quinn, J. P. Zhang, and P. E. Hargraves (1998), An effect of dissolved nutrient concentrations on alkenone-based temperature estimates, Paleoceanography, 13(2), 122-126. Harada, N., M. Sato, A. Shiraishi, and M. C. Honda (2006), Characteristics of alkenone distributions in suspended and sinking particles in the northwestern North Pacific, Geochimica et Cosmochimica Acta, 70(8), 2045-2062. Harada, N., K. H. Shin, A. Murata, M. Uchida, and T. Nakatani (2003), Characteristics of alkenones synthesized by a bloom of Emiliania huxleyi in the Bering Sea, Geochimica et Cosmochimica Acta, 67(8), 1507-1519. Herbert, T. D. (2003), Alkenone paleotemperature determinations, in Treatise in Marine Geochemistry, edited by H. Elderfield, pp. 391-432, Elsevier, Amsterdam. 74 Krumholz, J. (2012), Spatial and temporal patterns in nutrient standing stock and mass- balance in response to load reductions in a temperate estuary, University of Rhode Island, Narragansett. Lee, K. E., B. K. Khim, S. Otosaka, and S. Noriki (2011), Sediment trap record of alkenones from the East Sea (Japan Sea), Organic Geochemistry, 42(3), 255-261. Levitus, S. (1982), Climatological Atlas of the World Ocean, 173 pp., U.S. Government Printing Office, Washington D.C. Marlowe, I. T., S. C. Brassell, G. Eglinton, and J. C. Green (1990), LONG-CHAIN ALKENONES AND ALKYL ALKENOATES AND THE FOSSIL COCCOLITH RECORD OF MARINE-SEDIMENTS, Chem Geol, 88(3-4), 349-375. Mercer, J. (2002), Extending the Spatial and Temporal Applications of the Uk37 Index as a Marine Paleotemperature Proxy, 129 pp, Dartmouth College, Hanover, New Hampshire. Mercer, J., M. Zhao, and S. Colman (2005), Seasonal variations of alkenones and Uk37 in the Chesapeake Bay water column, Estuarine, Coastal and Shelf Science, 63(4), 675- 682. Müller, P. J., G. Kirst, G. Ruhland, I. von Storch, and A. Rosell-Melé (1998), Calibration of the alkenone paleotemperature index U37K′ based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S), Geochimica et Cosmochimica Acta, 62(10), 1757-1772. Murray, D. W., W. L. Prell, C. E. Rincon, and E. Saarman (2007), Physical Property and Chemical Characteristics of Surface Sediment Grab Samples from Narragansett Bay 75 and the Providence and Seekonk Rivers, a summary of the Brown University Narragansett Bay Sediment Project (BUNBSP). Narragansett Bay Estuary Program Report, NBEP-07-127Rep. Nanninga, H. J., and T. Tyrrell (1996), Importance of light for the formation of algal blooms by Emiliania huxleyi, Marine Ecology Progress Series, 136(1-3), 195-203. Ono, M., K. Sawada, Y. Shiraiwa, and M. Kubota (2012), Changes in alkenone and alkenoate distributions during acclimatization to salinity change in Isochrysis galbana: Implication for alkenone-based paleosalinity and paleothermometry, Geochemical Journal, 46(3), 235-247. Oviatt, C. A., and S. Nixon (1974), Sediment Resuspension and Deposition in Narragansett Bay, Estuarine and Coastal Marine Science, 3, 201-217. Prahl, F. G., and S. G. Wakeham (1987), Calibration of Unsaturation Patterns in Long- Chain Ketone Compositions for Paleotemperature Assessment, Nature, 330(6146), 367- 369. Prahl, F. G., L. A. Muehlhausen, and D. L. Zahnle (1988), Further evaluation of long- chain alkenones as indicators of paleoceanographic conditions, Geochimica et Cosmochimica Acta, 52(9), 2303-2310. Prahl, F. G., C. H. Pilskaln, and M. A. Sparrow (2001), Seasonal record for alkenones in sedimentary particles from the Gulf of Maine, Deep-Sea Res. Part I-Oceanogr. Res. Pap., 48(2), 515-528. Prahl, F. G., G. V. Wolfe, and M. A. Sparrow (2003), Physiological impacts on alkenone paleothermometry, Paleoceanography, 18(2). 76 Prahl, F. G., J. F. Rontani, N. Zabeti, S. E. Walinsky, and M. A. Sparrow (2010), Systematic pattern in U37K′ – Temperature residuals for surface sediments from high latitude and other oceanographic settings, Geochimica et Cosmochimica Acta, 74(1), 131-143. Rhode Island Department of Environmental Management (2012), Narragansett Bay Fixed-Site Monitoring Network (NBFSMN), edited, p. Data available at www.dem.ri.gov/bart, Office of Water Resources. Rosell-Mele, A. (1998), Interhemispheric appraisal of the value of alkenone indices as temperature and salinity proxies in high-latitude locations, Paleoceanography, 13(6), 694-703. Salacup, J., T. D. Herbert, and W. Prell (in prep-a), Alkenones in Narragansett Bay (RI) reveal and high-resolution record of North Atlantic Common Era climate. Salacup, J., J. Farmer, T. D. Herbert, and W. L. Prell (in prep-b), Alkenone production with open ocean characteristics in a near-shore setting: Narragansett Bay, R.I. Schulz, H.-H., Schoener, A., and Emeis, K. C. (2000), Long-chain alkenone patterns in the Baltic Sea an ocean-freshwater transition, Geochimica et Cosmochimica Acta, 64, 469-477. Schwab, V. F., and J. P. Sachs (2011), Hydrogen isotopes in individual alkenones from the Chesapeake Bay estuary, Geochimica et Cosmochimica Acta, 75(23), 7552-7565. Seki, O., T. Nakatsuka, K. Kawamura, S. I. Saitoh, and M. Wakatsuchi (2007), Time- series sediment trap record of alkenones from the western Sea of Okhotsk, Mar Chem, 104(3-4), 253-265. 77 Sicre, M. A., E. Bard, U. Ezat, and F. Rostek (2002), Alkenone distributions in the North Atlantic and Nordic sea surface waters, Geochem. Geophys. Geosyst., 3. Sikes, E. L., T. O'Leary, S. D. Nodder, and J. K. Volkman (2005), Alkenone temperature records and biomarker flux at the subtropical front on the chatham rise, SW Pacific Ocean, Deep-Sea Res. Part I-Oceanogr. Res. Pap., 52(5), 721-748. Ternois, Y., M. A. Sicre, A. Boireau, M. H. Conte, and G. Eglinton (1997), Evaluation of long-chain alkenones as paleo-temperature indicators in the Mediterranean Sea, Deep- Sea Res. Part I-Oceanogr. Res. Pap., 44(2), 271-286. Versteegh, G. J. M., R. Riegman, J. W. de Leeuw, and J. H. F. Jansen (2001), U(37)(K ')values for Isochrysis galbana as a function of culture temperature, light intensity and nutrient concentrations, Organic Geochemistry, 32(6), 785-794. Volkman, J. K., S. M. Barrett, S. I. Blackburn, and E. L. Sikes (1995), Alkenones in Gephyrocapsa-Oceanica - Implications for Studies of Paleoclimate, Geochimica et Cosmochimica Acta, 59(3), 513-520. Yamamoto, M., A. Shimamoto, T. Fukuhara, H. Naraoka, Y. Tanaka, and A. Nishimura (2007), Seasonal and depth variations in molecular and isotopic alkenone composition of sinking particles from the western North Pacific, Deep-Sea Res. Part I-Oceanogr. Res. Pap., 54(9), 1571-1592. 78 CHAPTER THREE _______________________________________________________ ALKENONES IN NARRAGANSETT BAY (RHODE ISLAND, U.S.A.) REVEAL A HIGH-RESOLUTION RECORD OF NORTH ATLANTIC COMMON ERA CLIMATE Salacup, J.M.1, Herbert, T.D.1, and Prell, W.L.1 1 Brown University, Department of Geosciences 79 3.1 Abstract Quantitative high resolution paleo-climate records provide context for current global warming and a target against which climate models are compared and judged. However, most high-resolution records of climate for the past 2000 years (the Common Era) are based on tree rings, and so are biased toward climate changes on land. Recent advances in the use of organic biomarker proxies for sea surface temperature (SST) can now provide data for coastal ocean environments. Here we present a Uk’37-inferred SST stack, based on three separate records from Narragansett Bay (Rhode Island, U.S.A.), which covers the past 1500 years with an average sample spacing of 4 years. A consistent SST offset between different cores sites reflects observed modern SST gradients, implying the Uk’37 signal is produced in the Bay. The records contain SST structure consistent with the Medieval Climate Anomaly, the Little Ice Age, and an anomalously high rate of modern warming (~1.9°C/100years). Decadal and centennial structure show evidence of a relationship with the North Atlantic and Arctic Oscillation Oscillations, respectively, in which relatively warm SSTs accompany NAO/AO positive periods, and vice versa. This positive relationship, and the timing of the Medieval Climate Anomaly and Little Ice Age in comparison to other North Atlantic records, highlights a north-south evolution in North Atlantic SSTs that is persistent on millennial timescales. 3.2 Introduction Understanding the evolution of North Atlantic sea surface temperature (SST) variability during the Common Era (last 2000 years) is important for predicting the impacts of global warming on marine and terrestrial ecosystems. The North Atlantic Ocean (Fig. 3.1) is an important gateway thru which heat collected in the ocean’s mixed layer is transported into the deep ocean. Thus, North Atlantic circulation and heat transport are implicated as important players in global climate change on a variety of 80 spatio-temporal scales. For example, the Atlantic Multidecadal Oscillation [AMO; Kerr, 2000] is a periodic (65-80 years) basin-wide SST anomaly detected in instrumental data that is correlated with changes in precipitation over the southeastern United States, Brazil, and the Sahel as well as the frequency of severe and increasingly expensive Atlantic hurricanes [Enfield et al., 2001; Ting et al., 2009]. In turn, atmospheric pressure anomalies such as the North Atlantic Oscillation [NAO; Hurrell, 1995; Visbeck et al., 2001] and the Arctic Oscillation [AO; Thompson and Wallace, 1998] influence North Atlantic SST patterns (Fig. 3.1 right). AO/NAO-like variability is associated with changes in Gulf Stream and North Atlantic Current intensity [Curry and McCartney, 2001], SST patterns [MERCINA, 2001; Pickart et al., 1999], cross-Atlantic wind speed and prevailing direction (and therefore northeastern United States and European storminess) [Hurrell, 1995; Visbeck et al., 2001], North Atlantic Deep Water production and export [Dickson et al., 1996], the mean position of the Gulf Stream [Taylor and Stepens, 1998], and marine ecosystem structure and productivity [Borkman and Smayda, 2009]. Kim and McCarl [1991] point out that the effects of the AO/NAO on European and North American agriculture are so powerful that the ability to forecast and exploit AO/NAO variability may be worth as much as 600 million to 1.1 billion U.S. dollars per year. The strength, pacing, and consequences of this natural climate variability is known to us through meteorological and economic data, but how, or even if [Booth et al., 2012], these modes of variability existed in the past requires a network of spatially and temporally detailed paleoclimate information. For example, paleo reconstructions of the Medieval Climate Anomaly (MCA) and the Little Ice Age [LIA; Jones and Mann, 2004], both lasting centuries, show they influenced hemispheric-to-global temperature variability, although the mechanisms, timing, duration, strength, and finer scale evolution of these anomalies are still poorly understood [Ahmed et al., 2013; Cronin et al., 2010; Mann et al., 2009]. 81 Until recently, high resolution temperature variability during the Common Era was almost exclusively described using tree rings [IPCC, 2007] and so is biased toward changes on land and the biases of that particular proxy. Paleo-reconstructions of important climate modes, such as the AMO [Gray et al., 2004] and AO/NAO [Cook, Edward R. et al., 2002; D' Arrigo et al., 2003], are often used as targets in modeling studies [Braconnot et al., 2012], but again are largely based on tree rings which shed little light on the ocean’s role in these modes of variability. Today, the world ocean absorbs and stores ~90% of the warming that has occurred since 1955 [Levitus et al., 2005]. A sink, source, and transporter of this much heat is a powerful climate force and thus demands a detailed characterization. New efforts are aimed at using organic biomarkers and biogenic carbonates produced and deposited in marine shelf and coastal environments to reconstruct SST because of the high sedimentation rates there [Cronin et al., 2010; DeLong et al., 2012; Salacup et al., in prep-b; Sejrup et al., 2011; Sicre et al., 2008; Wanamaker et al., 2011]. Unfortunately, carbonate-based proxies, such as the oxygen isotope [e.g. Wanamaker et al., 2011] or magnesium-to-calcium ratios in biogenic carbonates [Cronin et al., 2010] often do not work well in estuarine settings, due to the lack of adequate sedimentary carbonate and/or the confounding effects of salinity [e.g. Dissard et al., 2010] . The development and application of organic biomarker SST proxies hold promise in these dynamic, carbonate-deprived, mud-rich settings if they can be shown to be robust. Such records would need to be reproducible and compared with other paleo-SST records to establish the validity of the reconstruction before more detailed environmental interpretation. Salacup et al. [in prep- b] recently showed the Uk’37 SST proxy, based on biomarkers called alkenones, passes these tests in Narragansett Bay (Fig. 3.1 inset). Here we report results on the application of Uk’37 to Narragansett Bay sediment cores in order to reconstruct a Common Era SST record for southern New England. 82 Narragansett Bay (Fig. 3.1) is an ideal place to monitor changes in North Atlantic climate as it is situated near the boundary of cool northern sourced Labrador Sea Water and the warm southern sourced Gulf Stream [Levitus, 1982] and is sensitive to changes in the AO/NAO [Borkman and Smayda, 2009]. Our reconstruction employs five sediment cores, from three locations, sampled at ≤ 2cm resolution (Fig. 3.1 inset). Multiple cores will allow us to empirically determine the precision and reproducibility of Uk’37 in Narragansett Bay in relationship to the relatively small temperature signals over the Common Era. 3.3 Chronology The sampling interval was adjusted to approximate sub-decadal sample spacing throughout the record. Core NB44v comes from the lower mid-Bay near Fox Island (Fig. 3.1, inset). Cores NB25 and NB42 (upper mid-Bay, Potter’s Cove), and NB12 and NB45 (upper-Bay, Greenwich Bay), were composited using bulk chemical properties. We took a multi-proxy approach to our age control (Table 3.1), using the characteristic increase in sedimentary Pb concentrations associated with industrialization at 1850 [Bricker, 1993; Corbin, 1989; this study], unsupported Pb210 activities, and well-established historical pollen horizons [~1700; Parshall et al., 2003]. In addition, we radiocarbon dated foraminiferal tests (Elphidium sp.) after picking, sonication, and low temperature drying (≤ 50°C). All measurements were made at the National Ocean Sciences Accelerator Mass Spectrometry facility (NOSAMS). Reported radiocarbon ages were corrected for the standard marine reservoir effect (~400 years) minus a local reservoir effect [~120 years; McNeely et al., 2006], resulting in a 280 year reservoir correction. All dates were calibrated using OxCal. Calibrated dates were then modeled (Fig. 3.2) using a mixed regression technique [Heegaard et al., 2005]. Age model error (2σ) for all cores was estimated using the mixed regression technique and generally increases downcore (Fig. 3.2 bottom): NB25/42 avg. error = ±120 (23-182) years; NB12/45 avg. error = ±78 (4- 83 163) years; NB44 avg. error = ±70 (10-140) years. Two foraminiferal based radiocarbon determinations from NB44v were discarded as outliers after initial modelling. Age control is detailed in Table 3.1. Sedimentation rates are variable through time in Narragansett Bay (Fig. 3.2, bottom). Average sedimentation rates over the past 1500 years are 0.26, 0.24, and 0.11cm/yr at Potter’s Cove, Greenwich Bay, and Fox Island, respectively, consistent with each site’s distance from the primary source of sediment in the upper Bay and Providence River (Fig. 3.1, inset). Sampling spacing was between 2 and 15 (avg = 6) years in Potter’s Cove, between 1 and 25 (avg = 10) years in Greenwich Bay, and between 5 and 49 (avg = 14) years near Fox Island. The amount of time homogenized in each 1cm sample was between 1 and 20 (avg = 7) years in Potter’s Cove, between 1 and 37 (avg = 8) years in Greenwich Bay, and between 3 and 19 (avg = 9) years near Fox Island. 3.4 Temperature Reconstruction Alkenone biomarkers, produced by photoautotropic haptophyte algae, form the foundation of the well-established Uk’37 SST proxy that is widely-applied for open-ocean paleothermometry [see Herbert, 2003 for review]. The proxy was initially developed and calibrated for use in open ocean sediments [Brassell et al., 1986] underlying regions where alkenone production is dominated by the coccolithophores Emiliania huxleyi and Gephyrocapsa oceanica [Conte et al., 1994; Volkman et al., 1995]. Since then, alkenones were detected in coastal [Mercer et al., 2005; Salacup et al., in prep-b] and lacustrine [D'Andrea and Huang, 2005; Theissen et al., 2005; Toney, J. L. et al., 2010; Zink et al., 2001] settings. Previous investigations of Uk’37 as a temperature proxy in an estuary (Chesapeake Bay) were apparently hampered by the effects of low salinity (≤20psu) on the composition the alkenone producing population [Mercer et al., 2005]. Because different species can have different relationships between Uk’37 and SST 84 [Versteegh et al., 2001], the alkenone SST proxy functioned poorly. Conversely, Salacup et al. [in prep-b] recently showed that 1) alkenone production in Narragansett Bay bears all the hallmarks of production by well-understood open-ocean species E. huxleyi and G. oceanica, 2) the relationship between Narragansett Bay alkenones and SST follows the open-ocean calibration of Prahl et al. [1988] (Uk’37 = 0.034(T) + 0.039), and 3) Uk’37 closely matches mean annual SST in Narragansett Bay with a small and consistent positive offset (1.8°C). The nominal calibration error of Uk’37 is ±1.4°C [Prahl et al., 1988], however the acquisition and comparison of multiple cores indicated that Uk’37- inferred SST variability over the past 300 years in NB is reproducible to within ±0.3°C [Salacup et al., in prep-b], just greater than our analytical error (±0.2°C). This is the so- called ‘reconstruction error’. The Uk’37 reconstruction presented here is based on three geographically separated records of Narragansett Bay SST (Fig. 3.3), stacked as follows. The mean SST of each record, calculated for the period during which all three records overlap (1150-present), was subtracted from that record to produce an SST anomaly for that core. The resulting time series was then divided by the standard deviation for that record so that reconstructions with stronger variability were not given undue power over the stack. The individual records were then combined to produce the stack (Fig. 3.4) with a resolution between 0.01 and 38 (avg = 4) years (Fig. 3.3). The nominal Uk’37 calibration error of ± 1.4°C is approximately the size of Common Era SST variability [Jones and Mann, 2004; Mann et al., 2009], complicating the interpretation of Uk’37 derived SST estimates during this period. The calibration assumes core top sediments used in the calibration have zero age and that the ocean atlases used accurately reflect instrumental mean annual SST. This implicitly assumes that core-top Uk’37 ratios everywhere actually reflect mean annual SST, which is not always the case [Prahl et al., 2010; this work]. Therefore, the nominal error is a worst case 85 scenario estimate in globally distributed surface sediments, of uncertain age, underlying waters hosting a wide range of alkenone producing populations, and thus may be unrealistic in a geologic context when many of the above assumptions are avoided. In order to approximate our reconstruction error we evaluate the average first difference (taken as an absolute value) between adjacent SST determinations in each core and in the stack. This should be ≥1.4°C if a random 1.4°C calibration error is affecting our SST determinations. We also use apparent SST offsets between individual records (Fig. 3.3), emplaced by modern in situ SST gradients, to evaluate the reproducibility of SST trends between separate cores. Again, the error on these offsets should be ≥ 1.4°C if a random 1.4°C calibration error is acting on our SST determinations. The average first differences calculated for Greenwich Bay, Potter’s Cove, and Fox Island are 0.4, 0.4, and 0.5°C, respectively, only slightly larger than our analytical error (±0.2°C). The average first difference on our stack is 1.0°C, but this estimate is pessimistic as it includes error imparted by the apparent SST offsets between core sites (Fig. 3.3, 3.4), because we only standardized over the period of overlap (1100-present). For example, the mean of the anomalies for the longer Greenwich Bay and Fox Island cores are ~0.3°C warmer than that for the shorter Potter’s Cove suggesting that the real average first difference for the stack is closer to 0.7°C. The stack-based estimate is also impacted by error in the different age models as they may lead to real Bay-wide positive and negative SST anomalies not necessarily lining up in time. The persistence of the near constant offset thru time (0.9 ± 0.6°C) between Greenwich Bay and Fox Island (the longest two records) highlights the reproducibility and precision of the reconstructions. Together, the first difference and offset-based results are much tighter than would be allowed by a random calibration error of ± 1.4°C. Instead, our results suggest a Narragansett Bay specific Uk’37 reconstruction error closer 86 to 0.5°C (the average of the above analyses). This estimate is still pessimistic as it is likely that some of the SST change between adjacent points is real SST variability and that stack-based estimates are impacted by errors associated with SST offsets and differing age models. The reconstruction error is much better than the nominal open- ocean 1.4°C calibration error, implying complicating factors such as the composition of the alkenone producing population, differing calibrations, and production seasonality do not corrupt the Uk’37 proxy in the Bay over the past 1500 years. The reproducibility of this Common-era alkenone record, based on three core sites, stands in contrast to records from other archives. For example, warm-water corals also require replication to improve the robustness of their chronologies and their Sr/Ca SST estimates, which may vary by more than 2°C in response to changing coral architecture and sampling strategy [DeLong et al., 2013]. Kinetic and vital effects often confound environmental data contained in cold-water corals [Marali et al., 2013]. The many methods of standardizing (detrending) and calibrating tree ring width and density data all have their own list of caveats and by their function may remove important structure from the climate reconstruction, particularly in the multi-decadal to centennial frequencies [Esper et al., 2004]. Further, the construction of tree ring paleoclimatologies often requires the coring of multiple holes per tree, of multiple trees per site, and of multiple sites per region to form a network in which signal-to-noise is maximized [Hughes, 2010]. In contrast, because of the relatively broad temperature structure of the ocean [Deser et al., 2010], robust regional Uk’37 SST reconstructions may be achieved with far fewer cores. Narragansett Bay SST (as inferred by our stack) was relatively warm between 720 to 1140 (avg = +0.8°C). This warmth was terminated by cooling between 1070 and 1150, coincident with the Oort solar minimum (Fig. 3.4). Another relatively warm interval between 1220 and 1640 (avg = +0.4°C) was punctuated by several cool periods 87 between 1260 and 1360, circa 1480, and between 1560 and 1600, coincident with the Wolf and Spörer Minima. After ~1620 SST cooled rapidly by ~2°C until ~1845, coincident with the Maunder and Dalton Minima. After ~1850, Narragansett Bay SST warmed rapidly to + 0.6°C at a rate of ~1.9°C/100 years (±0.2°C; between 1900 and 2000), almost three times the global mean rate (0.74°C) [IPCC, 2007]. This is consistent with rates based on a long regional instrumental record and is likely due to the southward advection of polar sourced waters bearing the imprint of the polar amplification of global warming [Shearman and Lentz, 2010]. Narragansett Bay SST variability bears resemblance to other records of North Atlantic climate over the past 1500 years (Fig. 3.1, 3.4), although leads and lags are noted. For the basis of comparison, the MCA at a given site is defined here as the multi- centennial period nearest the ‘canonical MCA’ [950-1250; Mann et al., 2009; Moberg et al., 2005] during which SST for a given site is above the sites’ 2000-700 mean SST. The beginning of relative warmth detected in Narragansett Bay between 720 and 1120 leads MCA warming in Greenland by ~180 years [Kobashi et al., 2011], in Iceland by ~260 years [Ran et al., 2010; Sicre et al., 2008], and on the Voring Plateau by ~430 years [Andersson et al., 2003]. This lead is well outside our age model error (Fig. 3.2). MCA warming in Narragansett Bay may lag that in Chesapeake Bay [Cronin et al., 2010] and the Gulf of Mexico [Richey et al., 2007] by ~70 and 150 years, respectively, but the short length of our record and age model error makes this claim speculative. Taken together, these results imply a northward migration of oceanic heat anomalies from the tropics to the Arctic over a span of ~600 years (Fig. 3.4). The several cool periods detected between 1070 and 1850 reflect the climate deterioration of the LIA [Bradley and Jones, 1992] coincident with the Wolf, Spörer, Maunder, and Dalton Minimum sunspot events. The timing of minimum Narragansett Bay SST at ~1850 (Fig. 3.4) coincides closely with minimum temperatures in Greenland 88 [Kobashi et al., 2011], Iceland [Ran et al., 2010; Sicre et al., 2008], the Voring Plateau [Andersson et al., 2003], Chesapeake Bay [Cronin et al., 2010] and the Gulf Of Mexico [Richey et al., 2007]. This cold diachronous spatio-temporal pattern stands in contrast to the asynchronous warm ‘evolutive’ pattern described for the MCA, hinting at separate forcings or mechanisms. The change in SST (ΔSST) between the MCA (720-1120) and LIA (1450-1850) in Narragansett Bay is ~1.2°C consistent with results based on a global climate proxy network [0.6 - 1.8°C ; Mann et al., 2009] even though their definitions of the MCA (950-1250) and LIA (1400-1700) are different than ours. As noted earlier, atmospheric AO/NAO dynamics can influence North Atlantic climate by changing subpolar and subtropical gyre intensity (and thus Labrador Current, Gulf Stream, and North Atlantic Current intensity) [Curry and McCartney, 2001], and Northwest Atlantic SSTs [MERCINA, 2001; Pickart et al., 1999]. In the open-ocean, Curry et al. [2001] found that Gulf Stream and North Atlantic Current intensity varies positively with the NAO (by as much as 25-33%) due to latitudinal shifts in the westerlies and the spin up of the sub-tropical gyre (Fig. 3.1). Closer to the coast, Pickart et al. [1999] suggest a coupled slope water system (CSWS, Fig. 3.1 right) in which warm salty primarily Gulf Stream-sourced Atlantic Temperate Slope Water (ATSW) competes with cold and relatively fresh Labrador Current-sourced Labrador Subarctic Slope Water (LSSW), for position on the continental slope. Pickart et al. [1999] proposed that contributions of warm ATSW increased in response to positive NAO events (and vice versa) resulting in a positive relationship between the NAO and Gulf of Maine SSTs on annual to multi-decadal timescales. Comparison of a nearby (Woods Hole, M.A.) instrumental SST record [Shearman and Lentz, 2010] with instrumental NAO measurements (provided by the Climate Analysis Section, NCAR, Boulder, USA) shows a significant positive relationship between southern New England SST and the NAO (p<0.01; Fig. 3.5). It is likely that much of the decadal-scale variability in our Uk’37-inferred 89 NB SST record can be attributed to NAO variability. However, the resolution of our time series, and the quality of our age control, does not allow us to unambiguously assign this variability to the NAO. Comparison of our Uk’37 Narragansett Bay SST record with a proxy-based AO reconstruction [D' Arrigo et al., 2003], which displays stronger low- frequency variability than the NAO, also highlights a significant positive relationship (p<0.0001; Fig. 3.5). Here, our sampling resolution and age model error does allow us to more confidently link low-frequency (~100 year) Narragansett Bay SST variability to an atmospheric anomaly, in this case the AO, via the mechanisms discussed above. Previous work suggests the MCA was dominated by a mean AO/NAO positive phase [Mann et al., 2009; Trouet et al., 2012]. If much of the SST structure reconstructed for Narragansett Bay during the MCA is driven by AO/NAO dynamics than we would predict the winter climate of southern New England was warm and wet because of the proximity of warm Gulf Stream, the increased contribution of warm ATSW on the shelf, and an intensified northward oceanic heat transport. Conversely, proxy and model-based evidence suggests the AO/NAO declined into a generally negative state in concert with climate deterioration into the LIA [Mann et al., 2009; Shindell et al., 2001; Trouet et al., 2012]. Again, if Narragansett Bay SST structure during the LIA is driven by AO/NAO dynamics we would predict cold and dry winters due to the southward migration of the Gulf Stream, an increasing influence of northern sourced LSSW, and decreased northward oceanic heat transport. This precipitation pattern, wet winters during the MCA and dry winters during the LIA, is consistent with results from peat bogs in Ireland [Charman et al., 2012], which should have the same climatological response the NAO variability as southern New England, suggesting a primary control of AO/NAO variability on mean annual temperature and winter precipitation. 90 Lastly, given their proximity, the near anti-phased relationship between sea surface temperatures in Narragansett Bay and Chesapeake Bay is interesting (Fig. 3.4). Such anti-phasing has been reported previously, albeit on shorter timescales. For example, over the past 150 years, SSTs north and south of Cape Hatteras warmed and cooled, respectively [Shearman and Lentz, 2010]. Anti-phasing is also evident between southern New England [Woods Hole, MA; Shearman and Lentz, 2010] and Chesapeake Bay in comparisons of more recent instrumental SST data (Fig. 3.5). Our results imply the existence of a ‘north-south’ component to the previously detected “east-to-west” North Atlantic provincialism [Cronin et al., 2010, and refernces therein], in which SST change (for example) is detected in the southwestern Atlantic Ocean before the northwestern. It may be that while temperate and higher latitude climate is heavily influenced by AO/NAO dynamics [Charman et al., 2012; Hurrell, 1995; MERCINA, 2001; Trouet et al., 2009; this work], tropical Atlantic climate is effectively blind to these atmospheric anomalies on the centennial timescale, perhaps because of tropical ocean thermostat effects [Clement et al., 1996]. However, based on detailed historical coastal SST data and a simple advective model, Shearman and Lentz [2010] proposed that southward migrating Labrador Sea SST anomalies would stop at or before Cape Hatteras and thus locations proximal or to the south of this landmark would not record this high latitude forced variability. The former ‘signal attenuation’ model does not need to invoke the tropical ocean thermostat. Indeed, the ΔSST at Narragansett Bay between the MCA and LIA (~1.5°C) is at the lower (and southern) end of a southward progression of decreasing ΔSST which is ~3.5°C at the Voring Plateau, ~2.0°C on the Icelandic coast and over Greenland (air temperature), consistent with the attenuation of a northern sourced SST anomaly. 91 3.5 Conclusions Three records of Common Era SST inferred from the Uk’37 proxy in Narragansett Bay display remarkable downcore similarity. An in situ SST gradient between the core sites is reflected in a persistent offset in Uk’37 SST between the cores. The relatively small number of cores required to produce a robust reconstruction for Narragansett Bay stands in sharp contrast to other high resolution paleoclimate proxies. Our reconstruction shows the local expression of the MCA took place between the years ~720 and 1120 before cooling into the LIA. The coldest time of the past ~1500 years took place in the mid-1800s before warming at a rate approximately thrice the North Hemisphere average. The timing of the MCA and LIA in Narragansett Bay compared to other North Atlantic sites highlights the idea that these two long climate phases evolved (spatially and temporally) in different ways, implying separate forcings and/or mechanisms. Strong positive relationships between the state of the AO/NAO and Narragansett Bay SST are consistent with other mid-to-high latitude evidence that the MCA and LIA reflected periods of prolonged AO/NAO positive and negative phases, respectively. Anti-phasing of Narragansett Bay and Chesapeake Bay and Gulf of Mexico SSTs suggests a N-S disconnect in which the more southern sites are not impacted, at least not in the same way, by AO/NAO dynamics. 92 3.6 Tables Table 3.1. Age-depth constraints used for NB cores Depth (cm) Age (cal. bp) Year (AD) Method NB 12/45v 0 -58 2008 core top 3 -52 2002 Pb-210 8 -41 1991 Pb-210 13 -27 1977 Pb-210 18 -1 1951 Pb-210 23 21 1929 Pb-210 28 33 1883 Pb-210 35.5 75 1850 Pb-210 35 75 1850 Industrial [Pb] 55 167 (±19) 1783 14 C foram 70 180 (±41) 1770 14 C foram 77 250 1700 pollen 14 183 977 (±80) 973 C mollusk 14 251 5513 (±73) -3563 C mollusk NB25/42v 0 -58 2008 core top 2 -53 2003 Pb-210 5 -46 1996 Pb-210 8 -38 1988 Pb-210 11 -30 1980 Pb-210 14 -23 1973 Pb-210 17 -17 1967 Pb-210 22 -13 1963 Pb-210 27 -9 1959 Pb-210 30 -6 1956 Pb-210 33 9 1941 Pb-210 36 14 1936 Pb-210 43 75 1850 Industrial [Pb] 63 250 1700 pollen 140 511 (±34) 1439 14 C foram 160 937 (±32) 1013 14 C foram 180 1112 (±62) 838 14 C foram 198 2023 (±98) -73 14 C foram 200 2589 (±153) -639 14 C mollusk 227 2610 (±150) -660 14 C mollusk 280 3451 (±108) -1501 14 C mollusk 93 Table 3.1 (cont.). Age-depth constraints used for NB cores Depth (cm) Age (cal. bp) Year (AD) Method NB44v 0 -60 2010 core-top 2 -52 2002 Pb-210 4 -39 1989 Pb-210 7 -21 1971 Pb-210 14 12 50 (±5) 1900 C foram 14 20 26 (±19) 1924 C foram 31 100 1850 Industrial [Pb] 14 40 732 (±66) 1218 C foram 14 50 506 (±24) 1444 C foram 14 70 609 (±59) 1341 C foram 14 110 1039 (±25) 911 C foram 94 3.7 Figures Figure 3.1. Left (left) Major surface currents of the North Atlantic Ocean: GS = Gulf Stream, AC = Azores Current, NAC = North Atlantic Current, LC = Labrador Current, IC= Irminger Current, EG = East Greenland Current, WG = West Greenland Current, LC = Labrador Current, CSWS = Coupled Slope Water System. White circles and diamonds show the locations of records discussed in the text. The white square shows the location of the coarsely resolved (~60 year spacing) Sargasso Sea record of Keigwin [1996] included in the IPCC compilation but not discussed here. Together, the circles (excluding Greenland and Icelend) and square reflect the totality, and paucity, of SST records at least 1000 years long that are included in the IPCC Common Era reconstruction. (inset) Location of Narragansett Bay cores (white circles). (right) Coupled Slope Water System [Pickart et al., 1999]: ATSW = Atlantic Temperate Sea Water, LSSW = Labrador Sub- Arctic Sea Water. Dashed lines and arrows depict the movement of the boundary between LSSW and ATSW in relation to changes in the AO/NAO. 95 Figure 3.2. (top) Age models for cores NB25/42 (left), NB12/45 (middle), NB44 (right) were established using 14C of mollusks and foraminifera (filled squares), pollen horizons (triangles), [Pb] increases (open squares), and 210Pb (circles). Black and red lines are the mean and 2σ error age determinations, respectively. (bottom) The evolution of sedimentation rate (black) and 2σ age error (red) through time. 96 Figure 3.3. Plots comparing our raw (top) SST reconstructions for cores NB25/42 (blue diamonds), NB12/45 (red squares), and NB 44v (green triangles). The individual NB SST records are offset by an amount consistent with in situ SST gradients. Sample spacing of the resulting SST stack averaged 4 years (dashed line). 97 Figure 3.4. Narragansett Bay SST reconstructions compared with other North Atlantic records (Fig. 3.1). Arranged from north to south are: Voring Plateau August SST [Anderssen et al. 2003], coastal Iceland warm season SST [Sicre et al., 2011], Greenland air temperature [Kobashi et al., 2011], Narragansett Bay [this work], Chesapeake Bay warm season SST [Cronin et al., 2010], and Gulf of Mexico mean annual SST [Richey et al., 2007]. The horizontal red bars and vertical blue arrows point out times corresponding to the Medieval Climate Anomaly and the maximum Little Ice Age cooling, respectively. Recent solar minima are shown with open rectangles at the top of the plot. 98 Figure 3.5. (A) Comparison of instrumental Woods Hole SST [top; Shearman and Lentz, 2010] with the North Atlantic Oscillation [middle; NAO; provided by the Climate Analysis Section, NCAR, Boulder, USA], and instrumental Chesapeake Bay SST [lower; Cronin personal comm.]. A significant (p<0.01) and positive relationship between the NAO and Woods Hole SST is evident over the past 50 years, consistent with findings from the nearby Gulf of Maine [Pickart et al., 1999]. SST at Woods Hole has warmed ~0.47°C since 1950 while Chesapeake Bay SST has cooled 0.75°C. (B) Comparison of Uk’37- inferred Narragansett Bay SST with a tree ring based reconstruction of the Arctic Oscillation [D’Arrigo et al., 2003] shows a significant (p<0.0001) and positive relationship between the two over the past ~350 years, consistent with NAO-based findings and suggesting high latitude atmospheric dynamics like the AO and NAO impact regional oceanography. 99 Ahmed, M., et al. (2013), Continental-scale temperature variability during the past two millennia, Nature Geoscience, 6(5), 339-346. Andersson, C., B. Risebrobakken, E. Jansen, and S. O. Dahl (2003), Late Holocene surface ocean conditions of the Norwegian Sea (Voring Plateau), Paleoceanography, 18(2). Booth, B. B. B., N. J. Dunstone, P. R. Halloran, T. Andrews, and N. Bellouin (2012), Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability (vol 484, pg 228, 2012), Nature, 485(7399), 534-534. Borkman, D. G., and T. Smayda (2009), Multidecadal (1959-1997) changes in Skeletonema abundance and seasonal bloom patterns in Narragansett Bay, Rhode Island, USA, J. Sea Res., 61(1-2), 84-94. Bower, A., S. Lozier, and S. Gary (2011), Export of Labrador Sea Water from the subpolar North Atlantic: A Lagrangian perspective, Deep-Sea Res Pt Ii, 58(17-18), 1798- 1818. Braconnot, P., S. P. Harrison, M. Kageyama, P. J. Bartlein, V. Masson-Delmotte, A. Abe-Ouchi, B. Otto-Bliesner, and Y. Zhao (2012), Evaluation of climate models using palaeoclimatic data, Nature Climate Change, 2(6), 417-424. Bradley, R. S. (1999), Paleoclimatology: Reconstructing Climate of the Quaternary, Second ed., 610 pp., Academic Press, San Diego, CA. Bradley, R. S., and P. D. Jones (1992), "When was the Little Ice Age?", in Proceedings of the International Symposium on the "Little Ice Age" Cliamte, edited by T. Mikami, Department of Geography, Tokyo Metropolitan University. Brassell, S. C., G. Eglinton, I. T. Marlowe, U. Pflaumann, and M. Sarnthein (1986), Molecular Stratigraphy - a New Tool for Climatic Assessment, Nature, 320(6058), 129- 133. Bricker, S. B. (1993), The History of Cu, Pb, and Zn Inputs to Narragansett Bay, Rhode- Island as Recorded by Salt-Marsh Sediments, Estuaries, 16(3B), 589-607. Charman, D. J., V. Hohl, A. Blundell, F. Mitchell, J. Newberry, and P. Oksanen (2012), A 1000-year reconstruction of summer precipitation from Ireland: Calibration of a peat- based palaeoclimate record, Quatern Int, 268, 87-97. Cionco, R. G., and R. H. Compagnucci (2012), Dynamical characterization of the last prolonged solar minima, Advances in Space Research, 50(10), 1434-1444. Conte, M. H., A. Thompson, and G. Eglinton (1994), Primary production of lipid biomarker compounds by Emiliania huxleyi: results from an experimental mesocosm study in fjords of southern Norway, Sarsia, 79, 319-332. 100 Cook, E. R., R. D. D'Arrigo, and M. E. Mann (2002), A Well-Verified, Multiproxy Reconstruction of the Winter North Atlantic Oscillation Index since a.d. 1400*, J Climate, 15(13), 1754-1764. Corbin, J. (1989), Recent and historical accumulation of trace metal contaminants in Narragansett Bay sediments, 295 pp, University of Rhode Island, Kingston, RI. Cronin, T. M., K. Hayo, R. C. Thunell, G. S. Dwyer, C. Saenger, and D. A. Willard (2010), The Medieval Climate Anomaly and Little Ice Age in Chesapeake Bay and the North Atlantic Ocean, Palaeogeography, Palaeoclimatology, Palaeoecology, 297(2), 299-310. Curry, R. G., and M. S. McCartney (2001), Ocean gyre circulation changes associated with the North Atlantic Oscillation, Journal of Physical Oceanography, 31(12), 3374- 3400. D' Arrigo, R. D., E. R. Cook, M. E. Mann, and G. C. Jacoby (2003), Tree-ring reconstructions of temperature and sea-level pressure variability associated with the warm-season Arctic Oscillation since AD 1650, Geophysical Research Letters, 30(11). D'Andrea, W. J., and Y. Huang (2005), Long chain alkenones in Greenland lake sediments: Low delta C-13 values and exceptional abundance, Organic Geochemistry, 36(9), 1234-1241. DeLong, K. L., T. M. Quinn, F. W. Taylor, K. Lin, and C. C. Shen (2012), Sea surface temperature variability in the southwest tropical Pacific since AD 1649, Nature Climate Change, 2(11), 799-804. DeLong, K. L., T. M. Quinn, F. W. Taylor, C. C. Shen, and K. Lin (2013), Improving coral-base paleoclimate reconstructions by replicating 350 years of coral Sr/Ca variations, Palaeogeogr Palaeocl, 373, 6-24. Deser, C., M. A. Alexander, S. P. Xie, and A. S. Phillips (2010), Sea Surface Temperature Variability: Patterns and Mechanisms, in Annual Review of Marine Science, edited, pp. 115-143. Dickson, R., J. Lazier, J. Meincke, P. Rhines, and J. Swift (1996), Long-term coordinated changes in the convective activity of the North Atlantic, Prog Oceanogr, 38(3), 241-295. Dissard, D., G. Nehrke, G. J. Reichart, and J. Bijma (2010), The impact of salinity on the Mg/Ca and Sr/Ca ratio in the benthic foraminifera Ammonia tepida: Results from culture experiments, Geochimica et Cosmochimica Acta, 74(3), 928-940. Enfield, D. B., A. M. Mestas-Nunez, and P. J. Trimble (2001), The Atlantic multidecadal oscillation and its relation to rainfall and river flows in the continental US, Geophysical Research Letters, 28(10), 2077-2080. Esper, J., D. C. Frank, and R. J. S. Wilson (2004), Climate reconstructions: Low- frequency ambition and high-frequency ratification, Eos, Transactions American Geophysical Union, 85(12), 113-120. 101 Gray, S. T., L. J. Graumlich, J. L. Betancourt, and G. T. Pederson (2004), A tree-ring based reconstruction of the Atlantic Multidecadal Oscillation since 1567 AD, Geophysical Research Letters, 31(12), -. Heegaard, E., H. J. B. Birks, and R. J. Telford (2005), Relationships between calibrated ages and depth in stratigraphical sequences: an estimation procedure by mixed-effect regression, Holocene, 15(4), 612-618. Herbert, T. D. (2003), Alkenone paleotemperature determinations, in Treatise in Marine Geochemistry, edited by H. Elderfield, pp. 391-432, Elsevier, Amsterdam. Hughes, M. K. (2010), Dendroclimatology in High-Resolution Paleoclimatology, in Dendroclimatology: Progress and Prospects, edited by M. K. Hughes, Swetnam, W., and Diaz, H.F., Springer, London. Hurrell, J. W. (1995), DECADAL TRENDS IN THE NORTH-ATLANTIC OSCILLATION - REGIONAL TEMPERATURES AND PRECIPITATION, Science, 269(5224), 676-679. IPCC (2007), Summary for Policymakers, in Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, edited by S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Jones, P. D., and M. E. Mann (2004), Climate over past millennia, Rev Geophys, 42(2), RG2002. Keigwin, L. D. (1996), The Little Ice Age and Medieval warm period in the Sargasso Sea, Science, 274(5292), 1504-1508. Kerr, R. A. (2000), A North Atlantic climate pacemaker for the centuries, Science, 288(5473), 1984-1986. Kobashi, T., K. Kawamura, J. P. Severinghaus, J. M. Barnola, T. Nakaegawa, B. M. Vinther, S. J. Johnsen, and J. E. Box (2011), High variability of Greenland surface temperature over the past 4000 years estimated from trapped air in an ice core, Geophysical Research Letters, 38. Levitus, S. (1982), Climatological Atlas of the World Ocean, 173 pp., U.S. Government Printing Office, Washington D.C. Levitus, S., J. Antonov, and T. Boyer (2005), Warming of the world ocean, 1955-2003, Geophysical Research Letters, 32(2). Mann, M. E., Z. Zhang, S. Rutherford, R. S. Bradley, M. K. Hughes, D. Shindell, C. Ammann, G. Faluvegi, and F. Ni (2009), Global signatures and dynamical origins of the Little Ice Age and Medieval Climate Anomaly, Science, 326(5957), 1256-1260. Marali, S., M. Wisshak, M. L. Correa, and A. Freiwald (2013), Skeletal microstructure and stable isotope signature of three bathyal solitary cold-water corals from the Azores, Palaeogeogr Palaeocl, 373, 25-38. 102 McNeely, R., A. S. Dyke, and J. R. Southon (2006), Canadian marine reservoir ages, preliminary data assessment, Open File 5049, pp. 3. Mercer, J., M. Zhao, and S. Colman (2005), Seasonal variations of alkenones and Uk37 in the Chesapeake Bay water column, Estuarine, Coastal and Shelf Science, 63(4), 675- 682. MERCINA (2001), Oceanographic Responses to Climate in the Northwest Atlantic, Oceanography, 14, 76-82. Nixon, S. W. (1991), A History of Metal Inputs to Narragansett Bay Rep., 69 pp, Narragansett Bay Project, Providence, RI. Parshall, T., D. R. Foster, E. Faison, D. MacDonald, and B. C. S. Hansen (2003), Long- term history of vegetation and fire in pitch pine-oak forests on Cape Cod, Massachusetts, Ecology, 84(3), 736-748. Pickart, R. S., T. K. McKee, D. J. Torres, and S. A. Harrington (1999), Mean structure and interannual variability of the slopewater system south of Newfoundland, Journal of Physical Oceanography, 29(10), 2541-2558. Prahl, F. G., L. A. Muehlhausen, and D. L. Zahnle (1988), Further evaluation of long- chain alkenones as indicators of paleoceanographic conditions, Geochimica et Cosmochimica Acta, 52(9), 2303-2310. Ran, L., H. C. Jiang, K. L. Knudsen, and J. Eiriksson (2010), Diatom-based reconstruction of palaeoceanographic changes on the North Icelandic shelf during the last millennium, Palaeogeogr Palaeocl, 302(1-2), 109-119. Richey, J. N., R. Z. Poore, B. P. Flower, and T. Quinn (2007), 1400 yr multiproxy record of climate variability from the northern Gulf of Mexico, Geology, 35(5), 423-426. Salacup, J., J. Farmer, T. D. Herbert, and W. L. Prell (in prep), Alkenone production with open ocean characteristics in a near-shore setting: Narragansett Bay, R.I. Sejrup, H. P., H. Haflidason, and J. T. Andrews (2011), A Holocene North Atlantic SST record and regional climate variability, Quaternary Science Reviews, 30(21-22), 3181- 3195. Shearman, R. K., and S. J. Lentz (2010), Long-Term Sea Surface Temperature Variability along the U.S. East Coast, Journal of Physical Oceanography, 40(5), 1004- 1017. Sicre, M. A., J. Jacob, U. Ezat, S. Rousse, C. Kissel, P. Yiou, J. Eiriksson, K. L. Knudsen, E. Jansen, and J. L. Turon (2008), Decadal variability of sea surface temperatures off North Iceland over the last 2000 years, Earth and Planetary Science Letters, 268(1-2), 137-142. Taylor, A. H., and J. A. Stepens (1998), The North Atlantic oscillation and the latitude of the Gulf Stream, Tellus A, 50(1), 134-142. 103 Theissen, K. M., D. A. Zinniker, J. M. Moldowan, R. B. Dunbar, and H. D. Rowe (2005), Pronounced occurrence of long-chain alkenones and dinosterol in a 25,000-year lipid molecular fossil record from Lake Titicaca, South America, Geochimica et Cosmochimica Acta, 69(3), 623-636. Thompson, D. W. J., and J. M. Wallace (1998), The Arctic Oscillation signature in the wintertime geopotential height and temperature fields, Geophysical Research Letters, 25(9), 1297-1300. Ting, M. F., Y. Kushnir, R. Seager, and C. H. Li (2009), Forced and Internal Twentieth- Century SST Trends in the North Atlantic, J Climate, 22(6), 1469-1481. Toney, J. L., Y. S. Huang, S. C. Fritz, P. A. Baker, E. Grimm, and P. Nyren (2010), Climatic and environmental controls on the occurrence and distributions of long chain alkenones in lakes of the interior United States, Geochimica et Cosmochimica Acta, 74(5), 1563-1578. Toole, J. M., R. G. Curry, T. M. Joyce, M. McCartney, and B. Pena-Molino (2011), Transport of the North Atlantic Deep Western Boundary Current about 39 degrees N, 70 degrees W: 2004-2008, Deep-Sea Res Pt Ii, 58(17-18), 1768-1780. Trouet, V., J. Esper, N. E. Graham, A. Baker, J. D. Scourse, and D. C. Frank (2009), Persistent Positive North Atlantic Oscillation Mode Dominated the Medieval Climate Anomaly, Science, 324(5923), 78-80. Versteegh, G. J. M., R. Riegman, J. W. de Leeuw, and J. H. F. Jansen (2001), U(37)(K ')values for Isochrysis galbana as a function of culture temperature, light intensity and nutrient concentrations, Organic Geochemistry, 32(6), 785-794. Visbeck, M. H., J. W. Hurrell, L. Polvani, and H. M. Cullen (2001), The North Atlantic Oscillation: past, present, and future, Proc Natl Acad Sci U S A, 98(23), 12876-12877. Volkman, J. K., S. M. Barrett, S. I. Blackburn, and E. L. Sikes (1995), Alkenones in Gephyrocapsa-Oceanica - Implications for Studies of Paleoclimate, Geochimica et Cosmochimica Acta, 59(3), 513-520. Wanamaker, A. D., K. J. Kreutz, B. R. Schone, and D. S. Introne (2011), Gulf of Maine shells reveal changes in seawater temperature seasonality during the Medieval Climate Anomaly and the Little Ice Age, Palaeogeogr Palaeocl, 302(1-2), 43-51. Zink, K. G., D. Leythaeuser, M. Melkonian, and L. Schwark (2001), Temperature dependency of long-chain alkenone distributions in Recent to fossil limnic sediments and in lake waters, Geochimica et Cosmochimica Acta, 65(2), 253-265. 104 CHAPTER FOUR _______________________________________________________ SEVEN HUNDRED YEAR RECORD OF LAND CLEARANCE, NUTRIENT MOBILIZATION, AND PRODUCTIVITY IN NARRAGANSETT BAY, RI Salacup, J.M.1, Altabet, M.2 Herbert, T.D.1, and Prell, W.L.1 1 Brown University, Department of Geosciences 2 University of Massachusetts, School for Marine Science and Technology 105 4.1 Abstract Watersheds around the world have been the scene of rapid population growth, wide spread resource exploitation, and environmental degradation over the past ~150 years. Processes like land clearance, soil mobilization, and nutrient loading can stress an ecosystem leading to shifts in benthic and pelagic communities. In order to properly characterize the ecosystem response to documented historical landscape disturbance it must be contextualized within a longer time period. Here, we analyze sediment cores from three locations in Narragansett Bay using branched glycerol-dialkyl glygerol- tetraethers (brGDGTs), isotopes of bulk sedimentary nitrogen (δ15N), pollen of the invasive weed Ambrosia, and the abundance of benthic foraminifera to investigate the ecosystem’s response to land clearance, soil mobilization, and nutrient loading in the Bay’s watershed over the past 700 years. Changing patterns in the contributions of the different GDGT groups (Ia, IIa, IIIa) suggest environmental reconstructions based on brGDGT concentrations are confounded by in situ production. Otherwise, modern north-south gradients in fresh water, sediment load, and nutrients have been a persistent, although changing, feature of the Bay over the past 700 years. Down core δ15N and Ambrosia pollen concentrations reveal the balance between marine and terrestrial organic matter sources to the Bay was interrupted three times: first in the 1600s by the addition of terrestrial nutrients after the extirpation of the native beaver population, second in the 1700s by the wide-spread land clearance practiced by the Colonists and, third in the mid-1800s by a massive increase in nutrient flux associated with Industrialization and Urbanization. The nutrient N load added to the Bay concurrent with these events directly affected the Bay’s ecosystem leading to increases in benthic productivity. Our results imply that humans began to influence the Bay’s ecosystem as soon as the early 1600s, but that the environmental 106 degradation that occurred over the past 150 years is unprecedented in Narragansett Bay over at least that past 700 years. 4.2 Introduction Narragansett Bay (Fig. 4.1) was first visited in 1524 by the Florentine explorer Giovanni da Verrazzano. However, not until the mid-to-late-1600s did European explorers and settlers come to populate and later dominate the watershed’s landscape. In 1793, Slater Mill, located on the Blackstone River at the head of the Bay (Fig. 4.1), was the first mill in the United States to spin yarn using water power, making it the birthplace of the American Industrial Revolution. The ensuing industrialization and urbanization drove the human population of the Bay’s watershed up from ~50,000 in 1790 to more than 2 million by the year 2000 [Hamburg et al., 2008]. The past ~300 years was a period of progressive and widespread resource exploitation, ecosystem degradation, and habitat destruction. Resource exploitation and land use change may disturb a coastal water mass in many ways. For example, since the onset of major urban development, harmful concentrations of nutrients have been delivered to coastal waters associated with eroded soils, excess fertilizer application, and human waste from sewage treatment plants and coastally located septic systems [Nixon et al., 2008]. Over the past decade, these excess nutrients have led to eutrophication and episodic oxygen depletion (hypoxia) occurring, in general, from late-June through August [Codiga et al., 2009; Deacutis, 2008; Saarman et al., 2008]. In 2003, a hypoxic event was responsible for killing more than a million fish [RIDEM, 2003]. The spatial and temporal extent of hypoxic events correlate well with spatial and temporal patterns of nutrients and productivity (Fig. 4.2) [Murray et al., 2007; Oviatt, 2008] suggesting that a human-induced five-fold and two-fold increase in nutrient nitrogen and phosphorus delivery to the Bay, respectively, are responsible for eutrophication-induced hypoxia which have documented impacts on 107 the ecosystem [Deacutis, 2008; Nixon et al., 2008; RIDEM, 2003], In the deeper past, there is growing evidence for concentrated Native American population centers [multiple 100-150 dwelling villages; Waller, 2000] along the Bay’s western shores, from Greenwich Bay south to Narragansett Country. Clearance of forest to accommodate these villages [Bernstein, 1993] and the later expansion of the Colonies [Parshall et al., 2003], likely disturbed ancient soils resulting in the delivery of this material, and associated nutrients, into local rivers emptying ultimately into the Bay. Did these ‘new’ nutrients affect the ecosystem of the Bay? Understanding the environmental evolution of the Bay, and its responses to past disturbance, is directly relevant to understanding the sensitivity of the Bay to future changes. Unfortunately, using the instrumental record alone, we are unable to answer these questions, or characterize recent (last 150 years) ecological disturbance within a long-term context. In order to develop a better understanding of the effects of land use change on sediment and nutrient delivery, and, in turn, on ecosystem productivity, we apply a suite of proxies sensitive to environmental, biogeochemical, and ecological change to sediment cores from Narragansett Bay. Downcore variability in these proxies is interpreted within a context provided by a survey of these same proxies in Narragansett Bay surface sediments as well as comparison with downcore elemental data sensitive to changes in terrestrial input (Fe, Ti, magnetic susceptibility). 4.2.1 Geological Setting Narragansett Bay (Fig. 4.1) is a high-salinity, relatively well-mixed, north-south oriented estuary located predominately in RI, USA, although 60% of its watershed lies in Massachusetts. Fresh water and suspended sediment enters the system from the north via the Blackstone, Pawtuxet, and Taunton River systems. Overall, the Bay displays estuarine and cyclonic non-tidal flow. Gradients of sediment load, nutrients, primary 108 productivity, and fresh water follow a north-south gradient of decreasing intensity (Fig. 4.2) [Murray et al., 2007; Oviatt, 2008]. 4.2.2 Pre-Contact (1300-1600) The intensity, extent, and importance of local Native American horticulture as a major modifier of the landscape is still actively disputed [Waller, 2000]. Harper [1918] proposed that up to 90%-95% of Rhode Island and Massachusetts was covered in forest in the 1600s. Such thick land cover would imply land clearance and widespread horticulture were uncommon. However, evidence does exist to support at least some pre-Contact Native American land clearance. Perhaps the most direct evidence is based on the historical accounts of Verrazano (1524), who said, ”… for that there are plaines 25. or 30. leagues broad, open and without any impediment of trees and such fruitfulness, that any seede being sowne therein, will bringforth most excellent fruite.” Verrazano seems to imply that large areas (~150km2) were clear of forest and ripe for planting. Sedimentary evidence for late pre-Contact landscape change [Anderson and Webb, 1980] comes from the pollen of the weed Ambrosia, used throughout New England as a clear sedimentary marker of the massive land disturbance attributed to Colonization [Parshall et al., 2003]. However, some studies have also noted a smaller disturbance signal preceding the Colonization signal [Anderson and Webb, 1980], primarily near the highly-productive coast where Native Americans were known to concentrate, perhaps in large semi-permanent settlements [Waller, 2000]. Evidence also suggests Native Americans performed land clearance using fire [Patterson and Sassman, 1988] to increase the carrying capacity of the land for white-tailed deer, an 109 important food source [Bernstein, 1993]. Estimates suggest that as much as 80 to 90% (~80-100 x106 moles N/yr) of the Bay’s nutrients were sourced from Rhode Island Sound and the Atlantic Ocean during this period [Nixon et al., 2008] and the remaining (~5-20 x106 moles N/yr) came from the landscape and its population. Did Native American land use have an influence on the Bay’s nutrient dynamics and productivity? 4.2.3 Colonization (1600-1800) In 1636 Roger Williams received land rights to the Providence area at the northern end of the Bay (Fig. 4.1). By the 1660s and 70s settlement of what is now Narragansett County began, and with the Native Americans’ defeat in King Phillips War (1676-1676) the spread and growth of European settlements throughout Massachusetts, Connecticut, and Rhode Island proceeded unhindered [Baines and Weber, 2009; Pastore, 2011; Pesch et al., 2012]. The European desire for beavers’ pelts in the 1600s led to their rapid extirpation in the watershed [Pastore, 2011]. Loss of the keystone beaver population, and the landscape stabilizing effect of their damns and ponds, may have led to the delivery of as much as 22 million m3 of sediment, and its resident nutrient load, into the Bay [Nixon, 1997] and triggered the widespread desiccation of the landscape [Pastore, 2011]. In addition, trees were quickly felled throughout the watershed in the late-1600s and 1700s to produce pasture- and farm-land [Defebaugh, 1907; Harper, 1918; Parshall et al., 2003]. By the mid-1700s, the spread of slave-based plantation farming, focused on the Bay’s southwestern coasts, peaked on commodities heavily biased towards livestock and dairy [Pastore, 2011]. Estimates suggest ~20-40 x106 moles N/yr was excreted by humans and their animals in close proximity of the Bay’s southern shores between 1600 and 1800 [Pastore, 2011], reflecting a doubling to quintupling of pre-Contact nutrient loads. This nutrient loading lead to the eutrophication of tidal lagoons and sheltered creeks in the 110 southern Bay, turning them green in some cases [Pastore, 2011]. In contrast, northern Bay estimates suggest an annual loading of only ~5.5 x106 moles N/yr [Nixon et al., 2008], implying a south to north gradient in nutrient concentrations, the reverse of today’s situation, may have existed between 1600 and 1800. 4.2.4 The Industrial Revolution and Urbanization (1800-present) The 1790s saw the birth of the American Industrial Revolution and with it the addition of pollutants into the Bay’s ecosystem. Industrially important metals (chromium, lead, copper, zinc) were deposited into small rivers adjacent to mills [Nixon, 1991]. The availability of jobs in the urban centers led to population growth concentrated along the northern coast of the Bay at a rate of 22-40% per decade (watershed average [Nixon et al., 2008], representing an important shift in population density away from the agricultural lands in the south. Forest cover reached its minimum circa 1850 [Defebaugh, 1907; Harper, 1918] but anthropogenic nutrient loads to the upper Bay likely remained fairly low until at least 1865 as indicated by the presence of eel grass north of Field’s Point (Fig. 4.1). The introduction of running water to the city of Providence in 1871, and the lack of a plan to remove the resulting waste water, led to the flowing of liquid waste over and through soils directly into the Bay [Nixon et al., 2008]. Similar issues followed in other urban centers and estimates suggest that this quickly led to the doubling or quadrupling of nutrient nitrogen concentrations in the Bay (≥ 50 x106 moles N/yr [Nixon et al., 2008]. Despite the building of organized sewerage systems, initially emptying into local rivers (~1871) but later (~1901) discharged from Field’s Point on the ebb tide, waste-based nutrient loads to the Bay topped ~500 x106 moles N/yr by 1949, and ~605 x106 moles N/yr by 1980, when legislative steps were finally taken to decrease the nutrient load discharged to the Bay in wastewater [Nixon, 1991]. 111 4.2.6 Proxies of environmental change A multi-proxy approach to the sediment record of the Bay provides an opportunity to investigate how the watershed has evolved in response to human pressures over the past several hundred years. Landscape disturbance can be tracked using the concentrations of pollen of the invasive weed Ambrosia [Parshall et al., 2003]. Delivery of terrestrial sediment to a basin can be tracked with magnetic susceptibility (MS) which is sensitive to concentrations of elements such as titanium (Ti) and iron (Fe), although the latter can be later mobilized in the sediments due to changes in redox chemistry causing elevated Fe/Ti ratios in overlying surface sediments due to authigenic iron precipitation at the redox boundary [as reviewed in Lyons and Severman, 2006]. Soil mobilization can be tracked using a group of organic biomarkers called branched glycerol-dialkyl glycerol-tetraethers (brGDGTs; Fig. 4.3). These compounds are primarily produced in soils [Hopmans et al., 2004; Weijers et al., 2006a] and peat [Weijers et al., 2006b], but can also be produced in oxic and anoxic lake and marine water columns and sediments [Chappe et al., 1982; Peterse et al., 2009; Tierney and Russell, 2009; Zhu et al., 2011]. Here we focus on distributions of three brGDGT homologues which are similar in all chemical aspects except the number of alkyl branches (4, 5, and 6 each for Group Ia, IIa, and IIIa, respectively; Fig. 4.3). All three groups are thought to be produced, at least in part, by certain classes of bacteria, specifically anaerobic Acidobacteria, although their distributions can vary according to source and environment [Damste et al., 2011; Weijers et al., 2009]. The concentration of brGDGTs in marine settings has been suggested as a more reliable proxy for terrestrial organic matter delivery than the previously proposed BIT Index (BIT Index = (area1050 + area1036 + area1022) / (area1292 + area1050 + area1036 + area1022; Fig. 4.3), based on the ratio of brGDGTs to crenarchaeol a GDGT produced by marine thaumarchaeota [Hopmans et al., 2004]), especially in productive coastal environments [Smith et al., 112 2012]. This preference is based on evidence that the BIT Index in these settings can be more sensitive to changes in marine production of crenarchaeol than delivery of terrestrial brGDGTs. In this work, we test the suitability of both proxies in Narragansett Bay. Nitrogen isotope ratios (δ15N) of bulk sediment can track changes in N sources through time. Soil organic matter in pre-Contact landscapes had a δ15N of ~2‰ [Elliott and Brush, 2006]. The δ15N of marine organic matter is approximately 7‰ [Benner et al., 1997; Sweeney and Kaplan, 1980] because of biologic fraction associated with N uptake and assimilation from ocean water [~4‰; Altabet et al., 1995]. In comparison, nutrient nitrogen in water sourced from cultivated-irrigated fields, livestock feedlots, and septic systems is relatively [~7.4 – 21.3‰; Burns and Kendall, 2002; Harrington et al., 1998; Komor and Anderson, 1993] allowing us to track the relative contributions of these end members to the ecosystem. Lastly, the effects of land use change and nutrient delivery on the Bay’s productivity will be estimated using the concentrations of benthic foraminifera, important single celled heterotrophs in the modern Bay that have been shown to closely track modern gradients of nutrients and overall productivity [Martin et al., 2013]. 4.3 Methods 4.3.1 Water-Column Survey To identify modern patterns of water column brGDGT concentrations, surface waters were analyzed for particulate organic matter at eight locations (Fig. 4.1, circles). Samples were collected on a National Marine Fisheries Service cruise in January 2010. One 20L carboy was filtered through a 1μm glass fiber filter (Pall Corp.), and then frozen for biomarker extraction. The entire filter was extracted for GDGT analysis (see below). Seven of the eight filters yielded measureable concentrations of brGDGTs. One sample contained brGDGTs but the chromatography was too poor to allow its analysis. 113 4.3.2 Nitrogen Isotopes The δ15N of bulk sediment was obtained, in the laboratory of Mark Altabet (UMass-Dartmouth), on powdered freeze-dried sediment approximately every 2cm following the procedure of Altabet et al. [1995] on a CN analyzer coupled to an isotope- ratio mass spectrometer using He as a carrier gas and are reported in standard “δ” notation relative to air. Precision (1σ) determined from analysis of replicate samples (n=10 triplicates) and standards (n=53) is the same, ±0.2‰. 4.3.3 Benthic Foraminiferal Counts Samples for foraminiferal analyses were taken approximately every 10cm, soaked in water for no more than 48 hours, wet sieved through a 63μm mesh sieve, then dried at 50⁰C before identifying individual benthic foraminifera. The abundance of benthic foraminifera is normalized per gram dry weight of sediment. 4.3.4 Ambrosia Pollen Samples for pollen were processed using the standard pollen techniques [Faegri and Iverson, 1989]. We selected sediment samples (~1 cm3) for analysis based on the preliminary [Pb]-based age model. They were treated with KOH (to remove organics), 10% HCl (to detect, and then if present, remove carbonates), HF (to remove silicates), and lastly, acetolysis (to remove organics and stain pollen residue). The samples were then transferred to vials, and amended with silicon oil. The average number of grains counted per sample was 120, ranging from 23 to 216. Results are presented as a percentage of total pollen. Pollen analyses were not accomplished on core NB44v. 4.3.5 GDGTs Powdered freeze-dried sediments, sampled at least every 2cm were extracted with dichloromethane on an Accelerated Solvent Extractor (Dionex, ASE200) at 150°C and 1500 psi to produce a total lipid extract (TLE). TLEs were then separated over silica gel using hexane, dichloromethane, and methanol to yield hydrocarbon, ketone, and 114 polar (GDGT) fractions, respectively. The polar fraction was dried under N2, redissolved in hexane/isopropanol (99:1) and filtered through a PTFE 0.22μm filter prior to analysis using high performance liquid chromatography/atmospheric pressure chemical ionization-mass spectrometry. Analysis was performed with an Agilent 1200 liquid chromatograph coupled to a 6130 single-quadrupole mass-selective detector, using the single-ion monitoring mode. Peak areas were integrated according the method described by Weijers et al. [2007]. Branched GDGTs were quantified via the addition of a C-40 GDGT standard. Analytical precision of sedimentary brGDGT concentration determined on replicate injections of the same sample was ± 5ng/g sediment (n=15). No replicates were analyzed on surface water brGDGT samples. Variability in GDGT concentrations do not always translate in to changes in the BIT Index because of non- linearities associated with the mass spectrometer. 4.3.6 Sediment Cores This study employs 5 sediment cores, sampled at ≤ 2cm resolution (Fig. 4.1; numbered stars). Cores 12 (1.75m) and 45 (2.55m) were taken from Greenwich Bay (GB), and cores 25 (1.25m) and 42 (2.72m) were taken from Potter’s Cove (PC). Cores 12 and 25 were obtained using piston push coring while cores 45 and 42 were obtained using a vibra corer. Core 44 (1.18m) was taken near Fox Island (FI) using a vibra corer. Sedimentary elemental profiles were acquired every ≤2 cm using an INNOV-X 4000 hand held XRF. The three core locations reflect different depositional and bathymetric environments. Greenwich Bay and Potter’s Cove are relatively shallow and protected environments with organic matter rich sediments that accumulate rapidly, while Fox Island is more exposed and tidally swept resulting lower total organic carbon concentration and sedimentation rate. 115 Age control is based on unsupported 210Pb activities (usually helpful over the past 100 years), the characteristic increase in Pb concentrations ([Pb]) deposited in Bay sediments circa 1850 related to industry [Bricker, 1993; Corbin, 1989; Prell, personal communication], Ambrosia pollen horizons diagnostic of landscape disturbance circa 1700 [Parshall et al., 2003], and 14C analyses of benthic foraminifera (Elphidium sp.) and mollusk shells. Radioacarbon ages were calibrated using Oxcal and a standard reservoir correction of 400 years minus a Narragansett Bay-specific correction of 120 years [McNeely et al., 2006]. Age and model error were estimated (Fig. 4.4) using a mixed regression technique [Heegaard et al., 2005]. The average sample spacing in each core was: NB12/45 (Greenwich Bay), 10 years; NB25/42 (Potter’s Cove), 6 years; NB44 (Fox Island), 26 years (Fig. 4.4). 4.4 Results 4.4.1 Sediment Accumulation Rates Sedimentation rates varied through time in NB. Average sedimentation rates over the past 700 years are 0.28, 0.23, and 0.12cm/yr at Potter’s Cove, Greenwich Bay, and Fox Island, respectively (Fig. 4.4, bottom), consistent with each sites’ local depositional environment and distance from the primary sources of sediment. Elevated sedimentation rates are detected in Potter’s Cove in the mid-1600s and mid-1900s, in Greenwich Bay in the mid-1800s and 1900s, and at Fox Island from the early-1800s to the mid-1900s. We note however that the timing of changes in sedimentation rate are dependent on the resolution of our age control, which is admittedly coarse before 1700 in Greenwich Bay and Potter’s Cove, and before 1850 near Fox Island. 4.4.2 GDGTs The concentration of GDGTs in surface water and sediments varies from north to south (Fig. 4.5). Surface water brGDGT concentrations decrease an order of magnitude from 196 to 9ng/L from north to south with distance from the sediment source. In 116 contrast, surface water crenarchaeol concentrations generally increase from 5 to 33ng/L from north to south. The interplay between brGDGT and crenarchaeol concentrations results in BIT Index values which decrease rapidly with distance down Bay from .97 to .24 (Fig. 4.5). The relative contributions of the different GDGT groups (Ia, IIa, and IIIa; Fig. 4.3) are fairly constant down Bay and are similar in the West and East Passages (Fig. 4.6). Surface sediment brGDGT concentration decreases rapidly down Bay from 2.6 to 0.13ug/g (Fig. 4.5). The brGDGT concentration in Greenwich Bay and Potter’s Cove is about twice those from more open Bay sites to the north and south. Surface sediment crenarchaeol concentration also decreases down Bay from 0.20 to 0.13ug/g, with slightly elevated concentrations in Greenwich Bay and Potter’s Cove. This stands in contrast to results from our surface water survey, in which crenarchaeol concentration increases down Bay (Fig. 4.5). The BIT Index in surface sediments decreases from 0.92 to 0.47 with Greenwich Bay and Potter’s Cove having elevated values compared to open Bay sites to the north and south. The relative contributions of the different GDGT groups changes in surface water and sediments down Bay (Fig. 4.6). For example, in surface sediments GDGT IIIa increases from 18 to 32%, GDGT IIa decreases from 46 to 36%, and GDGT Ia decreases only slightly from 35 to 32% with distance from the main source of sediment. Additionally, large differences are noted between surface water and sediment surface samples (Fig. 4.6). In surface waters, GDGT IIIa only accounts for ~6% of the total GDGTs, but in surface sediments its contribution is ~27%. Similarly, GDGT IIa and Ia contribute 42 and 52% to total GDGTs in the surface water, respectively, but only 40 and 33% in the sediments. Downcore brGDGT concentrations, calculated for the entirety of each record, were an order of magnitude higher in Greenwich Bay (avg = 0.33ug/g sed) than near 117 Fox Island (avg = 0.075ug/g sed) or Potter’s Cove (avg = 0.047ug/g sed) (Fig. 4.7). Downcore crenarchaeol concentrations were also higher in Greenwich Bay (avg = 0.13ug/g sed) then in Potter’s Cove (avg = 0.047ug/g sed) or near Fox Island (avg = 0.095ug/g sed). The BIT Index was higher in Greenwich Bay (avg = 0.71) than in Potter’s Cove (avg = 0.55) or near Fox Island (avg = 0.44). The distributions of the different GDGT groups (Ia, IIa, IIIa) also changed downcore (Fig. 4.8), particularly in Greenwich Bay. A stepwise and permanent increase (decrease) in the contribution of GDGTs Ia and IIa (IIIa) initiated circa 1700 in Greenwich Bay, later reversed circa 1900 and then oscillated to the present. Brief transient excursions were also detected in Greenwich Bay in the mid-1400s and 1800s, in Potter’s Cove circa the mid-1600s and 1700s, and near Fox Island in the mid-1800s. Correlation coefficients and significance levels were computed between downcore brGDGT and crenarchaeol concentrations and other environmental data (n = 106): δ15N, [Fe], [Ti], [Fe] / [Ti] (a proxy for sediment anoxia), and MS (Table 4.1). Notable significant correlations were detected between all GDGT moieties and Uk’37 inferred SST, alkenone inferred productivity, [Ti], [Fe] / [Ti], and MS. A primary component analysis (PCA) of the same data (Fig. 4.9) shows that component 1 describes covariance between GDGT concentrations and [Fe] / [Ti], [Ti], and alkenone inferred productivity, and explains 53% of the variability in the dataset. Component 2 explains 18% of the variability and is determined by covariance between nitrogen isotopes and magnetic susceptibility. 4.4.3 Ambrosia Pollen Ambrosia pollen accounted for between 0 and 14% of total pollen in Potter’s Cove and Greenwich Bay (Fig. 4.10). Pollen was not analyzed on the Fox Island core. Sharp increases in Ambrosia pollen, diagnostic for the Colonial increase in regional land clearance circa 1700, are noted at Potter’s Cove and Greenwich Bay and form the basis 118 of age control in these parts of the cores. Additionally, a smaller increase in Ambrosia pollen concentration was detected deeper in both cores in the mid-1600s. 4.4.4 Magnetic Susceptibility Magnetic susceptibility (MS) is generally stable at our three cores sites until the mid-1700s when it begins to increase until the 1900s (Fig 4.10). At Greenwich Bay and Potter’s Cove, peak MS values are followed by a rapid decrease into the present that is not seen at Fox Island. 4.4.5 Nitrogen Isotopes Stable-isotopes of nitrogen (δ15N; Fig. 4.10) are more enriched and less variable at Fox Island (7.3-9.7‰, avg 8.1‰) than at Greenwich Bay (5.9-8.6‰, avg 6.8‰) or Potter’s Cove (4.4-8.2‰, avg 6.2‰). The average δ15N at our core sites (± 1σ) was 6.8 ± 0.9‰ from 1300-1600, 6.1 ± 0.8‰ from 1600-1800, and 7.8 ± 0.7‰ from 1800 to present. 4.4.6 Benthic Foraminifera Benthic foraminiferal abundance (Fig. 4.10) varies between 0 and ~600 foraminifera/g sed. The absence of benthic foraminifera in Greenwich Bay before 1827 may reflect dissolution of their tests in the organic rich sediments. Abundance at both Greenwich Bay and Fox Island increases sharply around 1870 coincident with enriched δ15N ratios. Foraminifera abundance decreases circa 1910 near Fox Island coincident with maximum abundances at Greenwich Bay. Abundances then decrease in Greenwich Bay circa 1980. The anti-phased relationship in abundance between Fox Island and Greenwich Bay may be caused by the increasingly inhospitable conditions in Greenwich Bay during the late 1900s or by incomplete capture of the most modern sediments there during coring, as has been discussed elsewhere [Salacup et al., in prep]. 119 4.5 Discussion 4.5.1 Sources of GDGTs to Narragansett Bay The use of GDGT-based proxies of soil delivery (BIT Index and brGDGT concentrations) is built on the idea that brGDGTs are only produced in continental environments. Previous work invalidates this premise in some lakes [Tierney and Russell, 2009], fjords [Peterse et al., 2009], and coastal/estuarine settings [Zhu et al., 2011] by showing that GDGT production can also occur in situ. Our results also suggest that GDGT-based records of soil delivery are confounded by multiple inputs in Narragansett Bay. At first glance, down-Bay decreases in surface water and sediment brGDGT concentrations appear to support the utility of brGDGTs and the BIT Index (Fig. 4.5). However, closer analysis implies this assumption is misled. Firstly, downcore concentrations of brGDGTs vary strongly and inversely with the concentration of titanium, a proxy for continental lithic material that should vary directly with brGDGT concentration if it is recording the delivery of soil OM (Table 4.1). Secondly, the ratio of the GDGT groups (Ia, IIa, IIIa) can act as a fingerprint to identify a given GDGT source. Following this logic, the ratio of each of the groups should be broadly similar in surface sediments as in the overlying water if brGDGTs delivered from the landscape pass thru a water column, and are deposited in an environment, that both lack GDGT production. Large changes in the relative proportions of the different groups in surface sediment with distance down Bay (Fig. 4.6), between water samples and underlying sediments (Fig. 4.6), and across apparent “events” in the sediment record (Fig. 4.8) implies the existence of multiple sources of brGDGTs to Narragansett Bay, thus destroying our ability to use these compounds as a tracer of any environmental variable through time in Narragansett Bay (Fig. 4.7). Changes in the proportions of GDGT groups down Bay and between surface waters and surface sediments may be due to the preferential preservation / degradation 120 of one group over the others. For example, results based on GDGT IIIa (Fig. 4.6) showing its relative enrichment both down Bay (from 7-11% in surface waters and from 18-32% in surface sediments), and from the surface waters to surface sediments, are consistent with its preferential preservation. However, GDGT Ia concentration increases slightly down Bay (from 44-47%) in the surface water, consistent with preferential preservation, while its concentration decreases from near 50% in surface waters to near 30% in surface sediments, suggesting its preferential degradation (Fig. 4.6). This variable behavior is inconsistent with preferential preservation / degradation as the primary driver of variability in GDGT group proportions. A simpler explanation evokes the existence of two or more brGDGT producing populations, each with its own fingerprint, in the Narragansett Bay watershed, one or more of which may live in the Bay’s water column or sediments. Branched GDGTs are produced, at least in part, by certain classes of bacteria. However, the scientific community has yet to isolate and culture microbes responsible for the production of the entire suite of brGDGTs [Schouten et al., 2013]. In the modern Bay, the major disparity in brGDGT proportions between surface waters and sediments suggests a sediment- hosted brGDGT-producing population that produces more GDGT IIIa and less GDGT Ia than the population in the overlying surface waters and/or the surrounding landscape (Fig. 4.6). Fe/Ti ratios, which should increase with increasing sediment anoxia, co-vary significantly with GDGT concentrations suggesting the sediment hosted population may prosper when sediments are more oxygenated (Table 4.1). However, this relationship could be driven by Ti concentration, which also has a strong negative and significant relationship with GDGT concentrations. Our downcore results from Greenwich Bay (Fig. 4.8) suggest a local change in the source(s) of GDGTs took place circa 1700, coincident with regional widespread land clearance. Pre-1700 brGDGT proportions were nearly equal between the three groups, 121 but after 1700, contributions of GDGT Ia and IIa increased at the expense of GDGT IIIa. This pattern later returned briefly to pre-1700 values in the mid-1900s before an apparent oscillation set up between pre and post-1700 brGDGT proportions (Fig. 4.8). This oscillatory behavior occurred during a period of extreme nitrogen enrichment and declining foraminifera abundance in Greenwich Bay (Fig. 4.9) suggesting a possible control of nutrient dynamics on GDGT distributions. This is supported by our correlation and PCA results which highlight significant covariance between GDGT concentrations and alkenone-inferred productivity, which is also sensitive to nutrient concentrations [Herbert, 2003]. The negative correlation between GDGT and alkenone concentrations may imply that whatever organisms produce GDGTs in Narragansett Bay compete with the alkenone-producing haptophyte algae for nutrients and/or niche space. 4.5.2. Pre-Contact (1300-1600) Landscape Stability The impacts, if any, of Native American land use practices on the pre-Contact ecosystem of Narragansett Bay is still debated [Waller, 2000, and references therein]. Archaeological evidence for the sustained habitation of Native Americans in concentrated settlements suggests their footprint may have been large, albeit on a very local scale. Our results do not provide any conclusive evidence regarding Native American activities on the Bay’s ecosystem. Ambrosia pollen concentrations, benthic foraminifera counts, and MS are low and relatively stable during this period implying regional landscape dynamics and water column nutrient concentrations remained largely undisturbed (Fig 4.10). The down Bay gradient in δ15N implied between our Greenwich Bay and Fox Island core sites at this time likely reflects a natural gradient between marine (enriched) and soil (depleted) sourced N at these sites consistent with their distance from the major sources of terrestrial material. In fact, assuming the δ15N of any small amount of marine organic matter (dissolved or suspended) entering the Bay from Rhode Island Sound is 122 ~8‰ (the average δ15N at Fox Island during this period and close to the 7‰ estimate provided by Benner et al, [1997] and Sweeney and Caplan [1980]), and the δ15N of terrestrial organic matter delivered from the pre-Contact landscape is ~2‰ [Elliott and Brush, 2006], an admittedly simple two end-member isotope mixing model can be made to estimate the contributions of marine and soil organic matter to Greenwich Bay given the δ15N of bulk sediment there averages ~6‰. We implicitly assume nitrogen utilization occurs to completion in the Bay, or is at steady state, between 1300 and 1600, and thus kinetic isotope effects associated with incomplete nutrient uptake can be ignored. The results of this model suggest ~70% of the N in Greenwich Bay was sourced from the Sound and ~30% from the landscape. This would be an overestimate of the landscape contribution because the soil end-member isotope value (~2‰) does not include the δ15N of nitrogen inputs from humans and animals living on the landscape, which would increase the δ15N of the terrestrial end-member via their waste. In this light, it is remarkably close to empirically-derived estimates placing the ratio at ~80% from the Sound and 20% from the landscape [Nixon et al., 2008] and a total pre-Industrial flux of ~16 x106 mole N/yr. 4.5.3. The Colonial Disturbance (1600-1800) Traditionally, the onset of truly widespread regional land clearance is placed around 1700, coincident with the rapid increase of Ambrosia pollen in regional sedimentary basins. As mentioned earlier, this increase was detected in our sediment cores, is coincident with increasing MS values, and constrains our age models at this time. However, smaller increases in Ambrosia pollen were discovered in Greenwich Bay and Potter’s Cove which date to the mid-1600s, coincident with locally high MS values in Potter’s Cove (Fig. 4.10). These smaller Ambrosia increases take place over several data points (5 in each core) and occur simultaneously in both cores suggesting the data are not spurious. 123 Instead, the timing of the early Ambrosia disturbance fits well with that of the extirpation of the watersheds’ beaver population in the 1600s. It has been argued that the removal of the beavers may have led to the delivery of as much as 22 x 106m3 of sediment, and its resident nutrient load, into the Bay [Nixon, 1997]. A 0.7‰ depletion in δ15N between the pre-Contact and Colonial Periods in our records may reflect the delivery of relatively depleted terrestrially sourced N during the emptying of beaver ponds and wide spread clearance and destabilization of the landscape in the watershed (Fig. 4.10). The strong enrichment δ15N detected at Potter’s Cove towards the end of this period may suggest animal waste from small farms, primarily dealing in livestock, that had been established on Prudence Island by this time [Baines and Weber, 2009], were augmenting the nutrient dynamics of this relatively restricted basin. In response to increases in nutrient supply between 1600 and 1800, foraminifera abundances at Fox Island more than double, from 59 to 147 counts/g sed, suggesting increasing nutrient supplies to the Bay’s ecosystem were beginning to alter the Bay’s ecosystem. Taken together, our results suggest the extirpation of the native beaver population and the wide spread clearance of forests directly impacted the nutrient dynamics and primary productivity of the Bay in a way not seen during the preceding 300 years. 4.5.5 Industrialization and Urbanization (1800-present) This relatively well-documented period in the watershed’s history is marked by land use change characterized by high concentrations of Ambrosia pollen, a rapid and persistent increase in nitrogen isotope ratios, a tripling of benthic foraminifera abundance, and elevated magnetic susceptibility (Fig. 4.10). Ambrosia pollen and magnetic susceptibility increase until approximately 1900 (the average of Greenwich Bay and Potter’s Cove) and the mid-1950s, respectively, consistent with evidence of maximum regional land clearance circa 1850 [Defebaugh, 1907; Harper, 1918] and the 124 post-Industrial increase in sedimentary metal pollution [Corbin, 1989]. Urbanization led to rapid population growth in the urban centers of the watershed [Baines and Weber, 2009; Pesch et al., 2012]. The δ15N enrichment, beginning broadly in the mid-1800s, is coincident with this population growth and an estimated 10 fold increase in human and animal-derived N delivery to the Bay between 1865 and 1925 (Fig. 4.10). The pattern of increasing (mid-1800s to late-1900s) and decreasing (late-1900 to present) foraminifera abundance is very closely timed with the enrichment in sedimentary δ15N and increases in waste-derived nitrogen (Fig. 4.9) estimated by Nixon et al. [2008]. These results imply this ‘new’ source of N led to spatially variable impacts on benthic productivity patterns, likely through the effects of oxygen stress on niche availability, a characteristic of the cultural eutrophication that has come to characterize the modern mid- and upper-Bay and Providence River [Codiga et al., 2009; Deacutis, 2008; RIDEM, 2003]. Taken together, our time series results suggest post-Industrial environmental change was different than anything the Bay experienced in the previous 700 years. 4.6 Conclusions Analysis of GDGT concentrations in Narragansett Bay suggest GDGT-based paleo-environmental reconstructions are confounded by in situ production. A sediment- hosted GDGT-generating population is suspected; characterized by higher GDGT IIIa and lower GDGT Ia than found in overlying waters. Nitrogen isotope ratios, Ambrosia pollen concetrations, and benthic foraminifera abundance in sediments from three geographically separated cores in Narragansett Bay suggest that modern north-south gradients in fresh water, sediment load, and nutrients have been a persistent, although changing, feature of the Bay over the past 700 years. The pre-Industrial balance between marine and terrestrial organic matter sources to the Bay was interrupted three times: first (1600s) by the addition of terrestrial nutrients after the extirpation of the native 125 beaver population, second by the wide spread land clearance and rapid growth in the watershed’s domestic animal population by the Colonists (1700s) and, third by the massive ~30 fold increase in nutrient flux from the northern Bay after the Industrial Revolution. The nutrient N load added to the Bay concurrent with these events directly affected the Bay’s ecosystem leading to an increase in its benthic productivity. This work implies that humans began to more heavily influence the Bay’s ecosystem as soon as the early 1600s, but that the environmental degradation that occurred over the past 150 years is unprecedented in Narragansett Bay over at least that past 700 years. 126 4.6 Tables Table 4.1. Results of correlation (r) and significance (p) tests between GDGTs and environmental variables on paired downcore samples from all cores in this study. [br] [cren] [IIIa] [IIa] [Ia] BIT ug/g ug/g ug/g ug/g ug/g SST* r 0.71 0.56 0.54 0.47 0.55 0.50 p 0.00 0.00 0.00 0.00 0.00 0.00 Productivity# r -0.71 -0.42 -0.19 -0.36 -0.39 -0.42 p 0.00 0.00 0.05 0.00 0.00 0.00 15 N r -0.49 -0.12 0.36 -0.07 -0.12 -0.17 p 0.00 0.24 0.00 0.51 0.22 0.09 Fe & r 0.09 0.18 0.07 0.19 0.16 0.17 p 0.38 0.06 0.48 0.05 0.10 0.08 Fe/Ti& r -0.75 -0.62 -0.57 -0.47 -0.62 -0.55 p 0.00 0.00 0.00 0.00 0.00 0.00 Ti& r -0.89 -0.55 -0.32 -0.39 -0.56 -0.52 p 0.00 0.00 0.00 0.00 0.00 0.00 MS r 0.77 0.39 0.10 0.27 0.40 0.40 p 0.00 0.00 0.33 0.01 0.00 0.00 *SST determinations are based on Uk'37 (Salacup et al, in prep) # Productivity determinations are based on sedimentary alkenone concentrations (Salacup et al, in prep) & Metal data from Murray et al. (2007) 127 4.7 Figures Figure 4.1. Map of Narragansett Bay and points of interest: core locations (stars; Greenwich Bay (NB12/45v); Potter’s Cove (NB25/42v); Fox Island (NB44v)), surface water POM sampling locations (open circles). Arrows represent major sources of fresh water. Stippled lines denote the functional boundary between sections of the Bay [Costa- Pierce and Desbonnet, 2008]. 128 Figure 4.2. The north-south gradients of nutrient induced productivity and pollution is evident in two plots showing the concentrations of organic matter (a proxy for productivity) and Pb (an industrial pollutant) in the surface sediments of Narragansett Bay. Maps adapted from Murray et al. [2007]. 129 Figure 4.3. Structures and representative m/z ion ratios for GDGTs analyzed in this study. Definition of the BIT Index is based on the area under the curve for each ion on the appropriate ion chromatogram. 130 Figure 4.4. Age models for cores NB25/42v (left), NB12/45v (middle), NB44v (right) were established using 14C of mollusks and foraminifera (filled squares), pollen horizons (triangles), [Pb] increases (open squares), and 210Pb (circles). Sediment accumulation rates are shown below the appropriate core. 131 Figure 4.5. Profiles of brGDGT concentrations and BIT Index in water column (left) and surface sediment (right) samples with latitude down Bay away from the major sources of terrestrial sediments at the Providence and Taunton Rivers. GDGT concentrations and BIT Index values in Greenwich Bay and Potter’s Cove samples, at right, are likely higher than those directly to the north and south due to their relatively restricted morphology and higher total organic carbon concentrations (Fig. 1.2). 132 Figure 4.6. Plots of the relative contributions of the three different branched GDGT groups in surface water (cirlces) and surface sediment (triangels) with latitude. 133 Year Figure 4.7. Timeseries of brGDGT concentration, crenarchaeol concentration, and the BIT Index at Greenwich Bay (cirlces), Potter’s Cove (squares), and Fox Island (triangles). Variability in GDGT concentrations do not always translate in to changes in the BIT Index because of non-linearities associated with the mass spectrometer. 134 Year Figure 4.8. Timeseries of the relative contributions of the three different branched GDGT groups. 135 PC2 PC1 Figure 4.9. Scatter plot of our PCA results. GDGTs are shown in black circles, environmental data in open circles. 136 Year Figure 4.10. Timeseries of (A) % Ambrosia pollen, (B) bulk sediment nitrogen isotopes (δ15N); (C) benthic foraminiferal abundance, and (D) magnetic susceptibility (MS). Nitrogen fluxes estimated by Nixon et al. [2008] are co-plotted in (C, dashed line). Dashed lines in (B) represent the average δ15N for the given interval of time. 137 4.8 References Ahmed, M., et al. (2013), Continental-scale temperature variability during the past two millennia, Nature Geoscience, 6(5), 339-346. Altabet, M. A., R. Francois, D. W. Murray, and W. L. Prell (1995), Climate-Related Variations in Denitrification in the Arabian Sea from Sediment N-15/N-14 Ratios, Nature, 373(6514), 506-509. Anderson, D. M., J. T. Overpeck, and A. K. Gupta (2002), Increase in the Asian southwest monsoon during the past four centuries, Science, 297(5581), 596-599. Anderson, P., and T. I. Webb (1980), Two short cores from Greenwich Cove area, Rhode Island, paper presented at 20th Annual meeting of the Northeastern Anthropological Association, Amherst, Ma. Andersson, C., B. Risebrobakken, E. Jansen, and S. O. Dahl (2003), Late Holocene surface ocean conditions of the Norwegian Sea (Voring Plateau), Paleoceanography, 18(2). Baines, J. L., and R. L. J. Weber (2009), Human and Land-Use, in An Ecological Profile of the Narragansett Bay National Estuarine Research Reserve, edited by K. B. Raposa and M. L. Schwartz, Rhode Island Sea Grant, NArragansett, RI. Bendle, J., A. Rosell-Mele, and P. Ziveri (2005), Variability of unusual distributions of alkenones in the surface waters of the Nordic seas, Paleoceanography, 20(2). Benner, R., B. Biddanda, B. Black, and M. McCarthy (1997), Abundance, size distribution, and stable carbon and nitrogen isotopic compositions of marine organic matter isolated by tangential-flow ultrafiltration, Mar Chem, 57(3–4), 243-263. Bernstein, D. J. (Ed.) (1993), Prehistoric Subsistence on the Southern New England Coast, the Record from Narragansett Bay, 188 pp. pp., Academic Press, Inc., San Diego. Blanz, T., K. C. Emeis, and H. Siegel (2005), Controls on alkenone unsaturation ratios along the salinity gradient between the open ocean and the Baltic Sea, Geochimica et Cosmochimica Acta, 69(14), 3589-3600. Booth, B. B. B., N. J. Dunstone, P. R. Halloran, T. Andrews, and N. Bellouin (2012), Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability (vol 484, pg 228, 2012), Nature, 485(7399), 534-534. Boothroyd, J. C. a. A., Peter, V. (2008), Geologic and Contemporary Landscapes of the Narragansett Bay Ecosystem, in Science for Ecosystem-Based Management, edited by A. D. a. B. A. Costa-Pierce, p. 570, Springer, New York, NY. Borkman, D. G., and T. Smayda (2009), Multidecadal (1959-1997) changes in Skeletonema abundance and seasonal bloom patterns in Narragansett Bay, Rhode Island, USA, J. Sea Res., 61(1-2), 84-94. 138 Braconnot, P., S. P. Harrison, M. Kageyama, P. J. Bartlein, V. Masson-Delmotte, A. Abe-Ouchi, B. Otto-Bliesner, and Y. Zhao (2012), Evaluation of climate models using palaeoclimatic data, Nature Climate Change, 2(6), 417-424. Bradley, R. S., and P. D. Jones (1992), "When was the Little Ice Age?", in Proceedings of the International Symposium on the "Little Ice Age" Cliamte, edited by T. Mikami, Department of Geography, Tokyo Metropolitan University. Brassell, S. C., G. Eglinton, I. T. Marlowe, U. Pflaumann, and M. Sarnthein (1986), Molecular Stratigraphy - a New Tool for Climatic Assessment, Nature, 320(6058), 129- 133. Bricker, S. B. (1993), The History of Cu, Pb, and Zn Inputs to Narragansett Bay, Rhode- Island as Recorded by Salt-Marsh Sediments, Estuaries, 16(3B), 589-607. Broerse, A., P. Ziveri, and S. Honjo (2000), Coccolithophore (CaCO3) flux in the Sea of Okhotsk: seasonality, settling and alteration processes, Mar Micropaleontol, 39, 179- 200. Burns, D. A., and C. Kendall (2002), Analysis of delta(15)N and delta(18)O to differentiate NO(3)(-) sources in runoff at two watersheds in the Catskill Mountains of New York, Water Resour Res, 38(5). Chappe, B., P. Albrecht, and W. Michaelis (1982), Polar Lipids of Archaebacteria in Sediments and Petroleums, Science, 217, 65-66. Charman, D. J., V. Hohl, A. Blundell, F. Mitchell, J. Newberry, and P. Oksanen (2012), A 1000-year reconstruction of summer precipitation from Ireland: Calibration of a peat- based palaeoclimate record, Quatern Int, 268, 87-97. Chu, G. Q., Q. Sun, S. Q. Li, M. P. Zheng, X. X. Jia, C. F. Lu, J. Q. Liu, and T. S. Liu (2005), Long-chain alkenone distributions and temperature dependence in lacustrine surface sediments from China, Geochimica et Cosmochimica Acta, 69(21), 4985-5003. Cleaveland, L. C., and T. D. Herbert (2009), Preservation of the alkenone paleotemperature proxy in uplifted marine sequences: A test from the Vrica outcrop, Crotone, Italy, Geology, 37(2), 179-182. Clement, A. C., R. Seager, M. A. Cane, and S. E. Zebiak (1996), An ocean dynamical thermostat, J Climate, 9(9), 2190-2196. Codiga, D. L., H. E. Stoffel, C. F. Deacutis, S. Kiernan, and C. A. Oviatt (2009), Narragansett Bay Hypoxic Event Characteristics Based on Fixed-Site Monitoring Network Time Series: Intermittency, Geographic Distribution, Spatial Synchronicity, and Interannual Variablity, Estuaries and Coasts, 32, 621-641. Conte, M. H., A. Thompson, and G. Eglinton (1994), Primary production of lipid biomarker compounds by Emiliania huxleyi: results from an experimental mesocosm study in fjords of southern Norway, Sarsia, 79, 319-332. 139 Conte, M. H., A. Thompson, D. Lesley, and R. P. Harris (1998), Genetic and physiological influences on the alkenone/alkenoate versus growth temperature relationship in Emiliania huxleyi and Gephyrocapsa oceanica, Geochimica et Cosmochimica Acta, 62(1), 51-68. Conte, M. H., J. C. Weber, L. L. King, and S. G. Wakeham (2001), The alkenone temperature signal in western North Atlantic surface waters, Geochimica et Cosmochimica Acta, 65(23), 4275-4287. Conte, M. H., M. A. Sicre, C. Ruhlemann, J. C. Weber, S. Schulte, D. Schulz-Bull, and T. Blanz (2006), Global temperature calibration of the alkenone unsaturation index (U-37(K ')) in surface waters and comparison with surface sediments, Geochemistry Geophysics Geosystems, 7, -. Cook, E. R., and P. Mayes (1985), Decadal-scale patters of climatic change over eastern North America inferred from tree-rings, in Abrupt Climatic Change: Evidence and Implications, edited by W. H. Berger and L. D. Labeyrie, pp. 61-66, NATO ASI Series. Cook, E. R., R. D. D'Arrigo, and M. E. Mann (2002), A Well-Verified, Multiproxy Reconstruction of the Winter North Atlantic Oscillation Index since a.d. 1400*, J Climate, 15(13), 1754-1764. Corbin, J. (1989), Recent and historical accumulation of trace metal contaminants in Narragansett Bay sediments, 295 pp, University of Rhode Island, Kingston, RI. Costa-Pierce, B. A., and A. Desbonnet (2008), An "Ecofunctional" Approach to Ecosystem-based Management for Narragansett Bay, in Science for Ecosystem-based Management, edited by A. Desbonnet and B. A. Costa-Pierce, pp. 545-562, Springer, New York, NY. Cronin, T. M., K. Hayo, R. C. Thunell, G. S. Dwyer, C. Saenger, and D. A. Willard (2010), The Medieval Climate Anomaly and Little Ice Age in Chesapeake Bay and the North Atlantic Ocean, Palaeogeography, Palaeoclimatology, Palaeoecology, 297(2), 299-310. Curry, R. G., and M. S. McCartney (2001), Ocean gyre circulation changes associated with the North Atlantic Oscillation, Journal of Physical Oceanography, 31(12), 3374- 3400. D' Arrigo, R. D., E. R. Cook, M. E. Mann, and G. C. Jacoby (2003), Tree-ring reconstructions of temperature and sea-level pressure variability associated with the warm-season Arctic Oscillation since AD 1650, Geophysical Research Letters, 30(11). D'Andrea, W. J., and Y. Huang (2005), Long chain alkenones in Greenland lake sediments: Low delta C-13 values and exceptional abundance, Organic Geochemistry, 36(9), 1234-1241. Damste, J. S. S., W. I. C. Rijpstra, E. C. Hopmans, J. W. H. Weijers, B. U. Foesel, J. Overmann, and S. N. Dedysh (2011), 13,16-Dimethyl Octacosanedioic Acid (iso-Diabolic Acid), a Common Membrane-Spanning Lipid of Acidobacteria Subdivisions 1 and 3, Appl Environ Microb, 77(12), 4147-4154. 140 De Leeuw, J. W., F. W. V. D. Meer, W. I. C. Rijpstra, and P. A. Schnencck (1980), On the occurrence and structural identification of long chain unsaturated ketones and hydrocarbons in sediments, in Advances in Organic Geochemistry, 1979, edited by E. D. Douglas and J. R. Maxwell, pp. 211-217, Pergamon, Oxford. Deacutis, C. F. (2008), Evidence of Ecological Impacts from Excess Nutrients in Upper Narragansett Bay, in Science for Ecosystem-Based Management, edited by A. a. C.-P. Desbonnet, B.A., pp. 349-381, Springer, New York, NY. Defebaugh, J. E. (1907), History of the lumber industry of America Volume 2, The American Lumberman. DeLong, K. L., T. M. Quinn, F. W. Taylor, K. Lin, and C. C. Shen (2012), Sea surface temperature variability in the southwest tropical Pacific since AD 1649, Nature Climate Change, 2(11), 799-804. DeLong, K. L., T. M. Quinn, F. W. Taylor, C. C. Shen, and K. Lin (2013), Improving coral-base paleoclimate reconstructions by replicating 350 years of coral Sr/Ca variations, Palaeogeogr Palaeocl, 373, 6-24. Deser, C., M. A. Alexander, S. P. Xie, and A. S. Phillips (2010), Sea Surface Temperature Variability: Patterns and Mechanisms, in Annual Review of Marine Science, edited, pp. 115-143. Dickson, R., J. Lazier, J. Meincke, P. Rhines, and J. Swift (1996), Long-term coordinated changes in the convective activity of the North Atlantic, Prog Oceanogr, 38(3), 241-295. Dissard, D., G. Nehrke, G. J. Reichart, and J. Bijma (2010), The impact of salinity on the Mg/Ca and Sr/Ca ratio in the benthic foraminifera Ammonia tepida: Results from culture experiments, Geochimica et Cosmochimica Acta, 74(3), 928-940. Dunbar, R. B., G. M. Wellington, M. W. Colgan, and P. W. Glynn (1994), EASTERN PACIFIC SEA-SURFACE TEMPERATURE SINCE 1600-AD - THE DELTA-O-18 RECORD OF CLIMATE VARIABILITY IN GALAPAGOS CORALS, Paleoceanography, 9(2), 291-315. Elliott, E. M., and G. S. Brush (2006), Sedimented organic nitrogen isotopes in freshwater wetlands record long-term changes in watershed nitrogen source and land use, Environ Sci Technol, 40(9), 2910-2916. Enfield, D. B., A. M. Mestas-Nunez, and P. J. Trimble (2001), The Atlantic multidecadal oscillation and its relation to rainfall and river flows in the continental US, Geophysical Research Letters, 28(10), 2077-2080. Epstein, B. L., S. D'Hondt, J. G. Quinn, J. P. Zhang, and P. E. Hargraves (1998), An effect of dissolved nutrient concentrations on alkenone-based temperature estimates, Paleoceanography, 13(2), 122-126. Esper, J., D. C. Frank, and R. J. S. Wilson (2004), Climate reconstructions: Low- frequency ambition and high-frequency ratification, Eos, Transactions American Geophysical Union, 85(12), 113-120. 141 Faegri, K., and J. Iverson (Eds.) (1989), Textbook of Pollen Analysis, Fourth Edition ed., 328 pp., John Wiley and Sons, Chichester. Gingerich, P. D. (2006), Environment and evolution through the Paleocene-Eocene thermal maximum, Trends in Ecology & Evolution, 21(5), 246-253. Gray, S. T., L. J. Graumlich, J. L. Betancourt, and G. T. Pederson (2004), A tree-ring based reconstruction of the Atlantic Multidecadal Oscillation since 1567 AD, Geophysical Research Letters, 31(12), -. Gupta, A. K., D. M. Anderson, and J. T. Overpeck (2003), Abrupt changes in the Asian southwest monsoon during the Holocene and their links to the North Atlantic Ocean, Nature, 421(6921), 354-357. Hamburg, S. P., D. Pryor, and M. A. Vadeboncoeur (2008), Nitrogen Inputs to Narragansett Bay: An Historical Perspective, in Science for Ecosystem-Based Management, edited by A. Desbonnet and B. A. Costa-Pierce, pp. 177-210, Springer, New York, NY. Harada, N., M. Sato, A. Shiraishi, and M. C. Honda (2006), Characteristics of alkenone distributions in suspended and sinking particles in the northwestern North Pacific, Geochimica et Cosmochimica Acta, 70(8), 2045-2062. Harada, N., K. H. Shin, A. Murata, M. Uchida, and T. Nakatani (2003), Characteristics of alkenones synthesized by a bloom of Emiliania huxleyi in the Bering Sea, Geochimica et Cosmochimica Acta, 67(8), 1507-1519. Harper, R. M. (1918), Changes in the forest area of New England in three centuries, Journal of Forestry, 16, 442-452. Harrington, R. R., B. P. Kennedy, C. P. Chamberlain, J. D. Blum, and C. L. Folt (1998), N-15 enrichment in agricultural catchments: field patterns and applications to tracking Atlantic salmon (Salmo salar), Chem Geol, 147(3-4), 281-294. Heegaard, E., H. J. B. Birks, and R. J. Telford (2005), Relationships between calibrated ages and depth in stratigraphical sequences: an estimation procedure by mixed-effect regression, Holocene, 15(4), 612-618. Heiss, G. A. (1994), Coral reefs in the Red Sea: Growth, production and stable isotopes, GEOMAR Rep., 1-143 pp. Herbert, T. D. (2003), Alkenone paleotemperature determinations, in Treatise in Marine Geochemistry, edited by H. Elderfield, pp. 391-432, Elsevier, Amsterdam. Herbert, T. D., L. C. Peterson, K. T. Lawrence, and Z. H. Liu (2010), Tropical Ocean Temperatures Over the Past 3.5 Million Years, Science, 328(5985), 1530-1534. Hopmans, E. C., J. W. H. Weijers, E. Schefuss, L. Herfort, J. S. S. Damste, and S. Schouten (2004), A novel proxy for terrestrial organic matter in sediments based on branched and isoprenoid tetraether lipids, Earth and Planetary Science Letters, 224(1- 2), 107-116. 142 Hubeny, J. B., J. W. King, and M. Cantwell (2008), Anthropogenic influences on estuarine sedimentation and ecology: examples from the varved sediments of the Pettaquamscutt River Estuary, Rhode Island, Journal of Paleolimnology, 41(2), 297-314. Hughes, M. K. (2010), Dendroclimatology in High-Resolution Paleoclimatology, in Dendroclimatology: Progress and Prospects, edited by M. K. Hughes, Swetnam, W., and Diaz, H.F., Springer, London. Hurrell, J. W. (1995), DECADAL TRENDS IN THE NORTH-ATLANTIC OSCILLATION - REGIONAL TEMPERATURES AND PRECIPITATION, Science, 269(5224), 676-679. Imbrie, J., et al. (1992), ON THE STRUCTURE AND ORIGIN OF MAJOR GLACIATION CYCLES 1. LINEAR RESPONSES TO MILANKOVITCH FORCING, Paleoceanography, 7(6), 701-738. IPCC (2007), Summary for Policymakers, in Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, edited by S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Jansen, E., J. Overpeck, K.R. Briffa, J.-C. Duplessy, F. Joos, V. Masson-Delmotte, D. Olago, B. Otto-Bliesner, W.R. Peltier, S. Rahmstorf, R. Ramesh, D. Raynaud, D. Rind, O. Solomina, R. Villalba and D. Zhang (2007), Palaeoclimate, in Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, edited by S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Jones, P. D., and M. E. Mann (2004), Climate over past millennia, Rev Geophys, 42(2), RG2002. Keigwin, L. D. (1996), The Little Ice Age and Medieval warm period in the Sargasso Sea, Science, 274(5292), 1504-1508. Kerr, R. A. (2000), A North Atlantic climate pacemaker for the centuries, Science, 288(5473), 1984-1986. Kobashi, T., K. Kawamura, J. P. Severinghaus, J. M. Barnola, T. Nakaegawa, B. M. Vinther, S. J. Johnsen, and J. E. Box (2011), High variability of Greenland surface temperature over the past 4000 years estimated from trapped air in an ice core, Geophysical Research Letters, 38. Komor, S. C., and H. W. Anderson (1993), NITROGEN ISOTOPES AS INDICATORS OF NITRATE SOURCES IN MINNESOTA SAND-PLAIN AQUIFERS, Ground Water, 31(2), 260-270. Krumholz, J. (2012), Spatial and temporal patterns in nutrient standing stock and mass- balance in response to load reductions in a temperate estuary, University of Rhode Island, Narragansett. 143 Lee, K. E., B. K. Khim, S. Otosaka, and S. Noriki (2011), Sediment trap record of alkenones from the East Sea (Japan Sea), Organic Geochemistry, 42(3), 255-261. Levitus, S. (1982), Climatological Atlas of the World Ocean, 173 pp., U.S. Government Printing Office, Washington D.C. Levitus, S., J. Antonov, and T. Boyer (2005), Warming of the world ocean, 1955-2003, Geophysical Research Letters, 32(2). Longo, W. M., J. T. Dillon, R. Tarozo, J. M. Salacup, and Y. Huang (2013), Unprecedented separation of long chain alkenones from gas chromatography with a poly(trifluoropropylmethylsiloxane) stationary phase, Organic Geochemistry, 65(0), 94- 102. Lough, J. M. (personal communication), edited by M. Mann. Mann, M. E., Z. H. Zhang, M. K. Hughes, R. S. Bradley, S. K. Miller, S. Rutherford, and F. B. Ni (2008), Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia, Proc Natl Acad Sci U S A, 105(36), 13252-13257. Mann, M. E., Z. Zhang, S. Rutherford, R. S. Bradley, M. K. Hughes, D. Shindell, C. Ammann, G. Faluvegi, and F. Ni (2009), Global signatures and dynamical origins of the Little Ice Age and Medieval Climate Anomaly, Science, 326(5957), 1256-1260. Marali, S., M. Wisshak, M. L. Correa, and A. Freiwald (2013), Skeletal microstructure and stable isotope signature of three bathyal solitary cold-water corals from the Azores, Palaeogeogr Palaeocl, 373, 25-38. Marlowe, I. T., S. C. Brassell, G. Eglinton, and J. C. Green (1990), LONG-CHAIN ALKENONES AND ALKYL ALKENOATES AND THE FOSSIL COCCOLITH RECORD OF MARINE-SEDIMENTS, Chem Geol, 88(3-4), 349-375. Marshall, H. G. (1994), Chesapeake Bay phytoplankton: 1 Composition, Procedings of the Biological Society of Washington, 107, 573-585. Martin, A., W. Prell, and D. Murray (2013), The spatial and temporal distribution of benthic foraminifera in Narragansett Bay, paper presented at New England Estuarine Research Society Spring Meeting, Portland, ME. McNeely, R., A. S. Dyke, and J. R. Southon (2006), Canadian marine reservoir ages, preliminary data assessment, Open File 5049, pp. 3. Mercer, J. (2002), Extending the Spatial and Temporal Applications of the Uk37 Index as a Marine Paleotemperature Proxy, 129 pp, Dartmouth College, Hanover, New Hampshire. Mercer, J., M. Zhao, and S. Colman (2005), Seasonal variations of alkenones and Uk37 in the Chesapeake Bay water column, Estuarine, Coastal and Shelf Science, 63(4), 675- 682. 144 MERCINA (2001), Oceanographic Responses to Climate in the Northwest Atlantic, Oceanography, 14, 76-82. Moberg, A., D. M. Sonechkin, K. Holmgren, N. M. Datsenko, and W. Karlen (2005), Highly variable Northern Hemisphere temperatures reconstructed from low- and high- resolution proxy data, Nature, 433(7026), 613-617. Müller, P. J., G. Kirst, G. Ruhland, I. von Storch, and A. Rosell-Melé (1998), Calibration of the alkenone paleotemperature index U37K′ based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S), Geochimica et Cosmochimica Acta, 62(10), 1757-1772. Murphy, M. T. J. (1969), Analytical Methods, in Organic Geochemistry, edited by G. Eglinton and M. T. J. Murphy, pp. 74-88, Springer, Berlin. Murray, D. W., W. L. Prell, C. E. Rincon, and E. Saarman (2007), Physical Property and Chemical Characteristics of Surface Sediment Grab Samples from Narragansett Bay and the Providence and Seekonk Rivers, a summary of the Brown University Narragansett Bay Sediment Project (BUNBSP). Narragansett Bay Estuary Program Report, NBEP-07-127Rep. Nakamura, H., K. Sawada, H. Araie, I. Suzuki, and Y. Shiraiwa (2014), Long chain alkenes, alkenones and alkenoates produced by the haptophyte alga Chrysotila lamellosa CCMP1307 isolated from a salt marsh, Organic Geochemistry, 66(0), 90-97. Nanninga, H. J., and T. Tyrrell (1996), Importance of light for the formation of algal blooms by Emiliania huxleyi, Marine Ecology Progress Series, 136(1-3), 195-203. Narragansett Bay Project Nu-Shuttle Transect Data, edited. NASA-GISS Surface Temperature Analysis, edited. Nixon, S. W. (1991), A History of Metal Inputs to Narragansett Bay Rep., 69 pp, Narragansett Bay Project, Providence, RI. Nixon, S. W. (1997), Prehistoric nutrient inputs and productivity in Narragansett Bay, Estuaries, 20(2), 253-261. Nixon, S. W., B. A. Buckley, S. L. Granger, L. A. Harris, A. J. Oczkowski, R. W. Fulweiler, and L. W. Cole (2008), Nitrogen and Phosphorus Inputs to Narragansett Bay: Past, Present, and Future, in Science for Ecosystem-based Management: Narragansett Bay in the 21st Century, edited by A. Desbonnet and B. A. Costa-Pierce, pp. 101-175, Springer, New York, NY. Nixon, S. X., S. Granger, B. A. Buckley, M. Lamont, and B. Rowell (2004), A one hundred and seventeen year coastal water temperature record from Woods Hole, Massachusetts, Estuaries, 27(3), 397-404. Nyberg, J., B. A. Malmgren, A. Kuijpers, and A. Winter (2002), A centennial-scale variability of tropical North Atlantic surface hydrography during the late Holocene, Palaeogeogr Palaeocl, 183(1-2), 25-41. 145 Ono, M., K. Sawada, Y. Shiraiwa, and M. Kubota (2012), Changes in alkenone and alkenoate distributions during acclimatization to salinity change in Isochrysis galbana: Implication for alkenone-based paleosalinity and paleothermometry, Geochemical Journal, 46(3), 235-247. Oviatt, C. A. (2008), Impacts of Nutrients on Narragansett Bay Productivity: A Gradient Approach, in Science for Ecosystem-based Management, Narragansett Bay in the 21st Century, edited by A. Desbonnet, pp. 523-543, Springer, New York. Oviatt, C. A., and S. Nixon (1974), Sediment Resuspension and Deposition in Narragansett Bay, Estuarine and Coastal Marine Science, 3, 201-217. Parshall, T., D. R. Foster, E. Faison, D. MacDonald, and B. C. S. Hansen (2003), Long- term history of vegetation and fire in pitch pine-oak forests on Cape Cod, Massachusetts, Ecology, 84(3), 736-748. Pastore, C. L. (2011), From Sweeetwater to Seawater: an Environmental History fo Narragansett Bay, 1636-1849, Univeristy of New Hampshire. Patterson, W. A., and K. E. Sassman (1988), Indian Fires in the prehistory of New England, in Holocene Human Ecology in Northeastern America, edited by G. P. Nichols, pp. 107-136, Plenum Press, New York. Pesch, C. E., M. A. Schumchenia, M. A. Charpertier, and M. C. Pelletier (2012), Imprint of the Past: Ecological History of Greenwich Bay, Rhode Island. EPA/600/R-12/050, United States Environmental Protection Agency, Office of Research and Development, National Heath and Environmental Efftects Laboratory, Atlantic Ecology DivisionRep., Narragansett, RI. Peterse, F., J. H. Kim, S. Schouten, D. K. Kristensen, N. Koc, and J. S. S. Damste (2009), Constraints on the application of the MBT/CBT palaeothermometer at high latitude environments (Svalbard, Norway), Organic Geochemistry, 40(6), 692-699. Pickart, R. S., T. K. McKee, D. J. Torres, and S. A. Harrington (1999), Mean structure and interannual variability of the slopewater system south of Newfoundland, Journal of Physical Oceanography, 29(10), 2541-2558. Prahl, F. G., and S. G. Wakeham (1987), Calibration of Unsaturation Patterns in Long- Chain Ketone Compositions for Paleotemperature Assessment, Nature, 330(6146), 367- 369. Prahl, F. G., L. A. Muehlhausen, and D. L. Zahnle (1988), Further evaluation of long- chain alkenones as indicators of paleoceanographic conditions, Geochimica et Cosmochimica Acta, 52(9), 2303-2310. Prahl, F. G., C. H. Pilskaln, and M. A. Sparrow (2001), Seasonal record for alkenones in sedimentary particles from the Gulf of Maine, Deep-Sea Res. Part I-Oceanogr. Res. Pap., 48(2), 515-528. Prahl, F. G., G. V. Wolfe, and M. A. Sparrow (2003), Physiological impacts on alkenone paleothermometry, Paleoceanography, 18(2). 146 Prahl, F. G., B. N. Popp, D. M. Karl, and M. A. Sparrow (2005), Ecology and biogeochemistry of alkenone production at Station ALOHA, Deep-Sea Res. Part I- Oceanogr. Res. Pap., 52(5), 699-719. Prahl, F. G., J. F. Rontani, N. Zabeti, S. E. Walinsky, and M. A. Sparrow (2010), Systematic pattern in U37K′ – Temperature residuals for surface sediments from high latitude and other oceanographic settings, Geochimica et Cosmochimica Acta, 74(1), 131-143. Quinn, T. M., F. W. Taylor, and T. J. Crowley (1993), A 173 YEAR STABLE-ISOTOPE RECORD FROM A TROPICAL SOUTH-PACIFIC CORAL, Quaternary Science Reviews, 12(6), 407-418. Ran, L., H. C. Jiang, K. L. Knudsen, and J. Eiriksson (2010), Diatom-based reconstruction of palaeoceanographic changes on the North Icelandic shelf during the last millennium, Palaeogeogr Palaeocl, 302(1-2), 109-119. Rhode Island Department of Environmental Management (2012), Narragansett Bay Fixed-Site Monitoring Network (NBFSMN), edited, p. Data available at www.dem.ri.gov/bart, Office of Water Resources. Richey, J. N., R. Z. Poore, B. P. Flower, and T. Quinn (2007), 1400 yr multiproxy record of climate variability from the northern Gulf of Mexico, Geology, 35(5), 423-426. RIDEM (2003), The Greenwich Bay Fish Kill – August 2003: Causes, Impacts, and Responses, edited by Rhode Island Department of Environmental Management. Rosell-Mele, A. (1998), Interhemispheric appraisal of the value of alkenone indices as temperature and salinity proxies in high-latitude locations, Paleoceanography, 13(6), 694-703. Saarman, E., W. P. Prell, D. W. Murray, and C. F. Deacutis (2008), Summer Bottom Water Dissolved Oxygen in Upper Narragansett Bay, in Science for Ecosystem-Based Management, edited by A. a. C.-P. Desbonnet, B.A., pp. 325-349, Springer, New York, NY. Salacup, J., T. D. Herbert, and W. Prell (in prep-a), Alkenones in Narragansett Bay (RI) reveal and high-resolution record of North Atlantic Common Era climate. Salacup, J., J. Farmer, T. D. Herbert, and W. L. Prell (in prep-b), Alkenone production with open ocean characteristics in a near-shore setting: Narragansett Bay, R.I. Salacup, J., S. Theroux, T. D. Herbert, and W. L. Prell (in prep-c), A four-year long time- series of water column alkenones and 18S rRNA suggests Uk’37 is a viable proxy in Narragansett Bay, RI. Schouten, S., E. C. Hopmans, and J. S. S. Damste (2013), The organic geochemistry of glycerol dialkyl glycerol tetraether lipids: A review, Organic Geochemistry, 54, 19-61. Schouten, S., E. C. Hopmans, E. Schefuss, and J. S. S. Damste (2002), Distributional variations in marine crenarchaeotal membrane lipids: a new tool for reconstructing 147 ancient sea water temperatures?, Earth and Planetary Science Letters, 204(1-2), 265- 274. Schulz, H.-H., Schoener, A., and Emeis, K. C. (2000), Long-chain alkenone patterns in the Baltic Sea an ocean-freshwater transition, Geochimica et Cosmochimica Acta, 64, 469-477. Schwab, V. F., and J. P. Sachs (2011), Hydrogen isotopes in individual alkenones from the Chesapeake Bay estuary, Geochimica et Cosmochimica Acta, 75(23), 7552-7565. Sejrup, H. P., H. Haflidason, and J. T. Andrews (2011), A Holocene North Atlantic SST record and regional climate variability, Quaternary Science Reviews, 30(21-22), 3181- 3195. Seki, O., T. Nakatsuka, K. Kawamura, S. I. Saitoh, and M. Wakatsuchi (2007), Time- series sediment trap record of alkenones from the western Sea of Okhotsk, Mar Chem, 104(3-4), 253-265. Shearman, R. K., and S. J. Lentz (2010), Long-Term Sea Surface Temperature Variability along the U.S. East Coast, Journal of Physical Oceanography, 40(5), 1004- 1017. Shindell, D. T., G. A. Schmidt, M. E. Mann, D. Rind, and A. Waple (2001), Solar forcing of regional climate change during the maunder minimum, Science, 294(5549), 2149- 2152. Sicre, M. A., E. Bard, U. Ezat, and F. Rostek (2002), Alkenone distributions in the North Atlantic and Nordic sea surface waters, Geochem. Geophys. Geosyst., 3. Sicre, M. A., J. Jacob, U. Ezat, S. Rousse, C. Kissel, P. Yiou, J. Eiriksson, K. L. Knudsen, E. Jansen, and J. L. Turon (2008), Decadal variability of sea surface temperatures off North Iceland over the last 2000 years, Earth and Planetary Science Letters, 268(1-2), 137-142. Sikes, E. L., T. O'Leary, S. D. Nodder, and J. K. Volkman (2005), Alkenone temperature records and biomarker flux at the subtropical front on the chatham rise, SW Pacific Ocean, Deep-Sea Res. Part I-Oceanogr. Res. Pap., 52(5), 721-748. Smith, R. W., T. S. Bianchi, and X. Li (2012), A re-evaluation of the use of branched GDGTs as terrestrial biomarkers: Implications for the BIT Index, Geochimica et Cosmochimica Acta, 80(0), 14-29. Sweeney, R. E., and I. R. Kaplan (1980), Natural abundances of 15N as a source indicator for near-shore marine sedimentary and dissolved nitrogen, Mar Chem, 9(2), 81- 94. Taylor, A. H., and J. A. Stepens (1998), The North Atlantic oscillation and the latitude of the Gulf Stream, Tellus A, 50(1), 134-142. 148 Ternois, Y., M. A. Sicre, A. Boireau, M. H. Conte, and G. Eglinton (1997), Evaluation of long-chain alkenones as paleo-temperature indicators in the Mediterranean Sea, Deep- Sea Res. Part I-Oceanogr. Res. Pap., 44(2), 271-286. Theissen, K. M., D. A. Zinniker, J. M. Moldowan, R. B. Dunbar, and H. D. Rowe (2005), Pronounced occurrence of long-chain alkenones and dinosterol in a 25,000-year lipid molecular fossil record from Lake Titicaca, South America, Geochimica et Cosmochimica Acta, 69(3), 623-636. Thompson, D. W. J., and J. M. Wallace (1998), The Arctic Oscillation signature in the wintertime geopotential height and temperature fields, Geophysical Research Letters, 25(9), 1297-1300. Tierney, J. E., and J. M. Russell (2009), Distributions of branched GDGTs in a tropical lake system: Implications for lacustrine application of the MBT/CBT paleoproxy, Organic Geochemistry, 40(9), 1032-1036. Ting, M. F., Y. Kushnir, R. Seager, and C. H. Li (2009), Forced and Internal Twentieth- Century SST Trends in the North Atlantic, J Climate, 22(6), 1469-1481. Toney, J. L., Y. S. Huang, S. C. Fritz, P. A. Baker, E. Grimm, and P. Nyren (2010), Climatic and environmental controls on the occurrence and distributions of long chain alkenones in lakes of the interior United States, Geochimica et Cosmochimica Acta, 74(5), 1563-1578. Toney, J. L., S. Theroux, R. A. Andersen, A. Coleman, L. A. Amaral-Zettler, and Y. Huang (2012), Culturing the first 37:4 predominanat lacustrine haptophyte: Geochemical, biochemical, and genetic implications, Geochimica et Cosmochimica Acta, 78(51-64). Trouet, V., J. D. Scourse, and C. C. Raible (2012), North Atlantic storminess and Atlantic Meridional Overturning Circulation during the last Millennium: Reconciling contradictory proxy records of NAO variability, Global Planet Change(0). Trouet, V., J. Esper, N. E. Graham, A. Baker, J. D. Scourse, and D. C. Frank (2009), Persistent Positive North Atlantic Oscillation Mode Dominated the Medieval Climate Anomaly, Science, 324(5923), 78-80. Versteegh, G. J. M., R. Riegman, J. W. de Leeuw, and J. H. F. Jansen (2001), U(37)(K ')values for Isochrysis galbana as a function of culture temperature, light intensity and nutrient concentrations, Organic Geochemistry, 32(6), 785-794. Visbeck, M. H., J. W. Hurrell, L. Polvani, and H. M. Cullen (2001), The North Atlantic Oscillation: past, present, and future, Proc Natl Acad Sci U S A, 98(23), 12876-12877. Volkman, J. K., S. M. Barrett, S. I. Blackburn, and E. L. Sikes (1995), Alkenones in Gephyrocapsa-Oceanica - Implications for Studies of Paleoclimate, Geochimica et Cosmochimica Acta, 59(3), 513-520. 149 Volkman, J. K., S. W. Jeffrey, P. D. Nichols, G. I. Rogers, and C. D. Garland (1989), Fatty-Acid and Lipid-Composition of 10 Species of Microalgae Used in Mariculture, J Exp Mar Biol Ecol, 128(3), 219-240. Waller, J. N. (2000), Late woodland settlement and subsistence in southern New England revisited: The evidence from coastal Rhode Island, North American Archaeologist, 21(2), 139-153. Wanamaker, A. D., K. J. Kreutz, B. R. Schone, and D. S. Introne (2011), Gulf of Maine shells reveal changes in seawater temperature seasonality during the Medieval Climate Anomaly and the Little Ice Age, Palaeogeogr Palaeocl, 302(1-2), 43-51. Webb, T. (1986), IS VEGETATION IN EQUILIBRIUM WITH CLIMATE - HOW TO INTERPRET LATE-QUATERNARY POLLEN DATA, Vegetatio, 67(2), 75-91. Weijers, J. W. H., S. Schouten, O. C. Spaargaren, and J. S. S. Damste (2006a), Occurrence and distribution of tetraether membrane lipids in soils: Implications for the use of the TEX86 proxy and the BIT index, Organic Geochemistry, 37(12), 1680-1693. Weijers, J. W. H., S. Schouten, J. C. van den Donker, E. C. Hopmans, and J. S. S. Damste (2007), Environmental controls on bacterial tetraether membrane lipid distribution in soils, Geochimica et Cosmochimica Acta, 71(3), 703-713. Weijers, J. W. H., S. Schouten, E. C. Hopmans, J. A. J. Geenevasen, O. R. P. David, J. M. Coleman, R. D. Pancost, and J. S. S. Damste (2006b), Membrane lipids of mesophilic anaerobic bacteria thriving in peats have typical archaeal traits, Environ Microbiol, 8(4), 648-657. Weijers, J. W. H., E. Panoto, J. van Bleijswijk, S. Schouten, W. I. C. Rijpstra, M. Balk, A. J. M. Stams, and J. S. S. Damste (2009), Constraints on the Biological Source(s) of the Orphan Branched Tetraether Membrane Lipids, Geomicrobiol J, 26(6), 402-414. Yamamoto, M., A. Shimamoto, T. Fukuhara, H. Naraoka, Y. Tanaka, and A. Nishimura (2007), Seasonal and depth variations in molecular and isotopic alkenone composition of sinking particles from the western North Pacific, Deep-Sea Res. Part I-Oceanogr. Res. Pap., 54(9), 1571-1592. Zachos, J., M. Pagani, L. Sloan, E. Thomas, and K. Billups (2001), Trends, rhythms, and aberrations in global climate 65 Ma to present, Science, 292(5517), 686-693. Zhu, C., J. W. H. Weijers, T. Wagner, J. M. Pan, J. F. Chen, and R. D. Pancost (2011), Sources and distributions of tetraether lipids in surface sediments across a large river- dominated continental margin, Organic Geochemistry, 42(4), 376-386. Zink, K. G., D. Leythaeuser, M. Melkonian, and L. Schwark (2001), Temperature dependency of long-chain alkenone distributions in Recent to fossil limnic sediments and in lake waters, Geochimica et Cosmochimica Acta, 65(2), 253-265. 150