Holocene Environmental Changes and Climate Variability in the Eastern

Dissertations in Physical Geography No. 2 Christos Katrantsiotis This thesis aims to contribute to a better understanding of climate evolution and the related drivers in the central-eastern Mediterranean Holocene environmental changes over the last 6000 years, a period of centennial-millennial climate changes whose nature is not well-understood. The study area is the Peloponnese peninsula, SW Greece, characterized by high climatic and climate variability in the heterogeneity, broad spectrum of natural archives and rich archaeological history. The identification of climate signal in fossil Eastern Mediterranean

records, however, might be complicated in this region due to the Holocene environmental changes and climate variability in Eastern Mediterraneanthe interaction of various environmental factors. The development of a Multiproxy sediment records from the Peloponnese peninsula, SW Greece denser network of multiproxy records can help to optimize the interpretative accuracy and identify synchronous changes attributed to a common climate mechanism. Here, a multiproxy approach mainly Christos Katrantsiotis based on diatom and biomarker analyses is applied on radiocarbon- dated sediment cores from the Agios Floros fen and the Gialova Lagoon in SW Peloponnese and the Ancient Lake Lerna in NE Peloponnese. The results highlight the interaction between climate, tectonics and human activities and reveal variations in the precipitation/temperature gradient over the Peloponnese, attributed to the interplay between different atmospheric circulation patterns. This interplay can further be elucidated by integrating the growing network of proxy records with paleoclimate modelling, which is also crucial to improve climate forecasting capabilities.

Christos Katrantsiotis holds BSc in Geology and MSc in Climate Science and Quaternary Geology. His main interest is the reconstruction of paleoenvironmental and paleoclimatic changes using diatoms and biomarker lipids in sediment cores. His current research focuses on the Mediterranean region and Northern Scandinavia. He also has special interest in diatom morphology and taxonomy.

ISBN 978-91-7797-664-6 ISSN 2003-2358

Department of Physical Geography

Doctoral Thesis in Physical Geography at Stockholm University, Sweden 2019 Holocene environmental changes and climate variability in the Eastern Mediterranean Multiproxy sediment records from the Peloponnese peninsula, SW Greece Christos Katrantsiotis Academic dissertation for the Degree of Doctor of Philosophy in Physical Geography at Stockholm University to be publicly defended on Friday 14 June 2019 at 13.00 in De Geersalen, Geovetenskapens hus, Svante Arrhenius väg 14.

Abstract This thesis presents multiproxy reconstructions of the mid to late Holocene climate and environmental changes in the Peloponnese peninsula, SW Greece. The combined dataset consists of diatom, biomarker and X-ray fluorescence spectrometry (XRF) elemental data in radiocarbon-dated sediment cores taken from the Agios Floros fen and the Gialova Lagoon in SW Peloponnese and the Ancient Lake Lerna in NE Peloponnese. Overall, the results highlight the complex interaction between climate, tectonics and human activities in the landscape development and further reveal changes in the W-E precipitation/temperature gradient over the peninsula connected to shifts in the large-scale atmospheric circulation patterns. The Agios Floros study provides a 6000-year hydrological record based on diatoms and hydrogen isotopic (δD) analysis of aquatic plant-derived n-C23 alkanes. The records indicate two decadal-long periods of deep water conditions at ca 5700 and 5300 cal BP, largely attributed to local tectonic processes and the hydrological anomalies of the nearby karst springs. A period of intermediate water level at ca 4600 cal BP is dominated by the new fossil species Cyclotella paradistinguenda described in this thesis. The gradual development of a fen at ca 4500 cal BP is attributed to a combination of human activities and drier conditions, the latter culminating in SW Peloponnese mainly after ca 4100 cal BP. From ca 2800 cal BP and onwards, there is evidence for flooding events probably related to marked rainfall seasonality. The n-alkane δD profiles and XRF data analyzed in the Gialova core co-vary with each other indicating a common climate signal during the last 3600 years, which resembles the Agios Floros record. The n-alkane δ13C values show high contribution of aquatic vegetation to sedimentary organic matter during wet/cold periods. The n-alkane δD signals from the Lake Lerna also exhibit a similar pattern to each other providing further evidence for precipitation/temperature changes over the last 5000 years. Comparison of the δD records reveals sometimes similar and sometimes opposing signals between NE and SW Peloponnese, which can be attributed to the relative dominance of high latitude and low latitude atmospheric patterns over the peninsula. The records show wet conditions at ca 5000-4600 cal BP likely associated with the weakening of the Hadley circulation. High humidity is also evident at ca 4500-4100, ca 3000-2600 (more unstable in SW) and after ca 700 cal BP with drier conditions at ca 4100-3900 and ca 1000-700 cal BP. These periods correspond to regional climate changes, when the North Atlantic Oscillation (NAO) likely exerted the main control with NAO (+) creating conditions of reduced moisture. A NE-SW climate see-saw with drier conditions in NE Peloponnese is evident at ca 4600-4500, ca 3200, ca 2600-1800 and ca 1200-1000 cal BP and a reversal at ca 3900-3300 ca 3200-3000 and ca 1800-1300 cal BP. The dipole pattern is likely driven by shifts in the North Sea–Caspian Atmospheric pattern (NCP), with NCP (+) leading to wetter and colder conditions in NE Peloponnese. The opposing signal can also be explained by changes in summer temperatures driven by the Asian monsoon intensity. Strong monsoonal periods coincide with cool summers in Lerna, due to the northerly winds (Etesians), in contrast to SW Peloponnese, located on the lee side of the mountain and most affected by the large-scale air subsidence.

Keywords: Mediterranean, Greece, Peloponnese, Holocene, sediments, diatoms, n-alkanes, stable isotopes, tectonics, climate variability, monsoons, NAO, NCP.

Stockholm 2019 http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-167747

ISBN 978-91-7797-664-6 ISBN 978-91-7797-665-3 ISSN 2003-2358

Department of Physical Geography

Stockholm University, 106 91 Stockholm

HOLOCENE ENVIRONMENTAL CHANGES AND CLIMATE VARIABILITY IN THE EASTERN MEDITERRANEAN

Christos Katrantsiotis

Holocene environmental changes and climate variability in the Eastern Mediterranean

Multiproxy sediment records from the Peloponnese peninsula, SW Greece

Christos Katrantsiotis ©Christos Katrantsiotis, Stockholm University 2019

ISBN print 978-91-7797-664-6 ISBN PDF 978-91-7797-665-3 ISSN 2003-2358

Paper I © The Holocene, SAGE Publications Paper II © Diatom Research, Taylor & Francis Online Paper III © Journal of Quaternary Science, John Wiley & Sons, Inc. Paper IV © Quaternary Science Reviews, Elsevier B.V. Paper V © Global and Planetary Change, Elsevier B.V.

Cover illustration: Gialova Lagoon. Photo by Christos Katrantsiotis, 2016 Printed in Sweden by Universitetsservice US-AB, Stockholm 2019 Distributor: Department of Physical Geography, Stockholm University Abstract

This thesis presents multiproxy reconstructions of the mid to late Holocene climate and environmental changes in the Peloponnese peninsula, SW Greece. The combined dataset consists of diatom, biomarker and X-ray fluorescence spectrometry (XRF) elemental data in radiocarbon-dated sediment cores taken from the Agios Floros fen and the Gialova Lagoon in SW Peloponnese and the Ancient Lake Lerna in NE Peloponnese. Overall, the results highlight the complex interaction between climate, tectonics and human activities in the landscape development and further reveal changes in the W-E precipitation/temperature gradient over the peninsula connected to shifts in the large-scale atmospheric circulation patterns. The Agios Floros study provides a 6000-year hydrological record based on diatoms and hydrogen isotopic (δD) analysis of aquatic plant‐derived n‐C23 alkanes. The records indicate two decadal-long periods of deep water conditions at ca 5700 and 5300 cal BP, largely attributed to local tectonic processes and the hydrological anomalies of the nearby karst springs. A period of intermediate water level at ca 4600 cal BP is dominated by the new fossil species Cyclotella paradistinguenda described in this thesis. The gradual development of a fen at ca 4500 cal BP is attributed to a combination of human activities and drier conditions, which culminate in SW Peloponnese mainly after ca 4100 cal BP. From ca 2800 cal BP and onwards, there is evidence for flooding events probably related to marked rainfall seasonality. The n-alkane δD profiles and XRF data analyzed in the Gialova core co-vary with each other indicating a common climate signal during the last 3600 years, which resembles the Agios Floros record. The n-alkane δ13C values show high contribution of aquatic vegetation to sedimentary organic matter during wet/cold periods. The n-alkane δD signals from the Lake Lerna also exhibit a similar pattern to each other providing further evidence for precipitation/temperature changes over the last 5000 years. Comparison of δD records reveals sometimes similar and sometimes opposing signals between NE and SW Peloponnese, which can be attributed to the relative dominance of high-latitude and low-latitude atmospheric patterns over the peninsula. The records show wet conditions at ca 5000-4600 cal BP likely associated with the weakening of the Hadley circulation. High humidity is also evident at ca 4500-4100, ca 3000-2600 (more unstable in SW) and after ca 700 cal BP with drier conditions at ca 4100-3900 and ca 1000-700 cal BP. These periods correspond to regional climate changes, during which the North Atlantic Oscillation (NAO) likely exerted the main control with NAO (+) creating conditions of reduced moisture. A NE-SW climate see-saw with drier conditions in NE Peloponnese is evident at ca 4600-4500, ca 3200, ca 2600-1800 and ca 1200-1000 cal BP and a reversal at ca 3900-3300 ca 3200-3000 and ca 1800-1300 cal BP. The dipole pattern is likely driven by shifts in the North Sea–Caspian Atmospheric pattern (NCP), with NCP (+) leading to wetter and colder conditions in NE Peloponnese. The opposing signal can also be explained by changes in summer temperatures driven by the Asian monsoon intensity. Strong monsoonal periods coincide with cool summers in Lerna, due to the northerly winds (the Etesians), in contrast to SW Peloponnese, located on the lee side of the mountain and most affected by the large-scale air subsidence.

Sammanfattning

Denna avhandling presenterar s.k. multiproxyrekonstruktioner av klimatvariationer under mitt- och sen- holocen på Peloponnesos-halvön, sydvästra Grekland. Rekonstruktionerna baseras på diatoméer, biomarkörer och XRF-data i kol14-daterade sedimentkärnor från Agios Floros-kärret och Gialova- lagunen på sydvästra Peloponnesos och den nu igenvuxna Lerna-sjön på nordöstra Peloponnesos. Resultaten belyser samspelet mellan klimat, tektonik och mänskliga aktiviteter i landskapsutvecklingen och påvisar dessutom förändringar i den väst-östliga nederbörds/temperaturgradienten över halvön i samband med förändringar i storskaliga atmosfärsmönster. Agios Floros-studien omfattar de senaste ca 6000 åren och med hjälp av diatoméanalys har paleohydrologiska förändringar rekonstruerats. Rekonstruktionen stöds av analyser av deuterium (δD) och n-C23-alkaner från akvatiska växter. Båda analysmetoderna anger två decennielånga perioder med djupa vattenförhållanden vid ca 5700 och 5300 kal BP, vilket huvudsakligen kan förklaras med tektoniska processer och hydrologiska anomalier i närliggande karstkällor. En period med intermediära vattennivåer vid ca 4600 kal BP domineras av en ny fossil diatoméart, Cyclotella paradistinguenda som beskrivs för första gången i denna avhandling. Agios Floros började utvecklas till ett kärr ca 4500 kal BP på grund av en kombination av mänskliga aktiviteter och de torrare förhållanden som rådde på sydvästra Peloponnesos, huvudsakligen efter ca 4100 kal BP. Med början ca 2800 kal BP finns det bevis för översvämningshändelser som sannolikt är förknippade med mer markerade säsongsvariationer i nederbörden. δD-analyser av n-alkaner och XRF-data som analyserats i kärnan från Gialova-lagunen samvarierar med varandra, vilket indikerar en gemensam klimatsignal under de senaste 3600 åren, liknande den som data från Agios Floros visar. n-Alkan-värden från δ13C från Gialova-kärnan visar högt inslag av vattenlevande vegetation i det organiska materialet under våta och kalla tidsperioder. δD-variationer i n- alkanerna från Lake Lerna visar ett nästan identiskt mönster som ger ytterligare bevis för nederbörds- och temperaturförändringar under de senaste 5000 åren. Jämförelser av variationer i δD visar ibland motsatta och ibland liknande signaler mellan nordöstra och sydvästra Peloponnesos, vilket kan förklaras med den relativa dominansen mellan atmosfäriska mönster på hög latitud eller låg latitud över halvön. δD-analyser visar fuktiga förhållanden vid ca 5000- 4600 kal BP som troligen är förknippade med försvagningen av Hadley-cirkulationen. Blötare förhållanden är också uppenbar vid ca 4500-4100, ca 3000-2600 (mer instabil i sydväst) och efter ca 700 kal BP med torrare perioder vid ca 4100-3900 och ca 1000-700 kal BP. Under dessa perioder med regionala klimatförändringar var den nordatlantiska oscillationen (NAO) dominerande och en positiv NAO skapade förhållanden med reducerad fuktighet. Ett nordostligt-sydvästligt motsatsförhållande med torra förhållanden på nordöstra Peloponnesos påvisas vid ca 4600-4500, ca 3200, ca 2600-1800 och ca 1200-1000 kal BP med en omkastning vid ca 3900-3300 ca 3200-3000 and ca 1800-1300 cal BP. Dipolmönstret drivs sannolikt av förändringar i det s.k. Nordsjö-Kaspiska atmosfäriska mönstret (NCP), där ett positivt NCP-index leder till fuktigare och kallare förhållanden på nordöstra Peloponnesos. Den motsatta signalen kan även förklaras av de sommartemperaturförändringar som drivs av den asiatiska monsunintensiteten. Starka monsunperioder sammanfaller med svala somrar i Lerna, på grund av de nordliga s.k. etesiska vindarna, i motsats till sydvästra Peloponnesos, som är beläget på läsidan av bergen och är mer påverkad av sjunkande luftmassor.

List of papers

This doctoral thesis consists of this summary and the following five papers, which are referred to by their Roman numerals in the text.

I. Katrantsiotis, C., Norström, E., Holmgren, K., Risberg, J., Skelton, A., 2015. High-resolution environmental reconstruction in SW Peloponnese Greece covering c. the last 6000 years: Evidence from Agios-Floros fen. The Holocene 26, 188–204. https://doi.org/10.1177/0959683615596838

II. Katrantsiotis, C., Risberg, J., Norström, E., Holmgren, K., 2016. Morphological study of Cyclotella distinguenda and a description of a new fossil species Cyclotella paradistinguenda sp. nov. from the Agios Floros fen, SW Peloponnese, Greece in relation to other Cyclotella species. Diatom Research 31, 243–267. https://doi.org/10.1080/0269249X.2016.1211178

III. Norström, E., Katrantsiotis, C, Finné, M., Risberg, J., Smittenberg, R., Bjursäter, S., 2018. Biomarker hydrogen isotope composition (δD) as proxy for Holocene hydro-climatic change and seismic activity in SW Peloponnese, Greece. Journal of Quaternary Science 33, 563–574. https://doi.org/10.1002/jqs.3036

IV. Katrantsiotis, C., Kylander, M., Smittenberg, R., Yamoah, K.K.A., Hättestrand, M., Avramidis, P., Strandberg, N.A., Norström, E., 2018. Eastern Mediterranean hydroclimate reconstruction over the last 3600 years based on sedimentary n-alkanes, their carbon and hydrogen isotope composition and XRF data from the Gialova Lagoon, SW Greece. Quaternary Science Reviews 194, 77–93. https://doi.org/10.1016/j.quascirev.2018.07.008

V. Katrantsiotis, C., Norström, E., Smittenberg, R., Finné, M., Weiberg, E., Hättestrand, M., Avramidis, P., Wastegård S., 2019. Climate changes in the Eastern Mediterranean over the last 5000 years and their links to the high-latitude atmospheric patterns and Asian monsoons. Global and Planetary Change 175, 36–51. https://doi.org/10.1016/j.gloplacha.2019.02.001

Author contributions

The following people have contributed to this thesis: Elin Norström (EN), Karin Holmgren (KH), Stefan Wastegård (SW), Jan Risberg (JR), Rienk Smittenberg (RS), Alasdair Skelton (AS), Malin Kylander (MK), Martina Hättestrand (MH), Pavlos Avramidis (PA), Kweku Yamoah (KY), Martin Finné (MF), Erika Weiberg (EW), Nichola Strandberg (NS), and Stefan Bjursäter (SB). All co-authors are acknowledged for reading and commenting on the manuscripts.

I. Conceived and designed by CK. CK and EN led the field work planning and performance. CK described and subsampled the Agios Floros sediment core at the Peloponnese University, Greece. CK performed the diatom analysis (sample preparation, diatom identification and counting), produced the age-depth model, prepared the figures, made the interpretation and wrote the paper. CK and EN prepared samples for bulk geochemical analysis. MK contributed with XRF analysis of subsamples. AS helped out with the tectonic interpretation.

II. Conceived and designed by CK. CK carried out the light and scanning electron microscopic analysis, prepared the figures, described the new diatom species, made the interpretation and wrote the paper.

III. Conceived and designed by EN. CK and EN led the field work planning and performance. CK described and subsampled the Agios Floros sediment core at the Peloponnese University, Greece. EN performed the n-alkane extraction and quantification as well as carried out the compound-specific isotope analysis of n-alkanes. EN prepared the figures, made the interpretation and wrote the paper. CK wrote the introduction and the regional settings as well as contributed to the interpretation. SB prepared the topographical map of the Messenian plain.

IV. Conceived and designed by CK. CK and EN led the field work planning and performance. CK described and subsampled the Gialova sediment core at the Peloponnese University, Greece. CK performed the n-alkane extraction and quantification and compound-specific isotope analysis of n-alkanes, prepared and ran samples for XRF analysis with the assistance of MK, produced the age-depth model, prepared the figures, made the interpretation and wrote the paper. MK contributed with the XRF interpretation.

V. Conceived and designed by CK. CK, EN and MH led the field work planning and performance. CK and MH described and subsampled the Lerna sediment core at Stockholm University. CK, with the assistance of RS, performed the n-alkane extraction and quantification. CK carried out the compound-specific isotope analysis of n-alkanes, construction of the age-depth model, prepared the figures, made the interpretation and wrote the paper.

Table of Contents

1. Introduction...... 1 2. Thesis objectives ...... 2 3. Eastern Mediterranean climate ...... 2 3.1. Winter atmospheric circulation patterns ...... 3 3.2. Summer atmospheric circulation patterns ...... 3 4. The Peloponnese ...... 4 4.1. Topography and climate ...... 4 4.2. Geology and lanscape development...... 4 4.3. Human occupation...... 5 5. Description of sites ...... 5 5.1. Messenian plain/Agios Floros Fen ...... 5 5.2. Gialova Lagoon ...... 6 5.3. Lake Lerna ...... 7 6. Material and methods ...... 8 6.1. Field work...... 8 6.2. Proxy description and laboratory work ...... 8 6.2.1. Radiocarbon dating ...... 8 6.2.2. Siliceous microfossils ...... 9 6.2.3. Organic geochemistry ...... 10 6.2.4. XRF core scanning ...... 15 7. Results ...... 16 7.1. Paper I ...... 16 7.2. Paper II ...... 18 7.3. Paper III ...... 19 7.4. Paper IV...... 20 7.5. Paper V...... 21 8. Discussion ...... 22 8.1. Chronology and resolution ...... 22 8.2. Proxy data ...... 23 8.2.1. Diatom preservation ...... 23 8.2.2. The new diatom species and its affinity ...... 24 8.2.3. Origins of n-alkanes...... 25 8.2.4. Factors behind δD variability ...... 26 8.3. The role of tectonics in abrupt hydrological changes in the Agios Floros fen ..... 27 9. Regional climate reconstruction ...... 29 10. Drivers of Holocene climate variability ...... 33 10.1. High-latitude atmospheric patterns ...... 33 10.1.1. Physical mechanism ...... 33 10.2. Effects of African and Asian monsoons ...... 35 11. Conclusions ...... 36 12. Future perspectives ...... 37 13. Financial support ...... 38 14. Acknowledgements ...... 38 15. References ...... 39

Abbreviations

ACL Average chain length AMS Accelerator mass spectrometry C/N Carbon/Nitrogen ratio cal BP Calibrated years before present (BP = 1950) CONISS Constrained incremental sum-of-squares CPI Carbon preference index EA/WR East Atlantic/West Russia GC-MC Gas chromatography-mass spectrometry GC-IRMS Gas chromatograph-isotope ratio mass spectrometer IPCC Intergovernmental panel on climate change IRD Ice rafted debris ITCZ Intertropical Convergence Zone LIA Little Ice Age LM Light microscope m a.s.l. Meters above sea level MCA Medieval Climate Anomaly MO Mediterranean Oscillation NAO North Atlantic Oscillation NCP North Sea-Caspian Pattern OM Organic matter OSL Optically stimulated luminescence PCA Principal component analysis SEM Scanning electron microscope TN Total nitrogen TOC Total organic carbon VPDB Vienna Pee Dee Belemnite VSMOW Vienna Standard Mean Ocean Water XRF X-ray fluorescence δ13C Stable carbon isotope δD Stable hydrogen isotope δ18O Stable oxygen isotope

1. Introduction The Holocene is generally characterized by centennial- millennial changes whose nature and chronological The global climate has been experiencing significant boundaries are not well-defined. A large body of dataset warming during the last decades which has primarily from the Mediterranean region provides evidence of been attributed to anthropogenic greenhouse forcing complex spatial and temporal climatic variability with (IPCC, 2014). This climate change is spatially and clear patterns from east to west and from north to south temporally non-uniform and affects some regions more during the Holocene (e.g. Dormoy et al., 2009; Finné et than others, with variable shifts in temperatures and al., 2011; Roberts et al., 2012; Magny et al., 2013; precipitation patterns (Ji et al., 2014). There is an Peyron et al., 2013). In addition to this climatic ongoing debate on whether an increasing risk of heterogeneity, limitations in dating accuracy hamper droughts exists worldwide, or whether wet areas are regional comparisons and discussions regarding the getting wetter and the arid regions drier under global causes and the effects. All these factors highlight the warming (Dore, 2005; Feng and Zhang, 2015). need to improve both the spatial coverage and quality of Natural climate variability operates in a similar way the existing data sets. The establishment of a denser on all timescales (different impacts on different areas) network of proxy records around the Mediterranean can and is superimposed on human-induced climate change. contribute to our understanding of regional climatic This complicates the accurate detection and attribution differences within the basin and its underlying of human influence on climate. More knowledge of past mechanisms as well as the relationship between the climatic fluctuation is required to understand the nature climate, the physical environment and human actions. and rate of the current climate change, to distinguish The Peloponnese peninsula in SW Greece at the anthropogenic effects, and to accurately project climate most south-eastern point of Europe offers the system response to future changes (Jansen et al., 2007). opportunity to explore the presence of such regional Instrumental records span only the last centuries, a climatic differences and/or common patterns. The period too short to understand the natural climatic peninsula is characterized by a west-east rainfall/ variability. Natural archives such as lake sediments, temperature gradient attributed to its topography and stalagmites, ice-cores and tree rings offer the the influence of different atmospheric circulation opportunity to extend this knowledge over longer, patterns from the mid-latitude and the low-latitude geological timescales. These archives contain a variety regions. Therefore, comparison of climate records from of proxies related to different climate variables and can the western and eastern parts of the peninsula can reveal preserve information about these variables at different shifts in the large-scale teleconnections that affect the time-scales. Thus, paleoenvironmental studies have Mediterranean region. In addition, the Peloponnese been characterized as the ultimate source of our peninsula offers a broad spectrum of natural archives understanding of the nature and rate of past climatic (e.g. speleothems, lake sediments) allowing climate variability and anthropogenic effects (Bradley, 2014). reconstructions with high temporal and spatial The Mediterranean region is particularly sensitive resolution. The peninsula is also rich in archaeological to climate changes due to its semi-arid conditions, its findings with numerous well-preserved and precisely intermediate position between the mid-latitude and dated archaeological sites and is thus highly interesting subtropical atmospheric systems and its high climatic for studies on how climate changes have impacted heterogeneity characterized by altitude gradients, different societies in the past (Weiberg et al., 2016). microclimates and extreme seasonality (Lionello et al., Palaeoenvironmental studies in the Peloponnese 2006; Ulbrich et al., 2012). In addition, physico- have a long tradition from a variety of disciplines with geographical parameters such as orography, land-sea special attention on the relationship between climate, interaction and the Mediterranean Sea, a source of landscape changes and past settlements. These studies moisture and a heat reservoir, make the regional have been based on biological (pollen, foraminifera) response to large-scale climate forcing rather complex and geochemical records from lagoons and lakes (Kraft (Lionello et al., 2006; Giorgi and Lionello, 2008). The et al., 1980; Jahns, 1993; Papazisimou et al., 2005; Intergovernmental Panel for Climate Change (IPCC) Cundy et al., 2006; Lazarova et al., 2012; Heymann et has identified the Mediterranean as a climatic hot spot al., 2013; Unkel et al., 2014). However, limitations where most countries are already experiencing a high related to dating accuracy and proxy preservation have often hampered the interpretations. Oxygen isotopes frequency of droughts, water scarcity, forest fires and 18 increasing desertification (IPCC, 2014; Loizidou et al., (δ Ο) in speleothems have recently been applied in 2016). These conditions are predicted to be exacerbated paleoclimate studies from the Peloponnese (Boyd, leading to dire consequences for ecosystems and human 2015; Finné et al., 2014, 2017). Although speleothem societies (IPCC, 2014; Guiot and Cramer, 2016). records are of high quality and resolution, they often The pronounced sensitivity of the Mediterranean cover limited periods and do show occasional hiatuses. region to climate changes has driven intense research Despite the abundance of paleoclimatic data, proxy on both the study of future and past climates. Special interpretations in the Peloponnese are complicated by focus has been placed on the Holocene, a period which the interaction of various environmental factors, i.e. involves possible human response to climate change tectonics, sea level changes and human activities in the (Kaniewski et al., 2013; Knapp and Manning, 2016). landscape development. These factors might cause

1 similar variations in the proxy records, which can 2. Thesis objectives impede the unambiguous separation of climate from non-climatic signals. In this case, a multiproxy The overall objectives of the work presented in this approach derived from independent sources will PhD thesis are 1) to increase the spatial coverage of potentially help to discriminate between different Holocene paleoclimate data from the Peloponnese and environmental factors, and particularly to identify also 2) to improve the understanding of the Holocene synchronous changes in proxy records attributed to a climate changes in the central-eastern Mediterranean. common climate mechanism. In addition to this, the The specific objectives of this PhD thesis are to: comparison of records from different sites of the Peloponnese can enable the separation of local  establish continuous paleoclimatic records from SW environmental changes from regional climate signals. and NE Peloponnese based on diatom and In this thesis, a multi-proxy approach is applied on geochemical analyses of three sediment cores: two three radio-carbon dated sediment cores retrieved from records from SW Peloponnese covering the last three climatic sensitive ecosystems in SW and NE 6000 and 3600 years, respectively and one record Peloponnese. The locations of sites on SW-NE transect from NE Peloponnese spanning the last 5000 years. aims to establish a small network of paleoclimatic data  discriminate between various environmental factors from different climatic zones of the Peloponnese which influence the multiproxy records, and peninsula. This can further enable us to stress local specifically to discern climate-induced changes. climate differences and provide a more accurate  identify short-term events by establishing high assessment of the regional climate variability and its resolution records, and to understand the nature of connection to the large scale atmospheric patterns. those events and their relation to tectonic activities. The cores were analyzed for different biological and  evaluate the applicability of n-alkane δD data as a geochemical proxies: diatoms, the distribution and paleoclimate proxy in this particular environmental compound specific carbon (13C) and hydrogen (D) setting by means of comparing the δD records with isotope compositions of leaf-wax n-alkanes, bulk other independent proxies and regional records. geochemistry and X-ray fluorescence spectrometry  trace the sources of sedimentary organic matter (XRF) elemental data. Diatoms are single-cell algae (OM) in the studied sites, which can further aid with that exhibit high diversity and well-defined ecological more accurate interpretation of n-alkane δD data. preferences. Diatom studies are rare in Greece (Wilson  examine diatom morphology by means of light and et al., 2008; Jones et al., 2013) while elsewhere this scanning electron microscope (LM and SEM). method have successfully been used for paleo-  present the first comprehensive list of fossil diatoms environmental reconstructions (e.g. Gasse, 1987,1994; observed in Greece and also to contribute to the Hassan, 2013; Cvetkoska et al., 2014). Organic geo- Mediterranean database of diatom paleotaxa. chemical proxies, so-called biomarkers, are preserved The goals related to paleoclimate investigation are to: in sediments over thousands of years, and along with their isotopic signal, they are useful markers for past  reconstruct shifts in the W-E rainfall gradient over hydroclimatic conditions given their clear biological the Peloponnese by comparing the δD signals with origin (Peters et al., 2005). In Greece, this approach has each other and with other Peloponnesian records. been applied on marine sediments (e.g. Triantaphyllou  produce a wider climate reconstruction for the et al., 2009; Gogou et al., 2016) and terrestrial material eastern Mediterranean region by comparing the combined with compound specific isotope analysis Peloponnesian records with other regional studies. (Schemmel et al., 2016; Norström et al., 2017). The  understand the driving mechanisms behind the XRF method is increasingly being used in stratigraphic observed climate changes, and particularly to studies to investigate changes in elemental composition, assess linkages with the high latitude atmospheric attributed to various environmental processes in lake patterns and the African/Asian monsoons. ecosystems and catchments (Croudace et al., 2006).

The continuous multiproxy records of climatic variability in the Peloponnese presented in this thesis will fill important gaps in the existing hydroclimate 3. Eastern Mediterranean climate studies from the central-eastern Mediterranean region. The growing network of paleoclimate research will The Mediterranean climate is characterized by large further contribute to a more accurate understanding of seasonal variations with hot and dry summers and mild the natural climate processes, so that current climate /cold and wet winters. This variation is a result of the change can be placed better in a historical context. Such interaction between the Subtrophic High and the mid- an approach will further allow testing the performance latitude westerlies leading to a distinct sensitivity of the of climate models. If a model can correctly predict local climate to the variability of supraregional trends in the past, one may expect it to predict future atmospheric patterns. In addition, the Mediterranean trends accurately as well. This is vital to responsibly climatic conditions are influenced by the complexity of inform policy makers and citizens about the potential the terrain as well as the interaction between land-sea, future climate changes in this drought-prone region. and the Mediterranean Sea itself (Lionello et al., 2006).

3.1. Winter atmospheric circulation patterns Winter climatic conditions in the Eastern Mediterranean are mainly dominated by the interplay between the mid- and high-latitude atmospheric patterns. This includes variations in the location, strength and spatial extension of the Azores High, the Icelandic Low and also the Siberia High (Fig. 1) (Trigo et al., 2006; Küttel et al., 2011). The precipitation regime is largely associated with moisture rich air carried Fig. 1. Atmospheric patterns affecting the Mediterranean climate dynamics. The square by the low pressure systems depicts the Peloponnese in the Eastern Mediterranean (after Katrantsiotis et al., 2019). originating from the North Atlantic region and the western Mediterranean (Ulbrich Turkey resulting in lower (higher) temperatures (Feidas, et al., 2012). The path of these systems is influenced by 2017). During positive (negative) NCP phases, the areas the North Atlantic Oscillation (NAO) modes defined as exposed to northern (southern) maritime trajectories, the normalized pressure difference between Iceland (the such as the eastern continental Greece and the Black Icelandic Low) and the Azores Portugal (the Azores Sea region (western Greece and southern Turkey), High). Positive (negative) NAO creates drier (wetter) receive higher precipitation than during negative conditions as less air from the North Atlantic penetrates (positive) NCP phases (Kutiel and Benaroch, 2002). into the region (Brandimarte et al., 2011). The Mediterranean region also favors the development of 3.2. Summer atmospheric circulation patterns cyclones with synoptic to mesoscale characteristics (Lionello et al., 2016). Cyclones are produced as a Summer conditions in the eastern Mediterranean region result of the advection of polar air masses over the are dominated by two processes. The first process is the warm sea as well as the passage of North Atlantic dynamic subsidence, which inhibits cloud formation synoptic systems and their interaction with topography and convection, resulting in a prolonged warm and dry (Trigo et al., 2002). Cyclogenesis takes place within the summers (Dayan et al., 2017). This has traditionally Mediterranean region in preferable places such as Gulf been associated with the African monsoons and the of Genova, Ionian Sea, Aegean Sea and Cyprus from subtropical descending branch of the Hadley cell, which which cyclones track SE-ward (Alpert et al., 1990). shifts over the southern Mediterranean (Ziv et al., 2004; Conditions for northerly flows of cold and dry air Lionello et al., 2006). Summer aridity and warming masses over the Mediterranean are mainly associated have also been linked to the Asian monsoon intensity with a westward ridging of the Siberian High and also (Rodwell and Hoskins, 1996). Deep convection over the the pressure difference between the western and eastern Asian monsoon region induces a westward propagating parts of the basin (Maheras et al., 1999). In particular, Rossby wave generating air masses descent over the Mediterranean Oscillation (MO) is defined by the Eastern Mediterranean (Ziv et al., 2004; Tyrlis et al., pressure differences between stations at Algiers and 2013). The second process is the dry and cool northerly Cairo and is linked to NAO modes (Fig. 1) (Conte et winds, i.e. the Etesians, whose ventilating effect al., 1989). Positive (negative) MO modes are dominated counteracts the adiabatic warming induced by the large- by an anticyclone (cyclone) in the Western scale subsidence. These winds are generated by a sharp Mediterranean and low (high) pressure systems in the pressure gradient, which builds up over the area due to Eastern Mediterranean leading to a southward the combination of the extended Asian thermal low and (northward) flow of cool (warm) and dry (wet) air over the Azores High (Rizou et al., 2018). Greece lies at the the Eastern Mediterranean (Feidas et al., 2007). In transitional zone between these two systems, resulting addition, the winter temperature and precipitation are in strong Etesians over the Aegean Sea (Fig. 1 and Fig. influenced by the pressure difference between the North 2). Thus, both subsidence and Etesians are related to the Sea and the Caspian Sea, the so-called North Sea- monsoonal activity and lead to dry summer conditions. Caspian Pattern (NCP) (Kutiel et al., 2002). An Despite this, sporadic rainfall events are induced by anticyclonic (cyclonic) upper circulation anomaly in the locally developed thermal lows associated with the North Sea and an cyclonic (anticyclonic) circulation diurnal fluctuations of inland surface temperature as pattern north of the Caspian Sea induce a southward well as by cut-off lows during reduced anticyclonic cold (northward warm) air stream over Balkans and activity periods (Trigo et al., 2002; Flocas et al., 2010).

4. The Peloponnese

4.1. Topography and Climate

The Peloponnese peninsula constitutes the southern part of the Greek mainland covering an area of 21.549 km² (Fig. 2 and Fig. 3). The terrain exhibits variable morphology characterized by rugged mountains, hills, broad plains and valleys. The peninsula is crossed by a large mountain range, which geologically is the extension of the Pindos Mountains stretching from the Greek-Albanian borders (NW) to central Greece. The average altitude in the peninsula is ca 543 m with the maximum elevation being ca 2400 m in the Taygetos mountain chain, in the central-southern Peloponnese. Climatic conditions are spatially and temporally diverse due to the irregular topography and the impacts of different atmospheric circulation patterns. Overall, the Peloponnese has mild, wet winters and hot, dry Fig. 2. Terrain map of Greece with the location of the Peloponnese peninsula in the inset box and the subduction summers. Most of the precipitation falls between zone to the south-southwest. Map modified after October and April. The NNW-SSE mountain range scilands GmbH (www.scilands.de). Scientific Landscape. across the peninsula causes a rain shadow effect on the leeward (eastward) side of the mountain with respect to the westerly winds (Dotsika et al., 2010). As a result, the eastern Peloponnese is dominated by drier and cooler conditions than SW Peloponnese (Fig. 3) (Fick and Hijmans, 2017). The annual mean precipitation and temperature in NE Peloponnese are 476 mm and 17.0°C respectively, with the monthly mean temperatures ranging from ca 27°C in July to ca 8°C in January (Pyrgela station, close to Lerna) (Hellenic National Meteorological Service, 2018). The annual mean precipitation and temperature in SW Peloponnese are 780 mm and 18.0°C respectively, with the monthly mean temperatures ranging from ca 26.5°C in July to ca 10 °C in January (Kalamata station, near Agios Floros) (Hellenic National Meteorological Service, 2018). The Peloponnese is located at the junction of different atmospheric circulation patterns (NAO, NCP, and Siberia High). During autumn, winter and early Fig. 3. Distribution of mean annual precipitation in the spring, frontal depressions from the Atlantic and the Peloponnese, for the period 1970-2000 (Fick and Hijmans, western Mediterranean produce significant precipitation 2017) with the study sites i.e. Gialova Lagoon, Agios Floros amounts associated with negative NAO and NCP fen and Lake Lerna as well as the location of other sites modes (Flocas and Giles, 1991). Saharan depressions discussed in the text (after Katrantsiotis et al., 2019). contribute to the rainfall totals as they move north– northeast and pick up moisture over the Mediterranean. The eastern Peloponnese receives high rainfall amounts 4.2. Geology and landscape development and occasional snowfall associated with cold northerly The Peloponnese displays a composite morphotectonic winds and cyclogenesis over the Aegean Sea during structure characterized by the presence of large grabens positive NCP periods. During summers, the northward and horsts bounded by E–W and NNW-SSE trending extension of the Hadley circulation causes atmospheric fault zones (Fig. 4) (Fountoulis and Mariolakos, 2008; stability. In addition, the area is affected by the Asian Vassilopoulou, 2010; Fountoulis et al., 2014). Inside monsoon system through modulating the intensity of these neotectonic macrostructures, there are many small large-scale subsidence and the strength of the Etesians morphological units of various directions. The grabens (Tyrlis et al., 2013). Strong Etesians trigger cooler were formed during the Pliocene-Early Pleistocene due conditions in NE Peloponnese compared to SW to the extensional deformations, which resulted in the Peloponnese located on the lee side of the mountain. fragmentation and the subsidence of the bedrock along Despite the summer aridity, shortwave troughs and cut- E–W and NNW-SSE normal fault zones (Zelilidis and off lows can sporadically cause enhanced instability and Doutsos, 1992; Zelilidis and Kontopoulos, 1994). This thus convective rainfall events (Xoplaki et al., 2003). further allowed sea level intrusion and deposition of

4.3. Human occupation The Peloponnese has a long history of human activity which extends back to the Paleolithic age. The peninsula was an important center for the establishment and evolution of ancient Greek civilization, comprising a wide and well-documented spectrum of early farming communities, palatial economies and city-states such as those of Mycenae and Olympia (Weiberg et al., 2016). The region is well-documented in historical sources. A large number of archaeological field surveys have been carried out in this region, tracing long-term sites and settlement patterns (Alcock and Cherry, 2004). Overall, the Peloponnese has been the focus of multi- disciplinary studies aiming to develop an integrated long-term narrative of the interactions between humans and environment in this region (Weiberg et al., 2016).

5. Description of study sites

Three sites were studied in this thesis; the Agios Floros Fig. 4. The main marginal fault zones of the post-Alpine fen and the Gialova Lagoon in SW Peloponnese and the basins in the Peloponnese (after Fountoulis et al., 2014). Lake Lerna in NE Peloponnese (Fig. 3). The selection of these sites was based on three criteria. (1) They are marine sediments in the subsided areas along the located in different climatic zones of the peninsula, coastal zones of the peninsula until the early-middle which is important to trace regional climatic differences Pleistocene (Athanassas and Fountoulis et al., 2013). and common patterns during the Holocene. 2) The Thereafter, an inversion of the kinematic regime caused studied sites are situated in geomorphologically the emergence of the formerly subsiding tectonic blocks sheltered positions, which favor sediment deposition at and the uplift and erosion of marine sediments at their a constant rate. These continuous deposits most likely present state (Fig. 4) (Marcopoulou-Diacantoni et al., contain uninterrupted records of the Holocene climatic 1989; Athanassas and Fountoulis et al., 2013; changes. (3) Hydrological conditions in these sites are Fountoulis et al., 2014). In addition, high rates of fluvial influenced by climate factors i.e. the balance between sedimentation, along with the retreating sea level, led to precipitation/freshwater input and evaporation rate. (4) the partial infilling of embayments and the formation of The sites are located close to important archaeological alluvial plains such as the Messenian plain, the places. The data from this thesis will help future studies Typhlomytis plain and the Argolis plain during the to place archeological findings in a climate perspective. Quaternary period (Fig. 4) (Kraft et al., 1975, 1977).

The Peloponnese is located in the direct vicinity of the Hellenic arc, i.e. the convergence of the African and 5.1. Messenian plain/Agios Floros fen the Eurasian plate, and is also characterized by a large number of active neotectonic faults (Fig. 2 and Fig. 4). The Agios Floros fen is located in the central-eastern These factors make the peninsula one of the most Messenian plain, which is a tectonic graben bounded by tectonically and seismically active areas in Europe. It NNW–SSE normal faults (Fig. 5). The bedrock consists has been subjected to destructive earthquakes like those of Mesozoic and Early Cenozoic limestones covered of 1875 (Magnitude=6, Kyparissia), 1886 (M=6.8, with Pliocene–early Pleistocene marine sediments, Filiatra), 1903 (M=6.6, Κythira) and in 1986 (M=6.0, Pleistocene alluvial deposits as well as Holocene peat Kalamata) (Fig. 4) (Papazachos and Papazachou, 2003; accumulations (Perrier, 1975). The fen was developed Fountoulis and Mavroulis, 2013; Papadopoulos et al., during the Holocene in a small shallow basin, with a 2014). In addition to high seismicity, the coastal areas maximum depth of 2 m and an outlet to the south via of the peninsula have been impacted by tsunamogenic the Pamisos River (Fig. 5). The fen is located less than events during the Holocene and beyond associated with one kilometer from the outlet of karst springs on the large magnitude earthquakes along the Hellenic arc western slopes of the Taygetos Mountain (Fig. 6). The (Vött, 2007; Scheffers et al., 2008; May et al., 2012; springs are connected to a NNW striking normal fault Willershäuser et al., 2015). For instance, historical system, and are the main water sources of the Pamisos accounts have described one such tsunami event that River (Fig. 5). The underlying karst aquifer acts as a occurred on 21 July 1635 (Seyfarth, 1971). The tsunami reservoir of rainfalls, and hence spring discharges are waves produced by a big earthquake near the west coast modulated by the annual precipitation amount. Water of Crete, south of the Peloponnese, reached the Methoni level changes in the fen are mainly attributed to surface settlement in Pylos causing inland inundation of 2 km. runoff, spring water outflow and the evaporation rate.

Nymphaeetum albae and Potamogeton natans. The hills are covered by dwarf-shrubs and sclerophyllous evergreen shrubs and trees (Papazisimou et al., 2005). The central Messenian plain is poorly investigated. Previous studies have mainly focused on the palaeo- geographical evolution of the lower Messenia plain based on pollen and foraminifera and with specific reference to archaeological evidence from settlements (Kraft et al., 1975; Engel et al., 2009). The only available study from the central Messenian plain is based on a low resolution pollen record providing evidence for Holocene vegetation shifts and peat formation in Agios Floros (Papazisimou et al., 2005).

5.2. Gialova Lagoon The Gialova Lagoon is a shallow coastal ecosystem located in the western Messenia, SW Peloponnese (Fig. 7). It is a part of a small active depression, which consists of the Navarino embayment to the south, and the Tifflomitis alluvial plain to the north of the lagoon (Fig. 7 and Fig. 8). The bedrock consists of Cretaceous and Paleocene/Eocene limestones (IGME, 1980). The post alpine deposits are composed of Pliocene marine F ig. 5. Geological map of the Messenian plain (after Perrier, sediments, covered with Holocene fluvial sediments. 1975 ) with the location of the Agios Floros fen and the coring The Gialova Lagoon is located at 0-1 m a.s.l. (meters above sea level) and linked to the adjacent site, close to the karst springs (blue dot). embayment via a narrow artificial inlet, which cuts through a sandy barrier system (Fig. 7 and Fig. 8). To the west, there is a belt of sand dunes, which separates the lagoon from the Voidokilia Bay and the Ionian Sea. The swamp area to the east has been drained for agricultural purposes. The catchment area to the north is mainly covered with floodplain sediments (Fig. 7). The lagoonal environment was established at ca 5300-4200 cal BP (calibrated years before present) and gradually replaced by a completely isolated lagoon until ca 3300 cal BP (Emmanouilidis et al., 2018). After the isolation, the physicochemical parameters of the lagoon have mainly been governed by the balance between precipitation/freshwater input via rivers and evaporation rate (Papakonstantinou, 2015). The lagoon has a surface area of ca 2.5 km2 with water depths ranging from 0.5 to 1m. The lowest water depth and the highest salinity occur during summers. The lagoon is also characterized as eutrophic in April and hypereutrophic in August with pH raging from 7 to 9 and the highest dissolved oxygen mainly occurring during winters (Koutsoubas et al., Fig. 6. The Agios Floros wetland with the location of the 1997; Bouzos et al., 2002; Papakonstantinou, 2015). coring site close to the flooded fields. The lagoonal bottom is sparsely covered with the salt tolerant grasses, Ruppia cirrhosa, and the seagrass, Until the middle of the last century, large areas of Cymodocea nodosa, mainly close to the communication the river valley were covered by swamps. Thereafter, channel with the Navarino marine embayment the valley was drained for agricultural activities (Lyras, (Tiniakos and Tiniakou, 1997; Papakonstantinou, 2011). The cultivated area currently consists of Olea 2015). The vegetation around the lagoon generally europaea. The vegetation in the fen and along the river consists of salt-tolerant plants (e.g. Salicornia europea), banks is characterized by reed (Phragmites australis, halophyte and rush plant communities (e.g. Schoenus Arundo donax), shrubs and trees e.g. Salix fragilis, nigricans and Juncus maritimus), rhizome geophyte Platanus orientalis, Nerium oleander. The springs and (e.g. Ammophila arenaria), phrygana communities (e.g. ponds are dominated by pure aquatic vegetation e.g. Sarcopoterium spinosum) evergreen shrubs and trees

(e.g. Juniperus phoenicea, Thymus capitatus), mastic 5.3. Lake Lerna tree (Pistacia lentiscus), perennial grasses, sedges and The Lerna site is located a few meters from the coast of plants (e.g. Phragmites australis) and cultivations (e.g. the Gulf of Argos, within the Argive Plain, in NE Olea europea) (Zangger et al., 1997; Korakis, 2006). Peloponnese (Fig. 9 and Fig. 10). The Gulf of Argos

and the Argive plain represent a fault-bounded NW-SE tectonic depression. The bedrock exposed to the east and west of the plain is composed of Mesozoic/Eocene limestones, ophiolites and flysch (van Andel et al., 1990). Plio-Pleistocene marls and fluvial conglomerates occur at lower elevations around the plain. The low- lying areas consist of Holocene coastal and alluvial deposits punctuated by marine sediments. The plain is bordered by NW-SE faults with secondary fault systems trending NE-SW and E-W (Maroukian et al., 2004).

Fig. 7. Geological map of the Navarino area (IGME, 1980) with the location of the Gialova Lagoon and the coring site in the central part of the lagoon.

W S

Fig. 9. Geological map of the Argive Plain. The black outline depicts the extension of the Ancient Lake Lerna (after van Andel et al., 1990). The black dot shows the location of the coring site.

Fig. 8. Gialova Lagoon. The Voidokilia Bay to the west. S N The sandy barrier and the Navarino embayment to the south.

The area surrounding Gialova Lagoon has a long history of human presence especially after 5000 cal BP (Kraft et al., 1980). The Navarino area functioned as one of the most important centers of the Mycenaean civilization at 3600-3000 cal BP (Zangger et al., 1997). Intensive settlement and land use are also inferred from Fig. 10. The Lerna wetland with the location of the coring site the Hellenistic and Roman times (2500-1400 cal BP). to the south inside the reed belt. Previous studies have mainly focused on the palaeoenvironmental evolution of the Navarino area The ancient Lake Lerna was formed behind a beach based on pollen and foraminifera in sediment cores barrier at ca 8300 cal BP, with maximum extension at (Kraft et al., 1980; Zangger et al., 1997; Avramidis et ca 6000 cal BP (Zangger, 1991). This barrier was al., 2015; Willershäuser et al. 2015; Emmanouilidis et created by the Inakhos River propagation and al., 2018). In addition, special attention has been paid to dispersion of sediment along the shore. The Lake was the interplay between landscape development, changes isolated from the sea and supplied with freshwater from in vegetation and societal dynamics in the area around the Erasinos River throughout the Holocene (Fig. 9) the lagoon (Kraft et al., 1980; Zangger et al., 1997). (Finke, 1988; Zangger, 1991). At ca 5000-4000 cal BP,

7 the eastern part of the lake was filled with sediments 6.1. Field work while the western part remained a shallow water body until modern times (Zangger, 1991; Jahns, 1993). Lake and lagoonal sediment sequences were collected The Lerna site is currently a wetland which from each studied site using a hand-operated Russian partially gets flooded during the wet season. The water peat corer. The coring was performed by recovering regime is also determined by numerous streams and contiguous core segments at successively greater depths springs. The vegetation in the wetland consists of reeds, on the same spots. In particular, during our field work (Phragmites australis and Arundo Donax), perennial in October 2013, a 7.5 m long core was retrieved from herbaceous plants (e.g. Typha latifolia) with emergent the Agios Floros fen, ca 2 km west of the homonymous and floating plants (Cladium mariscus and Nymphaea village, close to karst springs (Fig. 5 and Fig. 6) (N sp.) mainly occurring in small ponds (Jahns, 1993; 37°10'05.25" E 22°01'05.61"; 13 m a.s.l.). During the Pantazopoulou, 2015). The area around the wetland is same field campaign, a 2.6 m long core was recovered dominated by orange tree and olive cultivations. The from the central part of the Gialova Lagoon (Fig. 7 and non-cultivated land comprises phrygana and evergreen Fig. 8) (N 36°57'48.59" E 21°40'16.65"; 1 m a.s.l). The shrubs (e.g. Quercus coccifera, Pistacia lentiscus) and recovery of the core was performed from a boat. The some riparian vegetation (e.g. Eucalyptus, Salix) uppermost 10 cm was lost during sampling due to the (Pantazopoulou, 2015). The mountain areas are mainly high water content of sediments. For the Agios Floros- dominated by the holm oak (Quercus ilex) and ever- Gialova campaign, the Russian corer had a size of 0.5 green trees (e.g. Abies cephalonica and Pinus nigra). m length and 4.5 cm diameter. Field work was The Argive plain constitutes an archeologically conducted at the ancient Lake Lerna in December 2016. important area with famous Bronze Age settlements, A 5 meters long core was sampled with a Russian corer such as Mycenae, Tiryns and Lerna (Jahns, 1993; Cline, of 1 m length and 4.5 cm diameter (Fig. 9 and Fig. 10) 2010; Weiberg et al., 2016) and the city-state of Argos (N 37°34'57.66" E 22°43'57.60"; 1 m a.s.l.). The during the historical periods (Pariente and Touchais, sequences were examined in the field for their physical 1998). The archaeological site of Lerna (ca 4 km SW of characteristics such as lithology, color, boundaries and the coring site) was used from the Neolithic period and internal structure. The lithological descriptions were throughout the Bronze Age, with some activities in the later discussed in relation to proxy data. Thereafter, the historical periods (Zangger, 1991; Jahns, 1993; cores were placed into rigid plastic half-tubes, wrapped Wiencke, 2010). The site is particularly known for its in plastic film and were transported to the University of Early Bronze Age remains (ca 4700-4200 cal BP). the Peloponnese, in Kalamata, Greece and to Stockholm Paleoenvironmental studies in the Argive plain are University for subsampling and laboratory analysis. relatively old (early 90s) focusing on the Holocene shore displacement of the Gulf of Argos and vegetation 6.2. Proxy description and laboratory work shifts around the Lerna site (Kraft et al., 1977; Zangger, 1991; Jahns, 1993). Attempts have also been made to The sediment cores were visually examined in more bring together information from different archives and detail for their physical characteristics in the laboratory archaeological records to explore links between societal and thereafter sub-sampled with a resolution of 1 cm. dynamics and landscape development (e.g. Zangger, The subsampling for the Agios Floros and Gialova 1991; Jahns, 1993; Cline, 2010; Weiberg et al., 2016). cores was performed at the laboratory of Archaeometry at the University of the Peloponnese. The sub-samples were sealed in plastic bags and transported to Stockholm University. The subsampling for the Lerna 6. Material and methods core was performed at the Department of Physical Geography, Stockholm University. All the samples The studies presented in this thesis are based on down- were kept in storage at 4°C until they were analyzed. core fossil data extracted from sediment cores. A multiproxy approach was applied to strengthen 6.2.1. Radiocarbon dating paleoclimatic interpretation. The methods involved several steps. The field work planning and performance Radiocarbon dating is the most widely used method to included the examination of sites for suitable coring establish an age-depth relationship for sediment cores. spots and the recovery of cores. The chronologies of the The principles of the radiocarbon dating have been cores were based on accelerator mass spectrometry discussed extensively (e.g. Broecker and Kulp, 1956; (AMS) radiocarbon dating of macrofossils and bulk Olsson, 1991; Björck and Wohlfarth, 2001). Briefly, the samples. The laboratory work further involved the use radioactive carbon isotope (14C) is formed in the upper of various techniques to extract and analyze biological atmosphere, where cosmic ray neutrons react with 14N and geochemical proxies. In particular, the proxy atoms. The newly formed 14C is rapidly oxidized to 13 14 methods included diatoms, n-alkanes and their δ C and CO2, dispersed through the atmosphere and δD composition, bulk geochemistry and XFR core subsequently enters the Earth's plant and animal life scanning data. A theoretical framework for each proxy through photosynthesis and the food chain. Plants and and further details for the laboratory and analytical animals, which take up 14C during their lifetimes, exist work are given both here and also in the papers (I-V). in equilibrium with the atmospheric 14C concentration.

When the organism dies, absorption of 14C ceases, and 6.2.2. Siliceous microfossils the amount of 14C gradually decays at a constant rate. Carbon-14 has a half-life of 5730±40 years, meaning Diatoms that the concentration halves every 5730 years. By Diatoms are microscopic unicellular algae encased in a measuring the 14C content left in a sample, and cell-wall, known as frustules. The cell wall consists of particularly the ratio of 14C against the stable forms of rigid amorphous silica which has high preservation carbons, the time since the organism died can be potential in sedimentary environments. The application estimated. However, the production of carbon isotopes of diatoms in paleoenvironmental studies has a long varies both spatially and temporally, and therefore tradition (Stoermer and Smol, 2010 and references calibration curves need to be used to covert radiocarbon therein). This is because the siliceous frustule displays dates into calendar years (Olsson, 1991; Fairbanks et specific morphological characteristics enabling diatom al., 2005). The radiocarbon calibration curve for the identification into lower taxonomic level (Round et al., northern hemisphere (IntCal13) is extended back to 1990). Besides, diatoms have well-defined ecological 50.000 cal BP and is based on 14C measurements of preferences and high species diversity. These single-cell known-age tree-ring, plant macrofossils, speleothems, organisms reproduce quickly and are sensitive to a corals, and foraminifera samples (Reimer et al., 2013). range of environmental factors such as salinity, water Macrofossils from terrestrial plants, which fix only depth and nutrient availability (Battarbee et al., 2001). atmospheric CO , are considered to be the most 2 Therefore, diatom species composition changes in fossil reliable material for 14C dating (Olsson, 1991; Björck records due to shifts in one of these parameters can be and Wohlfarth, 2001). However, a source of uncertainty used to reconstruct paleo-hydrological conditions. is sometimes related to the potential input of younger or Although diatoms are considered as excellent older terrestrial plant macrofossils into the sediments indicators of past environmental conditions, the quality through erosion and bioturbation (Björck and of reconstruction usually relies on the degree of diatom Wohlfarth, 2001). In the absence of terrestrial remains, preservation, which can be a major taphonomic issue aquatic plant macrofossils and bulk sediments can be (Ryves et al., 2013). Diatom valves are sometimes used for 14C dating with the caveat that these materials absent or poorly preserved in sediments. The can be influenced by the reservoir effect and/or hard composition of fossil assemblages can be affected by water effect. These effects are associated with the 14C- various chemical and physical processes including depleted carbon in water compared to atmospheric CO 2 dissolution, resuspension and bioturbation (Snoeijs and and/or the input of pre-aged carbon via rivers and Weckström, 2010). Poor preservation of diatom species groundwater. The assimilation of old carbon by aquatic due to silica dissolution and valve fragmentation is plants results in too old apparent ages (Olsson, 1983). common in shallow freshwater and saline ecosystems

(Ryves et al., 2006). This could bias diatom Core chronology – Age depth modelling assemblages to more resistant taxa (Barker, 1992). Other potential limitations when interpreting diatom Sediment samples throughout the cores were soaked in records include misidentification of taxa, taxonomic 10% KOH and washed with a fine jet of water through problems and limited ecological knowledge of species. a 0.250 mm sieve. The residues were dispersed in a petri dish filled with distilled water. Macrofossils and charcoal pieces were extracted under a stereo Phytolith/Diatom ratio microscope using soft tweezers. The dating materials Phytoliths are siliceous remnants of the decomposition were placed in glass vials filled with distilled water. In of plant material with highly distinctive shape that can the Lerna core, bulk sediments were also dated in some be identified to species level (Piperno, 2006). Upon depths due to the absence of seeds. Accelerator mass plant death and decay, phytoliths are preserved in the spectrometry (AMS) 14C dating of the samples was depositional environment and can be used to examine performed in Poznan Radiocarbon Laboratory, Poland, vegetation history (e.g. Rovner, 1971; Delhon et al., Beta analytics and in the Tandem Laboratory, Uppsala. 2003; Finné et al., 2010). The phytolith/diatom ration The calibration of 14C ages and the construction of has been used as an indicator of water level changes age depth-models were performed with 2σ confidence and shifts in the distance between sampling site and intervals using the CLAM 2.2 program (Blaauw, 2010) shore (Chalié and Gasse, 2002). Small distances and and the IntCal13 calibration curve (Reimer et al., 2013) shallow water levels likely favor the high abundance of in the software R.3.1.1 (R Development Core Team, phytoliths in the site (Holmgren et al., 2012). Deep lake 2013). All calibrated ages are given in years before and large distance likely prevent phytoliths of terrestrial present (cal BP). Age-depth models were constructed and telmatic origin from reaching the sampling site. The by using linear interpolation between the midpoints of phytolith production rate, however, depends on the 2-sigma calibrated age ranges and extrapolating to the vegetation type in the catchment area. Therefore, the base and top of the cores. Weighted average estimates phytolith/diatom ration should be used as a complement and accumulations rates (cm-1) were obtained from tool to the reconstruction based on the habitat diatom every depth based on the entire calibrated distribution groups. The latter provides more qualitative information of all dates and also the applied model (Blaauw, 2010). on the past lake level changes (Chalié and Gasse, 2002).

Siliceous microfossil extraction and analysis Holmgren et al., 2012). The results are presented as stratigraphic diagrams depicting the relative abundance The sediment cores retrieved from the three studied of diatom species, the dissolution and phytolith indices sites were first examined for diatom presence. Sediment as well as the clastic debris content. The graphs were samples were taken and analyzed from each litho- constructed using the Tilia program version 1.7.16, stratigraphic unit throughout the cores. The samples (Grimm, 1991, 2004). Zones of the diatom assemblage were prepared according to the standard method were defined performing constrained incremental sum- described below (Battarbee, 1986). The Agios Floros of-squares cluster analysis (CONISS; Grimm, 1987). sediment core contained abundant diatoms, and therefore was proceeded for detailed analysis. The samples from the Gialova Lagoon and the Lake Lerna 6.2.3. Organic geochemistry were deprived of diatoms and the cores from these two Organic matter (OM) in sediments consists of mixture sites were analyzed for other geochemical proxies. of biogeochemical components, derived from various Around 2 g of wet sediment with a resolution of 4– autochthonous and allochthonous sources in and around 6 cm were used for diatom analysis in the Agios Floros aquatic ecosystems (Meyers and Teranes, 2001). These core. The samples were treated with 10 % HCl to substances retain paleoenvironmental information remove carbonates. Organic compounds were oxidized regarding the sedimentary OM origins and also the by using 17% H O . To minimize vivid chemical 2 2 abundance of different biota from which the organic reactions, the samples were left at room temperature compounds are derived (Meyers, 2003). Therefore, the overnight and boiled on water-bath the following day. identification of different OM sources is important for After oxidation, the suspensions were repeatedly paleoenvironmental reconstructions. The study of OM decanted at two-hour intervals from 100 ml beakers can be performed in bulk samples that contain a mixture until most of the clay particles were removed. Distilled of organic compounds, and/or at molecular level where water with NH was added to disperse the remaining 3 each compound can be attributed to a specific source. clay particles. The residues were transferred to 5 ml tubes and a volume of the residue was extracted and n-Alkanes spread out evenly on the cover-glass. The prepared samples were dried onto cover-slips and mounted on a The analysis of lipid biomarkers, particularly aliphatic microscope slide using Naphrax® to increase refraction saturated hydrocarbons (n-alkanes), has become a index. Counting were carried out under a light widely applied tool in paleo-environmental studies (e.g. microscope at a magnification of X1000 using oil Pu et al., 2010; Zech et al., 2011; Aichner et al., 2018). immersion to further increase the refraction index. This is because n-alkanes are characteristic of specific In total, 125 samples were analyzed throughout the sources, constitute the most abundant biomarkers in Agios Floros core and a minimum number of 300 sediments and also are relatively easy to purify and valves were counted in each slide. Detailed taxonomic analyze (Peters et al., 2005; Grice and Eiserbeck, 2014). descriptions were performed using SEM. Diatom Besides, due to their structure consisting of strong and species were grouped into classes based on their water localized C–C and C–H bonds, n-alkanes are the most depth preferences (benthic/planktonic/subaerophilous). resistant lipid compounds to degradation, and hence are Identification and ecological classification were preserved in sediments over geological time-scales primarily based on Krammer and Lange-Bertalot (1986, (Labinger and Bercaw, 2002; Peters et al., 2005). 1988, 1991a, 1991b). To investigate the impact of Different organisms produce n-alkanes with distinct dissolution in the diatom assemblage, a dissolution carbon chain-lengths (Bush and McInerney, 2013 and index was calculated after classifying the valves as references therein). The down-core carbon chain-length intact versus dissolved (F index = dissolved/intact + distributions can therefore provide evidence for the dissolved valves; Currás et al., 2012; Ryves et al., sedimentary OM origin (Giger et al., 1980; Meyers, 2001). Dissolved valves were considered all the valves 1997). Aquatic algae and bacteria are the primary showing sign of dissolution under the LM; valves with producers of short-chain n-alkanes (C17 to C21) margins and striation missing, transparent and (Cranwell et al., 1987). Mid-chain homologues (C23 to featureless valves and also broken valves with signs of C25) are mainly attributed to submerged aquatic dissolution (dissolved fragments) (Barker, 1992). The macrophytes and/or mosses (Baas et al., 2000; Ficken et dissolution index varies between 0 and 1; with a score al., 2000). Long chain n-alkanes (C27 to C35) are of 1 representing samples dominated by dissolved typically produced as part of the epicuticular leaf waxes valves. In addition, the number of phytoliths was of terrestrial higher plants (Eglinton and Hamilton, counted along the same transects. No attempt was made 1967). In particular, woody plants (trees and shrubs) to distinguish the different phytolith morphotypes. A have been shown to produce C27 and C29 n-alkanes, phytolith index (phytolith/diatom + phytolith) was while herbaceous plants and grasses tend to contain a calculated with values close to 1 representing samples relatively higher amount of C31 or C33 n-alkanes dominated by high phytolith amounts. The phytolith (Cranwell, 1973; Meyers and Ishiwatari, 1993). index, along with the visual inspection of clastic debris However, there is a large variation in the chain-length content in each sample, was used as an additional distribution within the different plant groups attributed indicator of lake level changes (Chalie and Gasse, 2002; to physiological and environmental factors (Bush and

McInerney, 2013). Some studies have shown that graphs showing the n-alkane abundances throughout the woody plants, such as Olea europea, can synthesize sediment cores. The graphs were constructed using the abundant n-C33-C 35 alkanes under moisture deficit (e.g. Tilia program version 1.7.16 (Grimm, 1991; 2004). Pierantozzi et al., 2013). Aquatic plants have also been Three indexes were calculated based on the C17-C35 found to produce long-chain n-alkanes in shallow lakes homologues. The carbon preference index (CPI) was (Aichner et al., 2010). Overall, the n-alkane distribution calculated as: CPI = Σodd / Σeven (Wang et al., 2003). The should not be used indiscriminately to constrain the OM average chain length (ACL) was determined as: ACL = 13 sources. This method should be combined with the δ C Σ (n * Cn) / Σ Cn where Cn is the concentration of odd analysis of n-alkanes and ideally complemented with a chain n-alkanes (Tanner et al., 2010). The Paq index survey of n-alkane production of the local vegetation. (proxy for aquatic macrophytes) was calculated as: Paq = (C23 + C25) / (C23 + C25 + C29 + C31) (Ficken et al., 2000). n-Alkane extraction and quantification Stable carbon and hydrogen isotopes Lipid extraction was performed on freeze-dried and homogenized sediment samples (1-15 g), which were Organic compounds mainly consist of molecules taken at every 5 cm throughout the sediment cores. The containing carbon and hydrogen, and therefore the samples were then mixed with organic solvents stable isotopes of these two elements are studied in (dichloromethane: methanol 9:1 v/v), placed in a sonic organic geochemistry (Brocks and Summons, 2014). bath for ca 20 min and centrifuged at 3000 rpm. Stable isotopes are atoms of the same chemical element Afterwards, the lipid extract was recovered into a glass with the same number of protons and electrons (i.e. tube. This procedure was repeated three times. The atomic number) but different number of neutrons (i.e. extracts were combined into a total lipid extract dried mass number). Carbon has two stable isotopes 13C and 12 1 under a nitrogen gas stream (N2), re-dissolved in hexane C, the latter being the lighter and most abundant. H and adsorbed onto precombusted silica gel. Solid phase (H) and 2H or D (deuterium) are stable isotopes of extraction was performed in order to obtain aliphatic hydrogen with deuterium being the heavier and rarer hydrocarbon fractions using pipette columns filled with (Wieser and Brand, 2017). Since the various isotopes of glass wool, silica gel and the (adsorbed) lipid an element have different mass, their chemical and compounds on top. The fractions were extracted by physical properties are slightly different, a phenomenon flushing the columns with hexane (3×column volume). called isotope effect. There are two types of isotope In some samples from the Agios Floros and Lerna cores effects - equilibrium and kinetic - which cause isotopic elemental sulfur was detected during the gas fractionation i.e. partial separation of isotopes of the chromatography-mass spectrometry (GC-MS) analysis same element between two substances during physical described below. To remove sulfur and obtain fractions or chemical processes (Kendall and Caldwell, 1998). with only saturated hydrocarbons, the aliphatic Heavy isotopes form stronger bonds and react, diffuse hydrocarbon fractions were eluted with hexane through or evaporate slower compared to light ones (kinetic pipette columns filled with glass wool, activated copper fractionation). At equilibrium, heavy isotopes tend to (prepared by rinsing with a sequence of HCl (1 M), be concentrated in phases with stronger chemical bonds isopropyl alcohol, dichloromethane and hexane) and (low energy state: solid phases) than the light isotopes, 10% silver nitrate (AgNO3)-coated silica gel. The final which are more volatile (equilibrium fractionation). fractions with hexane were then stored in a freezer. Due to the low abundance of the heavy isotopes, For the n-alkane identification and quantification stable isotopes are expressed as the ratio of the rare hydrocarbon fractions were analyzed by a Shimadzu (heavy) isotope to the common (light) isotope compared GCMS-QP2010 Ultra gas chromatography-mass to the same ratio in an international standard using spectrometry (GC-MS). The instrument was equipped the delta notation (equation 1). The isotopic with an AOC-20i autosampler and a split-splitless composition is reported as parts per thousand or 'per injector operated in splitless mode. For separation, a mil' (‰) with positive values indicating that the sample GC column with the following dimensions was used: has a higher ratio than the standard and vice versa.

30m×0.25mm×0.25 μm (Zebron ZB-5HT Inferno). The GC oven was programmed for a temperature increase δ = (Rsample / Rstandard -1) * 1000 (1) from 60 °C to 180 °C by 20 °C/min, then up to 320 °C by 4 °C/min, followed by a hold time of 20 min. The internationally reference standards are: Vienna Pee Quantification was performed with an external Dee Belemnite (VPDB) for carbon isotopes (δ13C, with calibration curve, produced by analyzing regularly a n- R=13C/12C) and Vienna Standard Mean Ocean Water C21-40 alkane standard solution (Fluka Analytical) at (VSMOW) for hydrogen isotopes (δD, with R=D/H). different concentrations. The software MS Solution 4.11 (Shimadzu) was used for compound identification δ13C values of photosynthetic organisms and area integration of the relevant peaks. Compound identification and area integration of the relevant peaks Plants in terrestrial environments absorb atmospheric were performed with the GC5MS Solution software CO2 as the sole carbon source for photosynthesis. The version 4.11 (Shimadzu). The results are reported in resulting isotopic composition of plants thus depends on

11 the 13C content of carbon source and the isotope effects during assimilation of CO2 and biosynthesis in the plants (Hayes, 1993; Chikaraishi, 2014). The variation 13 in the C content of atmospheric CO2 was minimal during the late Quaternary, mainly ranging between - 6‰ and -8‰ (Jasper and Hayes, 1990; Schmitt et al., 2012; Bauska et al., 2016). Therefore, changes in the isotopic composition of terrestrial plants can largely be associated with fractionation primarily attributed to kinetic isotope effect. This takes places in two steps and results in plant biomass being more depleted in 13C than CO2 (Fig. 11) (Farquhar et al., 1989; Chikaraishi, 12 2014). In the first step, CO2 preferentially diffuses through the stomata into plant cells due to its lighter mass and higher velocity. In the seconds step, the isotopic fractionation is associated with fixation of inorganic carbon via carboxylation. The fractionation extent depends on the photosynthetic pathway (O’Leary, 1988). In particular, bulk biomass of C3 Fig. 11. δ13C values for different plants, their leaf waxes plants fixes CO2 via the Calvin-cycle, which shows a and carbon sources (after Mouk, 2001 and Aichner, 2009). 13 greater discrimination against C than the Hatch-Slack- Cycle. The latter is used by C4 plants, which 3.6‰ on average compared to bulk biomass (Fig. 11) consequently have less negative bulk δ13C values (from (Park and Epstein, 1961; Collister et al., 1994). The 13 −9 to −17‰) than C3 plants (between −23 and −34‰) δ C composition of individual n-alkanes is sensitive to (O’Leary, 1988; Chikaraishi and Naraoka, 2001; Tipple plant type, mainly due to the differences in the isotopic and Pagani, 2007). Some plants utilize a third photo- fractionation extent between growth forms (e.g. synthetic pathway, the so-called Crassulacean acid angiosperms v. gymnosperms, trees v. grasses) and 13 metabolism (CAM), producing intermediate δ C values photosynthetic pathways (e.g. C3 v. C4) (Collister et al., 13 between those of C3 and C4 plants (Bender, 1971). 1994; Diefendorf and Freimuth, 2017). The C content Algae and aquatic plants can selectively use of individual n-alkanes can thus provide important − dissolved CO2 and/or bicarbonate (HCO3 ) as carbon information regarding the OM sources. Shifts in the sources (Chikaraishi, 2014). This carbon csignal is carbon source and abundance of aquatic macrophytes altered by fractionation processes leading to more will be recorded by the mid-chain n-alkanes and negative δ13C values compared to the carbon sourse changes in the algal and microbial primary production (Fig. 11). Moreover, there is a substantial difference in mainly by the short-chain n-alkanes. Analogously, − the isotopic composition between CO2 and HCO3 due terrestrial vegetation changes will be recorded by the to equlibrium effect. Carbon isotopic fractionation in an δ13C profiles of long-chain homologues. In the − equilibrium system CO2 (gas), HCO3 (dissolved) and Mediterranean region, C4 grasses are generally rare 13 CaCO3 (solid) is dependent on the temperature; the (Cross, 1980). Consequently, shifts to higher δ C magnitude of fractionation decreases with increasing values of these n-alkanes (> −26‰) in lake sediments temperature (O’Neil, 1986). In this case, the heavier can be interpreted as an increased contribution of isotope tends to remain in the liquid phase (Gonfiantini, aquatic plants to OM producing high amount of long- 1986). The equilibrium reaction between CO2 and chain homologues (Fig. 11) (Aichner, 2009; Liu et al., − 13 − HCO3 leads to the C enrichment of HCO3 by as high 2015). In addition, other environmental conditions can 13 as 10% compared to CO2 (Fig. 11) (Peters et al., 2005; influence the n-alkane δ C values. Plants close their Chikaraishi, 2014). Aquatic plants and algae can switch stomata under highly saline or dry conditions to − from assimilation of dissolved CO2 to HCO3 , if the preserve water. This results in a decreased rate of 13 former decreases (under high productivity or alkalinity). photosynthetic CO2 assimilation leading to higher δ C This results in higher δ13C values of the assimilating values (Wooller et al., 2007; Douglas et al., 2012). organism (Prins and Elzenga, 1989; Mead et al., 2005). δD in precipitation δ13C composition of n-alkanes Hydrogen isotopes are the principal components of The δ13C compositions of individual n-alkanes can be various compounds in the hydro-, geo- and biosphere. used to differentiate between different plant types and Their conservation in proxy archives provides better track the OM sources (Chikaraishi and Naraoka, information of past hydroclimate changes (Gat, 1996; 2003; Mead et al., 2005; Fang et al., 2014). In addition Darling, 2011; Sessions, 2016). The basis for these to fractionation process related to plant type, another paleoclimate investigations relies on the fact that the fractionation step is included during lipid synthesis. D/H ratio of precipitation (δDprecip) as well as in aquatic This results in more negative δ13C values of lipids by ecosystems varies substantially over space and time in response to various climatic and environmental changes

Fig. 12. Hydrogen isotope fractionation in the water cycle (after Mügler, 2008). Various aquatic and terrestrial plants use different water pools for biosynthesis tracking the δD variability of source water. A convective system is also illustrated with two processes contributing to the amount effect and modulating the isotopic composition of water vapor (after Risi et al., 2008).

(Dansgaard, 1964). The stable isotopes of hydrogen wet seasons, the δDprecip values are mainly influenced exhibit the largest relative mass difference, and by the amount-effect. As rainfall intensities increase and consequently have different physical and chemical the heavier isotope rains out, isotopic fractionation properties; e.g. the heavier isotope has lower mobility. decreases with the ongoing condensation leading to This leads to isotopic fractionation processes, and hence lower D/H ratio values of meteoric water (Dansgaard, in a large variation in the D/H ratio accompanying the 1964; Gourcy et al., 2005). In addition, two processes physical conversion of water between the solid, liquid contribute to the amount effect (Fig. 12) (Risi et al., and vapor phases (Fig. 12) (Craig, 1961; Gourcy et al., 2008). The first is the vapor recycling process which 2005). In particular, evaporation discriminates against includes the input of already depleted vapor into the the heavier isotope (lighter isotope evaporates faster) sub-cloud layer from the unsaturated downdrafts. The resulting in more negative D/H ratios in water vapor latter becomes more intense with stronger convection. compared to original water source. On the contrary, the In the second process, the amount effect is attributed to process of condensation discriminates in favor of the low re-evaporation rate of falling rain and the more the heavier isotope leading to more positive D/H ratio efficient diffusive exchanges between raindrops and the in water vapor. As the heavier isotope is preferentially surrounding vapor at high precipitation rates (Risi et al., condensed to rain droplets, the resulting rain is enriched 2008). Model simulations suggest that the amount in D relative to the remaining vapor, which becomes effect in low-latitude regions is observed if precipitation more depleted in D with subsequent rainfall events. variations result from changes in the large-scale vertical Several geographical and climatic factors, such as motions (Bony et al., 2008). This is further confirmed the amount effect and the temperature effect, the by modern observations showing a link between elevation and distance from the shore, can influence the monthly δ18O with large-scale, not local, precipitation, δDprecip values (Fig. 12) (Dansgaard, 1964). Decreasing and especially changes in the location and convection temperatures with increasing elevations lead to intensity in the source regions of moisture (Conroy et enhanced condensation, and therefore to progressive al., 2013; Tang et al., 2015). The amount effect has also depletion of precipitation in heavy isotopes. This been shown to play an important role in water isotopes altitude effect accounts for -1 to -4 ‰ per 100 m for at the middle latitudes, with lower δD values mainly deuterium (Holdsworth et al., 1991). Additionally, resulting from higher amounts of precipitation during progressive depletion in heavy isotopes with increasing summer (Dansgaard, 1964; Lawrence and White, 1991). distance from the ocean predominantly occurs in mid- All these effects, along with others influencing the latitudes and this is known as the continental effect. source and transport of atmospheric moisture, generate (Fig. 12). For Europe, the inland gradient of D content characteristic δD signatures of water pools recorded by amounts to -3.3 ‰ per 100 km in winter and -1.3‰ per the photosynthetic organisms (Fig. 12) (Sachse et al., 100 km in summer (Rozanski et al., 1982). In high and 2012). Local climatic factors can further alter the δD middle latitude continental regions dominated by strong signal of water ecosystems (δDlake) through enhanced temperature variability, the δDprecip values are primarily evaporation rate or increased river inflow. Therefore, influenced by the temperature effect (Bowen, 2008). understanding the factors influencing the δDprecip and Lower mean air temperatures lead to lower δDprec also δDlake is crucial for the accurate interpretation of values. In tropical region dominated by distinct dry and the mechanism that drives the n-alkane δD variability.

δD composition of n-alkanes The hydrogen isotopic composition of individual n– alkanes has become an established proxy for palaeohydrological research (e.g. Eglinton and Eglinton, 2008; Aichner et al., 2015). Photosynthetic organisms use their ambient water as their primary source of hydrogen to synthesize organic compounds such as lipids (n- alkanes). Therefore, the n-alkane δD values track, in principle, the δD variability of source water modulated by climate factors as discussed above (Fig. 12) (Sachse et al., 2006; Garcin et al., 2012; Sessions, 2016). However, there is a mean offset of ca -130‰ to -160‰ between source water and lipids, attributed to biochemical isotopic fractionations during biosynthesis of n-alkanes (Fig. 13) (Sessions et al., 1999; Sachse et al., 2004). In addition, this signal can be altered by isotopic Fig. 13. Processes affecting the n-alkane δD values (adopted and modified from Aichner, 2009 and Mügler, 2008). fractionations occurring during evaporation of lake, soil water evaporation and leaf-water transpiration (Fig. 13). salinity and vegetation changes (Liu et al., 2006; Duan The δD values of short and middle chain n-alkanes and He, 2011; Yang et al., 2011; Aichner et al., 2017). produced by aquatic organisms have been shown to In particular, isotopic fractionation has been shown to correlate with lake-water δD, modulated by the balance increase in some aquatic plant n-alkanes under high between freshwater input and evaporation in close salinity potentially negating the water isotope signal ecosystems (Huang et al., 2004; Mügler et al., 2008; (Ladd and Sachs, 2012; Aichner et al., 2017; He et al., Guenther et al., 2013). Lower δD values are expected 2017). This can lead to an inverse relationship between by the enhanced input of D-depleted precipitation and salinity and n-alkane δD values (higher salinity, lower river flows, while strong evaporation can cause δD values). Moreover, the effects of vegetation changes significant D-enrichment of lake water relative to on n-alkane δD values are largely attributed to different precipitation. Analogously, the δD values of the plant physiology and biochemistry (e.g. biosynthetic terrigenous long-chain homologues (C27-C31) have been fractionation). In particular, apparent fractionation shown to reflect the isotopic signature of δDprec linked between source water and lipids for individual species to climate parameters (Daansgaard, 1964; Huang et al., has been shown to range from −204‰ to −34‰ 2004; Sachse et al., 2004, 2012 and references therein). resulting in different δD values (Sachse et al., 2102). In The offset between the δD values of terrestrial and addition, the δD values of leaf waxes has been shown to aquatic lipids (Dterr-Daq) has been proposed as a differ by as much as 70‰ for various growth forms, proxy for reconstructing terrestrial evapotranspiration with trees and ferns having the highest values compared and palaeo-hydrological conditions in lakes (Sachse et to grasses (Hou et al., 2007). This δD variability al., 2004; Mügler et al., 2008). In addition to δDprec and between different forms has further been attributed to biosynthetic fractionation, δD values of terrestrial n- various groundwater sources used by the associated alkanes are affected by soil water evaporation and leaf- organisms and the effects of micro-climate around the water transpiration (Fig. 13) (Smith and Freeman, 2006; tree and grasses leading to different degrees of evapo- Kahmen et al., 2013). This results in an isotopic transpiration (Hou et al., 2007; Sachse et al., 2012). difference between aquatic and terrestrial n-alkanes Therefore, the use of n-alkane δD values for paleo- which is mainly dependent on hydroclimatic conditions climate reconstruction is based on two assumptions: 1) (Mügler et al., 2008; Duan et al., 2011). In humid These values are correlated to the source water δD climates with positive water balances, aquatic n-alkanes values, and 2) modulation of the primary climate signal have been shown to have more negative δD values by vegetation change is limited or well-constrained because they are not affected by evapotranspirative (Sachse et al., 2012). If these two pieces of information enrichment in contrast to terrestrial n-alkanes which are not available then a complementary multiproxy show more enriched δD values. In more arid climates, analysis can help with more accurate δD interpretation. additional isotopic enrichment of the lake water will lead to the enrichment of aquatic n-alkanes compared to the precipitation source. However, the D-enrichment in δ13C and δD analysis of n-alkanes leaf water recorded in the δD values can be quite The stable δ13C and δD isotopic compositions of n- variable (from 18% and 100%) depending on climate alkanes were measured on a Thermo Finnigan Delta V conditions and leaf physiology (Kahmen et al., 2013). gas chromatograph-isotope ratio mass spectrometer Because of these confounding effects, the aquatic- (GC-IRMS) coupled to a ThermoTrace GC 2000 with a terrestrial offset approach needs to be used with care. ConFlo IV and GC-isolink system. For the δ13C Variations in the δD values of n‐alkanes are not analysis, the GC oven temperature was programmed to only caused by changes in the δDprecip itself, but are also start at 100°C increasing with a rate of 20°C/min to influenced by several environmental factors such as 250°C followed by a rate of 5°C/min until 340°C,

14 where it was held stationary for 18 mins. For the δD mortar. Aliquots of the samples were weighed and analysis, the GC oven program was set to: 100 to 250°C wrapped into tin capsules. The TOC amount was at a ramp of 15°C/min, and then another ramp of measured on samples decarbonated with HCl (10%) and 10°C/min until 320°C, which remained for 9 mins. Two dried overnight. For isotope analysis, the samples were μl of the sample were injected into the GC and helium combusted with a Carlo Erba NC2500 analyser was used as a carrier gas at a constant rate. The connected via a split interface to reduce the gas volume compounds were separated on a Zebron ZB-5HT to a Finnigan MAT Delta V mass spectrometer. The Inferno GC column. A standard mixture of n-alkanes reproducibility was calculated to be better than 0.15‰ with known isotopic compositions (reference mixture The TOC amount and the total nitrogen (TN) values A6, provided by Arndt Schimmelmann, Indiana were determined when measuring the isotope ratios. University, USA) was analyzed several times daily to The relative error was <1% for both measurements. check instrument performance and to calibrate the reference gas pulses (CO and H ) against which the 2 2 6.2.4. XRF core scanning isotope compositions were measured. The H3+ factor was determined on a daily basis and typically accounted X-ray fluorescence (XRF) core scanning has become a for 3.4-3.5 ppm mv-1. The standard deviation of the standard analytical technique for non-destructive and samples for the δ13C and δD analyses was <1 and <5 rapid analysis of sediment cores at sub-millimeter ‰, respectively. Duplicates for δ13C and triplicates for scales (Croudace et al., 2006; Francus et al., 2009). The δD were measured. The isotopic values were calculated Itrax core scanner produces high-resolution elemental with ISODAT software and are reported as mean values profiles and optical and X-radiographic images of (against VPDB for δ13C and against VSMOW for δD). cores. The scanner measures the elemental composition as element intensities in total counts second-1 (cps). The measurements reflect the relative concentration of each Bulk geochemical analysis element. Variations in elemental composition can be The bulk geochemical properties studied in this thesis connected to various processes within lakes and in included the total organic carbon (TOC) content, the catchments including shifts in minerogenic input, water carbon to nitrogen ratio (C/N) and the stable carbon level changes and redox processes (e.g. Moreno et al., isotope composition (δ13C). TOC measurements were 2007; Kylander et al., 2011). The XRF technique has performed to quantify the amount of OM contained in successfully been applied in lacustrine and marine sediments. C/N ratio was applied in the Agios Floros sediments for studying paleoenvironmental changes core to distinguish between algal and land-plant origins (Francus et al., 2009; Croudace and Rothwell, 2010). of sedimentary OM. Protein-rich algae are relatively In the present study, XRF scanning analysis was enriched in nitrogen (N) and thus tend to have C/N performed on subsamples taken throughout the Gialova ratios of <10 whereas cellulose-rich and protein-poor core to complement the biomarker results. Few samples vascular plants has C/N ratio of >20 (Meyers and were also analyzed from the lower part of Agios Floros Teranes, 2001). Intermediate C/N values indicate OM core. All the samples were placed in cubes and scanned dominated by macrophytes, or a combination of at the SLAM Lab at the Department of Geological terrestrial plants and macrophytes (Kendall et al., Sciences, Stockholm University using an ITRAX XRF 2001). Bulk δ13C composition has previously been used Core Scanner from Cox Analytical Systems to distinguish potentially sources of sedimentary OM (Gothenburg, Sweden; Croudace et al., 2006). The XRF (e.g. C3 versus C4 plants, marine versus freshwater scanner was equipped with a Molybdenum tube set at algae and macrophytes versus phytoplankton) (Meyers, 30 KV and 45 mA. The analysis was performed with a 1994). The bulk isotope approach, however, can be step size of 300 mm and integration time of 60 sec per problematic for two reasons. First, bulk isotope values measurement. Approximately five measurements were potentially record inputs of carbon from terrestrial and made on each cube surface and an average of these aquatic plants making the distinction between different measurements was calculated to acquire one value for sources dubious (Tanner et al., 2010). For instance, each sample. The elemental data were normalized by phytoplankton (C3 algae) use dissolved CO2 which is incoherent and coherent scattering to remove effects of usually in equilibrium with the atmosphere and is changes in the water content, density of the sediments therefore isotopically indistinguishable from organic during analysis and by several instrumental parameters matter produced by C3 plants in the surrounding (e.g. dead time tube aging) (Kylander et al., 2011). catchment (Meyers and Teranes, 2001; Sifeddine et al., Data interpretation for the Gialova samples was 2004). Secondly, diagenetic effects on isotopic values aided by performing principal component analysis are not well constrained, which can further complicate (PCA). The PCA is a useful tool for understanding the the interpretations of records (Tanner et al., 2010). co-variation between variables and identifying patterns In this study, the bulk geochemical data were used of correlations within the data sets. In particular, PCA as a complementary approach to the main proxy identifies those elements that show the same variability analysis. Sediment samples (resolution of 10 cm for the and are thus likely controlled by the same processes. Agios Floros core, and 5 cm for the Gialova and Lerna PCA calculations were performed using JMP Pro 12 cores) were freeze-dried and powdered with porcelain software in correlation mode with a Varimax rotation.

7. Results diatomaceous silty bands. The diatom assemblage is dominated by freshwater and indifferent salt-tolerant The results of this thesis are presented as a summary of taxa which thrive in alkaline (pH>7) waters. In terms of five papers. Paper I provides diatom-based evidence habitat preferences, there are distinct stratigraphic for tectonic- and climate-induced hydrological changes changes between sub-aerophilous, benthic and in the central Messenian plain over the last ca 6000 planktonic taxa. This implies that water level changes years. Paper II presents a detailed morphological study have played an important role in the development of the of the diatom species, Cyclotella distinguenda with a Agios Floros basin over the last 6000 years. description of a new fossil diatom species, Cyclotella Cyclotella distinguenda is the dominant planktonic paradistinguenda from the Agios Floros sequence in species which mainly occurs in the laminated sequence relation to other Cyclotella taxa. Paper III investigates (Fig. 14). It is also observed with lower amounts in the further the hydroclimate changes in Agios Floros based organic-rich sediments where diatom valves are highly on the δD values of aquatic plant‐derived n‐C alkanes dissolved. Cyclotella distinguenda exhibits large 23 morphological variability throughout the sequence. This (δD23). Paper IV presents a reconstruction of the Eastern Mediterranean hydro-climate during the last has resulted in the identification of a new species 3600 year mainly based on n-alkanes, their δ13C and δD described in the paper II. Cyclotella distinguenda values and XRF data in a sediment core from the indicates the development of lake stages in the Agios Gialova Lagoon. Paper V presents an n-alkane δD Floros basin with maximum water depth and extension. record from the Lake Lerna NE Peloponnese covering Cocconeis placentula occurs in the gyttja with a the last 5000 years. This record is compared with other wide range of variation in length, width and shape. n-alkane δD studies from SW Peloponnese providing Several varieties of this species (var. placentula, var. evidence for changes in the E-W climate gradient over euglypta and var. lineata) are encountered (Fig. 14). the peninsula during the last 5000 years. These changes Denticula kuetzingii exhibits high abundances in the are linked to shifts in the high-latitude atmospheric laminated clay and in the gyttja. This taxon is observed patterns and the effects of the African/Asian monsoons. as single, linear to elliptical valves and in chains with valve faces abutting. The most prominent morpho- characteritic is the fibulae extended to the distal margin 7.1. Paper I of the valves, forming partitions (Fig. 14). In terms of ecology, Cocconeis placentula and Denticula kuetzingii Katrantsiotis, C., Norström, E., Holmgren, K., Risberg, are here classified as benthic. The latter has a wider J., Skelton, A., 2015. High-resolution environmental range of ecology, which is also evident from its co- reconstruction in SW Peloponnese Greece covering the occurrence with Cyclotella distinguenda and Cocconeis last c. 6000 years: Evidence from Agios-Floros fen. The placentula. Denticula kuetzingii can grow in shallow, Holocene 26, 188–204. relatively deep and open water with high electrolyte https://doi.org/10.1177/0959683615596838 content and nitrate concentrations (Huntsman-Mapila et al., 2006; Matlala et al., 2008; Mackay et al., 2012). This paper presents a stratigraphic investigation of a 7.5 Ellerbeckia arenaria is abundant in the organic-rich sediment core retrieved from the Agios Floros fen, in sediments, mainly underneath peat. Valves of this the central-eastern Messenian plain, SW Peloponnese. species are usually highly fragmented and/ or dissolved The sequence was analyzed for lithostratigraphy, with outer parts missing (Fig. 14). Ellerbeckia arenaria diatoms and bulk geochemistry (δ13C, TOC, C/N) shows distinct stratigraphic changes, and therefore is complemented with XRF data obtained from selected classified as subaerophilous species to separate it from subsamples at the bottom of the sequence. The aim of the common benthic taxa such as Cocconeis placentula. this research was to reconstruct the hydrological shifts The Agios Floros sequence contains a number of in the Agios Floros paleolake over the last 6000 years benthic species that occur with lower abundances. For and to elucidate the driving mechanisms behind these instance, Gomphonema angustatum, Cymbella affinis, changes. Special focus was placed to the study of fossil Achnanthes ploenensis var. gessneri and some Eunotia diatom flora as diatom records from Greece are rare, species are present in the organic sections with <5%. due to the alkaline (carbonate) environment that hamper In summary, there is a clear stratigraphic change their preservation. Therefore, the Agios Floros study between four main species in response to water level provides the first diatom record from the Peloponnese changes during the last 6000 years. Shifts in the diatom and further offers the opportunity to improve the assemblage are further accompanied by changes in the knowledge on the Holocene diatom taxa from Greece. amount of phytoliths and the diatom preservation. In The Agios Floros sequence covers the last ca 6000 general, valves are well-preserved and the phytolith years on the basis of 18 radiocarbon dates. High amount is high during periods of higher water level and accumulation rates (up to 23 mm/yr) are recorded increased accumulation rates. These shifts are also between two decadal-long periods centered at ca 5700 accompanied by changes in the geochemical proxies. 13 and ca 5300 cal BP. The sequence consists of fen peat Negative excursions in the TOC amount, δ Corg values, in the uppermost section. This is underlain by lacustrine as well as in the C/N ratio generally correspond to more sediments, punctuated by two layers of clay with positive peaks in the planktonic taxon and vice versa.

a b c

d e f

Fig. 14. SEM photos of diatom species found in the Agios Floros sequence: (a) Cyclotella distinguenda, (b) Cocconeis placentula, (c) Denticula kuetzingii, (d) Ellerbeckia arenaria, (e) Cymbella affinis, (f) Achnanthes ploenensis var. gessneri (right) and Eunotia species (left).

The Agios Floros diatom and geochemical data vertical displacement of 1 m for each event. An offer a new insight into the paleoenvironmental increase in the flow rate of springs associated with evolution of a region where a few high resolution seismic stress in bedrock should also have contributed records are available. The results suggest that the Agios to high water levels before, during and/or after these Floros basin underwent fluctuations from marshy to events. Moreover, high amount of Denticula kuetzingii deep and open water environments during the last 6000 likely indicates changes in the lake chemistry, related to years (Fig. 15). These changes can be attributed to a increased electrolyte content and nitrate concentrations combination of tectonic, climatic and human factors. in the water discharging from the karst aquifers. At ca 6000-5800 cal BP, the basin was occupied by Apart from these tectonic processes, abrupt climatic a shallow water body with low oxygen/anoxic changes with enhanced precipitation might have played conditions. This is evident from the absence of diatoms, a role as this was a period of increased humidity in the and the presence of a coherent sandy clay layer with eastern Mediterranean region. The two deep water high peaks in Fe and S. Between ca 5700-4600 cal BP, phases were terminated at ca 5650 and ca 5280 cal BP, there is evidence for rapid water level changes in the probably due to high accumulation rates. Thereafter, Agios Floros basin. Two decade-long periods of a rapid another two events of shallow to intermediate lake development of deep lake conditions with an open levels were identified at ca 5200 and ca 4600 cal BP. water environment were identified at ca 5700 cal BP After ca 4500 cal BP, the development of peat and and ca 5300 cal BP. These two periods correspond to the disappearance of diatoms can be attributed to both diatomaceous laminated clay dominated by Cyclotella human and climate factors. This period in the Agios distinguenda, and high accumulation rates. The lakes Floros area is associated with enhanced human likely had a depth of ≤10 m based on the ecological activities (Simpson, 2014), which likely caused optima of Cyclotella distinguenda from other shallowing of the basin initiating vegetation succession. Mediterranean studies and the Alpine training set At the same time, a shift to drier conditions is also (Wunsam et al., 1995; Currás et al., 2012). These events evident from other regional records (e.g. Finné et al., can be connected to local tectonic processes, and 2017). After ca 2800 cal BP, the appearance of highly particularly, to a two-stage deepening of the basin in fragmented valves infers a return to wetter conditions as response to earthquakes along the western flank of the a response to marked rainfall seasonality, flooding fault-bounded Taygetos Mountain. Using empirical events and/or human activities. These conditions led to relationships, the possible moment magnitude of these the development of the present-day environment with earthquakes was estimated to be MW=6.7 causing a cultivated lands and seasonally semi-flooded fields.

Fig. 15. The development of the Agios Floros fen during the last 6000 years. Water in bedrock is under artesian conditions flowing through fractures and discharging as flowing artesian springs. Seismic stress at 5300 and 5700 ca BP likely caused high flow rate of the springs further contributed to deep water conditions.

7.2. Paper II tentatively classified as two different morphotypes within Cyclotella distinguenda. The need for more Katrantsiotis, C., Risberg. J., Norström, E., Holmgren, detailed investigation emerged from the uncertainties K., 2016. Morphological study of Cyclotella associated with casual microscopical observations and distinguenda and a description of a new fossil species the absence of recent studies on this species. In this Cyclotella paradistinguenda sp. nov. from the Agios study, a closer investigation of these cells was carried Floros fen, SW Peloponnese, Greece in relation to other out following detailed LM and SEM observations on a Cyclotella species. Diatom Research 31, 243–267. large number of valves. Morphological observations https://doi.org/10.1080/0269249X.2016.1211178 were combined with a simple statistical approach. The results show that there are constant and distinct This study was initiated by the results of the previous differences in the morphology and ultrastructure paleoenvironmental research from the Agios Floros fen. between the two morphotypes. These differences, along In particular, during the analysis of the diatom with their distinct stratigraphic distribution, provide succession, large amounts of two types of Cyclotella sufficient evidence to recognize a new species. On this cells were encountered with specific stratigraphical base, one morphotype is here described as a new taxon, distribution in the Agios Floros core. These cells were Cyclotella paradistinguenda. The other morphotype is generally characterized by similar valve ornamentation assigned to Cyclotella distinguenda and is further but had distinct size and structure of central areas and consistent with the original description of this taxon. stria densities. Based on routine microscopical The two species can be distinguished by a combination observations and the existing literature, they were of distinct differences in the structure and size of central

a b e

c d f

Fig. 16. (a-d) Cyclotella paradistinguenda. (a-b) External valve view and (c-d) Internal valve view. (e-f) Cyclotella distinguenda. (e) External valve view and (f) Internal valve view. The arrows show the position of marginal fultoportulae and the circles indicate the position of marginal rimoportulae. areas, the structure and number of striae, the 7.3. Paper III arrangement of marginal fultoportulae/density of costae Norström, E., Katrantsiotis, C, Finné, M., Risberg, J., between fultoportulae, the shape of alveolar chambers Smittenberg, R., Bjursäter, S., 2018. Biomarker and rimoportulae (Fig. 16). The two species also have hydrogen isotope composition (δD) as proxy for in common, striation of almost equal length, central Holocene hydro-climatic change and seismic activity in area lacking fultoportulae and a single rimoportula SW Peloponnese, Greece. Journal of Quaternary situated on a costa within the ring of marginal Science 33, 563–574. fultoportulae. Based on these similarities, the Cyclotella https://doi.org/10.1002/jqs.3036 distinguenda complex species is further proposed as a new morphological group which includes the closely- related taxa Cyclotella distinguenda and its varieties, This paper is mainly based on the analysis of the n- Cyclotella paradistinguenda and Cyclotella plitvicensis. alkane distributions and δD composition of aquatic Cyclotella paradistinguenda is only known from plant-derived n-C23 alkanes (δD23) in the Agios Floros Agios Floros, and thus its ecological interpretation can core. The goal is to complement our previous diatom- be inferred from its co-occurrence with other taxa of based study and to fill gaps in the record where diatoms known ecology and lithostratigraphic changes. are absent. This multiproxy method establishes a Cyclotella paradistinguenda occurs underneath peat, continuous, hydro-climate record for the last 6000 where the diatom abundance is low and is mainly years, integrated with other paleoclimate data from the dominated by dissolved valves of the subaerophilous Peloponnese and the Eastern Mediterranean. Another species, Ellerbeckia arenaria. Therefore, it can be aim is to validate the n-alkane δD application for this concluded that Cyclotella paradistinguenda is tolerant particular environment by comparing the outcome of of shallower conditions with lower nutrient availability the present research with our previous diatom-based and/or higher pH compared to Cyclotella distinguenda. study. This further offers the opportunity to test and re- This research could be used as a guide for relevant evaluate the water-level changes at 6000-4500 cal BP. diatom studies in the Mediterranean region helping out The carbon isotope composition of n-C23 alkane 13 13 with the identification and ecological classification of (δ C23) shows more stable values than δ C of long Cyclotella species. High taxonomical resolution and chain n-alkanes, mainly after 4500 cal BP. This ecological separation of closely-related species are also indicates a stable source for n-C23 suggesting that any essential to improve environmental monitoring and substantial effects of vegetation changes have a minor produce accurate paleolimnological reconstructions. impact on the δD23 signal. Thus, the observed

19 variability of the δD23 profile is attributed to changes in quantitative analysis of plant remains. The goal of this source water δD, modulated by the isotope signal in study is to trace the sources of sedimentary OM in the precipitation and evaporative effects on the fen water. Gialova Lagoon and to extract paleo-climate data over The δD23 signal and the n-alkane distribution support the last 3600 years. Another goal is to produce a climate the paleolimnic conditions inferred by the diatom data, reconstruction for the Eastern Mediterranean by with maximum aquatic biomarker abundance at ca 5300 comparing the Gialova data with other regional records, and ca 4600 cal BP (Fig. 17). The deep lake at ca 5700 and to explore the drivers behind the observed changes. cal BP is reflected by a slight increase in the aquatic n- The results show that the sedimentary OM has alkane abundance, likely due to low organic content in largely been derived from the local aquatic vegetation. this part of the core. Shifts in δD23 at ca 5700 and ca High TOC content coincides with high abundances of 5300 cal BP are abrupt and short-lived. They also macrophyte remains, mainly seeds from Ruppia correspond to distinct changes in the diatom maritima, and a long-term trend towards positive δ13C composition dominated by Cyclotella distinguenda values of aquatic plant-derived mid-chain n-C23-C25 13 13 (Fig. 17) as well as to high accumulation rates. It is thus alkanes. These high δ C23-δ C25 values (up to -19‰) proposed that the δD23 anomalies are caused by tectonic can mainly be attributed to carbon-limiting conditions events that led to temporal mixing of groundwater from during periods of aquatic vegetation expansion, causing − different aquifer systems, thereby altering the isotope plants to assimilate isotopically-enriched HCO3 . The 13 13 13 signal of the fen water during annual to multi-annual δ C23-δ C25 profiles co-vary with the δ C signals of periods. The δD23 signal at ca 4600 cal BP corresponds the long-chain n-alkanes, and especially with the 13 13 to lower accumulation rates, organic-rich sediments and δ C27-δ C29 records, which exhibit values of up to –20 the dominance of tychoplanktonic/benthic species. A and –22‰, respectively. It is likely that the n-C27-C29 short-lived drop in δD23 is also evident at ca 2800 cal alkanes consist of different degrees of mixing between BP. Similar hydrological changes at ca 4600 and ca aquatic and terrestrial plants triggering shifts towards 13 13 − 2800 cal BP are also observed at other sites in the positive δ C27- δ C29values due to the HCO3 effect. 13 Peloponnese and beyond. This suggests that these two The δ C record of algae-bacteria derived n-C17 events are likely attributed to large-scale climate shifts. alkane is stable throughout the sequence indicating a From a long-term perspective, the Agios Floros stable source organism (not shifting between dissolved − δD23 record indicates wetter conditions between ca CO2 and HCO3 ). The δD17 profile co-varies with δD of 6000-4500 cal BP, followed by an aridification trend all n-alkanes (δD23-31) implying that they all reflect until ca 2900 cal BP, in agreement with other eastern changes in the source water isotope composition, driven Mediterranean records. Thereafter, the δD23 signal by hydroclimate variability. In addition, the n-C31 13 indicates a return to humid conditions followed by a alkane shows lower δ C variability compared to n-C27- period of enhanced aridity and stronger seasonality C29 alkanes whereas the δD signal of the same n-alkane contrasts from 2000-1500 cal BP until modern times. exhibits the widest isotopic range in the record with an The present study confirms the ability of n-alkane average value of ca -190‰. This value is more negative δD approach to reconstruct climate changes in this by ca 20‰ compared to that of aquatic-derived alkanes particular environmental setting. The combined use of indicating that the δDprec-driven δD31 signal is further n-alkane δD approach with other proxies is also enhanced by the additive effect of evapotranspiration. promising for the identification of past seismic events. Principle Component Analysis (PCA) of the XRF data indicates shifts between dry and wet conditions. 7.4. Paper IV PC1 explains 53% of the observed variance. K, Al, Rb, Fe, Ti and Si associated with clastic sediments load Katrantsiotis, C., Kylander, M., Smittenberg, R.H., positively on PC1. These elements can be transported Yamoah, K.K.A., Hättestrand, M., Avramidis, P., by floods and rivers during wet periods. Ca and Sr Strandberg, N.A., Norström, E., 2018. Eastern driven by carbonate precipitation during dry periods Mediterranean hydroclimate reconstruction over the last load negatively on PC1. The δD17-δD31 signals 3600 years based on sedimentary n-alkanes, their generally covary with PC1 (Fig. 17). This suggests that carbon and hydrogen isotope composition and XRF both proxies are driven by a common climate signal. 13 13 data from the Gialova Lagoon, SW Greece. Quaternary Over longer time-scales, the δ C23-δ C29 records Science Reviews 194, 77–93. are generally anti-correlated with the n-alkane δD https://doi.org/10.1016/j.quascirev.2018.07.008 (δD17-31) profiles and the PC1 axis. This indicates that shifts in the expansion of aquatic vegetation are related This paper presents a 3600-year multiproxy geo- to climatic variability. Overall, macrophyte-dominated chemical record of climatic changes from the Gialova periods correspond to wet and/or cold episodes during Lagoon, SW Peloponnese. The analysis is based on a which low salinity and higher water level likely created 2.6 m radiocarbon-dated sediment core retrieved from favorable conditions for plant growth in this lagoon. the central part of the lagoon. The dataset consists of n- The records provide evidence for shifts from dry alkane distribution, their δ13C and δD values, bulk (warm) periods at ca 3600-3000, ca 1700-1300 and organic geochemistry and a XRF core scanning profile. 900-700 cal BP (late Medieval Climate Anomaly) to This approach was complemented with a semi- wet (cold) episodes at ca 3000-2700, ca 1300-

Fig. 17 Comparisons of XRF and δD31 records from the Gialova Lagoon with δD23 and the abundance of planktonic /tychoplanktonic species from the Agios Floros fen as well as with the Lake Lerna δD23, ACL and Paq records.

-900 (early Medieval Climate Anomaly) and ca 700- sources of OM in this coastal environment, and the 200 cal BP (Little Ice Age). This climatic fluctuation factors that influence its distribution. This is crucial for can be explained by the relative dominance of high- more accurate quantification of regional and global latitude and low-latitude atmospheric patterns over SW carbon budget and for forecasting changes in the carbon Peloponnese. The Gialova record between 3600 and cycling induced by climatic and anthropogenic impacts. 2000 cal BP and during the LIA resembles other central-western Mediterranean records. This suggests 7.5. Paper V that the NAO exerted the main control in SW Peloponnese during these periods. Conversely, the Katrantsiotis, C., Norström, E., Smittenberg, R.H., climatic signal from 2000 to 900 cal BP appears Finne M., Weiberg E., Hättestrand, M., Pavlos synchronous with Anatolian records, indicating that Avramidis Wastegård, S., 2019. Climate changes in the other regional atmospheric patterns such as the NCP Eastern Mediterranean over the last 5000 years and had the dominant influence for this time-interval. In their links to the high-latitude atmospheric patterns and addition, the climate variability revealed in the Gialova Asian monsoons. Global and Planetary Change 175, record is almost synchronous with shifts in the intensity 36–51. https://doi.org/10.1016/j.gloplacha.2019.02.001 of East African-Asian monsoons; drier (wetter) conditions in Gialova are associated with periods of enhanced (weak) monsoons. This can primarily be This paper presents a 5000 year record of n-alkane attributed to the more northward (southward) position distributions and their δD compositions as well as bulk of the Intertropical Convergence Zone (ITCZ) and the organic geochemistry (δ13C, TOC) from the ancient associated Subtropical High. The large-scale Lake Lerna in NE Peloponnese. The specific goals are teleconnection with Asian monsoons would also result to extract paleoclimate data from the Lerna core, and in stronger subsidence and more intense Etesians, also to compare these data with other n-alkane δD which act as katabatic winds in the western records from SW Peloponnese. Through this Peloponnese, further enhancing the summer warming. comparison, shifts in the W-E rainfall/temperature Overall, the results from the present study and their gradient over the Peloponnese peninsula are examined comparison with other paleoclimatic records from the during the last 5000 years. Another goal of this study is Mediterranean and beyond improve our knowledge of to explore the nature of the observed changes, and to the late Holocene regional climate evolution, and the assess linkages between regional climatic variability related drivers. The present study also sheds light on the and shifts in the large-scale atmospheric patterns.

The results show that there is an anti-correlation pattern 8. Discussion between the n-alkane abundances of the mid-chain and long-chain homologues. This indicates two distinct The main goal of this project was to increase the sources, which is also evident from the more positive knowledge of the environmental and climate changes in average δD23-25 values than the δD29-31values. The mid- the Peloponnese, SW Greece during the middle and late chain n-alkanes are attributed to aquatic plants Holocene (the last ca 6000 years) and to contribute with (submerged-floating). The n-C27–C29 alkanes likely new paleoclimate data for the Eastern Mediterranean. originate from the fen and the n-C31 alkane is mostly To achieve this, a multiproxy analysis, including derived from the terrestrial vegetation around the fen. diatoms, n-alkanes and their isotope composition as The n-alkane δD profiles (C23 to C31) covary with well as XRF data, was applied on three sediment cores each other indicating that they all reflect changes in the from the SW and NE Peloponnese, two areas with isotopic composition of source water. The δD records opposite climatic conditions. The results from the also covary with Paq, which has been used a proxy for multi-proxy studies include; i) reconstruction of long- lake level changes (Fig. 17). High Paq values indicate term climatic trends and identification of specific increased contribution from aquatic macrophytes and events, ii) chronological correlation of the sequences thus higher water level, which coincides with more with other records from the Mediterranean and beyond negative δD values and vice versa. This interpretation is iii) identification of non-climatic drivers influencing the supported by the higher ACL values during periods of landscape development in this region, iv) identification low Paq and positive δD trends. In the Peloponnese, of sedimentary OM sources in the Gialova Lagoon, and plants with high drought adaptive capacity have been (v) description and ecological separation of a new fossil shown to produce longer chain n-alkanes (Norström et diatom species as well as possibly association between al., 2017). This indicates that during dry periods the niche environmental evolution and diversification. sedimentary n-alkanes likely originate from drought- adapted plants and moisture stressed vegetation. Despite the general co-variability of the n-alkane 8.1. Chronology and resolution δD records, short-term terrestrial vegetation changes, The analyzed sediment cores are comparatively well- known to have taken place in this region, might bias δD dated (high number of dates) taking into account the of long chain n-alkanes over shorter timescales. long time-intervals they cover. The age-depth models Therefore, the climate reconstruction is mainly based are based on dates derived from seeds of emergent and on the predominantly macrophyte (submerged terrestrial plants as well as of floating plants that utilize /floating)-derived δD which shows the highest long- 23 atmospheric carbon. Some bulk sediments were also term variability in the record compared to δD . 29-31 dated in the Agios Floros and Lerna cores. In the The comparison of the δD records from SW and Peloponnese, further uncertainties are introduced when NE Peloponnese reveals sometimes similar and performing 14C dating on bulk samples or aquatic plant sometimes opposite climate signals between the two remains since the bedrock is made up of limestones and areas. The records infer high humidity in the peninsula will contribute with old carbon to the sediments. at ca 5000-4600, ca 3000-2600 (more unstable in SW) Aquatic plants will, during photosynthesis, take up this and after ca 700 cal BP with drier periods at ca 4100- old carbon resulting in considerably older radiocarbon 3900 and 1000-700 cal BP. On the other hand, a W-E dates (Olsson, 1991). In addition, bulk sediments can opposite climate signal is evident at ca 4600-4500, ca contain a mixture of components with widely differing 3200, ca 2600-1800 and ca 1200-1000 when the Lerna dates, related to hard water effect or inwash of old signal shows the highest δD values with a reversal at ca detrital OM from the catchment. Despite these 3900-3300, ca 3200-3000 and ca 1800-1300 cal BP. uncertainties, the dates from the Gialova and Lerna sites Concluding, this paper provides evidence for shifts appear in stratigraphic order. In addition, the Agios in the W-E precipitation gradient over the Peloponnese Floros age-depth model is robust, mainly above 200 during the last 5000 years. These changes can be cm, with some outliers between 200 and 400 cm. explained by the relative dominance of high-latitude The resolution is also an important factor when atmospheric patterns, and particularly the NAO and the comparing records with uneven temporal spacing of the NCP, over the Peloponnese. In addition, the data from samples resulting from variations in accumulation rates this paper support our previous outcome that the (Rehfeld and Kurths, 2014). In this study, the sampling African and Asian monsoonal systems have played a was carried out at a resolution of ca 5 cm. For Agios major role in the summer climatic conditions in the Floros δD , the sampling was performed at every ca 10 Eastern Mediterranean. Summer temperature changes 23 cm. Hence, the resolution is low during periods of low are particularly associated with shifts in the intensity of accumulation rates, which holds especially true for the Asian monsoons. These changes are potentially Agios Floros δD and Gialova records. This irregular superimposed on the precipitation signal reflected in the 23 sampling in time hampers the identification of short- δD records. The physical mechanism underlying the term events in SW Peloponnese. In the Lerna study, large-scale teleconnections with high latitude and low continuous accumulation rates as indicated by the age- latitude atmospheric patterns is further explored in this depth model and also the gradual boundaries between paper and the discussion part of the present summary. layers resulted in a relatively high resolution record.

The dating and resolution issues are described 8.2 Proxy data below highlighting the periods of low resolution data. Multiproxy records obtained from the same studied (i) The Agios Floros diatom record has an average sites can provide both complimentary and comparative temporal resolution of ca 52 years per sample. The information. This approach aims to identify possible highest resolution (ca 8-10 years) is observed below effects of confounding variables (i.e. local, non- 350 cm (before ca 5100 cal BP) (Fig. 17). The climatic factors). However, different proxies can be resolution above this depth ranges from 40 and 90 influenced by various environmental factors at a range years. The δD23 record has an average resolution of ca of spatial scales, and therefore show different strengths 125 years per sample, with the lowest resolution (ca and weaknesses (Birks and Birks, 2006). In this study, 230 years) above 50 cm (after 2500 cal BP). The age- special attention is given to the diatom preservation and depth model is based on 15 dates with an average how this affects the interpretation, the ecology of new dating uncertainty of ±63 years (min±18, max±160). diatom species as well as the origin of n-alkanes that There is a general overlap in the 2σ confidence intervals can influence the δD interpretation. The causes behind below 200 cm while five dates were excluded as they the variability of proxy records are also discussed, and show reversals. These dates were obtained from seeds particularly whether these changes are associated with of Cladium mariscus (emergent plant) and bulk regional climate shifts or local environmental factors. sediments and correspond to periods of abrupt water level changes. Thus, the reversals could be related to incorporation of aquatic carbon by the emergent 8.2.1. Diatom preservation vegetation or reworked material. Nevertheless, some of The availability of modern diatom data in Greece is the dates (689, 366-374, and mainly 338 cm) were limited due to the scarcity of relevant studies and the based on terrestrial seeds and appear in a stratigraphic lack of long-term ecological monitoring dataset (Ziller order. This indicates that despite some uncertainties, and Montesanto, 2004; Solak and Àcs, 2011). In our core chronology should be considered reliable. In addition, diatom-based paleo-environmental studies in addition, the comparative dating on terrestrial seeds and the central-eastern Mediterranean, and especially in bulk sediments above 200 cm indicates that there is no Greece, are relatively rare (Wilson et al., 2008; age difference, and bulk material can be considered Cvetkoska et al., 2014). In addition, there are no suitable for age determination within the peat section. previous diatom studies in the Peloponnese. This can be

(ii) The Gialova record has an average resolution of 71 attributed to unfavorable conditions for diatom growth years per sample ranging from ca 100-160 years per and preservation due to the carbonate bedrock and/or sample at the depth of 150-90 cm (2600-800 cal BP) to low silica availability. Therefore, the limited regional ca 20-70 years per samples in the rest of the sequence. diatom database and the variable degree of valve The age-depth model is based on seven dates with an preservation in sediments potentially lead to some average uncertainty of ±50 years (min ±35, max ±88). uncertainties regarding the interpretation of changes in The dates were mainly obtained from seeds of the the composition of fossil flora in the present study. rooted-floating or free-floating plant, Najas, which The Agios Floros diatom record shows profound changes in the dominant species abundances indicating takes up CO2 from the atmosphere (Prasad et al., 1997). Two dates are based on charcoal and terrestrial seeds lake level shifts in this small tectonic basin over the last respectively, and they all appear in stratigraphic order ca 6000 years. Overall, the Agios Floros study provides which strengthens the core chronology. Dates obtained a unique diatom record from the Peloponnse, which is from seeds of the aquatic plant, Ruppia maritima show also evident from the absence of diatoms in the Gialova a deviation of ca 100 years from the model interpreted and Lerna cores. However, valve preservation in the Agios Floros core is poor to moderate. High amount of as an indication of a minor potential reservoir effect. dissolved valves, including dissolved fragments, mainly (iii) The Lerna record has an average temporal occurs in the organic-rich accumulations (gyttja and resolution of 48 years/sample. This is the highest peat) compared to the diatom-rich clay (Fig. 18). achieved resolution for isotope/microfossil-based study Poor preservation can be related to both physical of a sediment core in the Peloponnese. The age-depth breakage and chemical dissolution prior to and during model was constructed using seven dates from seeds of the accumulation of valves in sediments. Highly emergent plants and bulk sediments. The average dating alkaline conditions can be one factor affecting diatom uncertainty is ±68 years (min±25, max±91) after preservation in this carbonate system. The comparison calibration. The dates appear in stratigraphic order, of dissolution index with accumulation rate indicates although two of them (one bulk sample and one sample that these two factors are related to each other and they with a charred appearance) showed reversals and were are both associated with the diatom abundance (Fig. excluded from the model. The reversed date from the 18). High diatom abundance (diatomaceous clay) and bulk sample at 222 cm corresponds to a clay layer low dissolution index (well-preserved valves) mainly which is barren of plant remains. This indicates that occur during periods of high accumulation rates. The even low contamination levels from old carbon could latter potentially promotes diatom preservation as the have larger effects on this date compared to the dates deposited valves are quickly removed from the related to organic-rich accumulations (Olsson, 1991). sediment–water interface and uppermost sediment

13 Fig. 18. Vertical profiles (against depth) of bulk sediment geochemistry, δ C values of n-C23 alkane, accumulation rate (AR), abundance of diatoms and dissolution index. The red line marks the depth with high abundances of Cyclotella paradistinguenda. layers where most of dissolution occurs (Flower, 1993; in the valve preservation can further be attributed to the Flower and Ryves, 2009). However, diatom valves shallowing of the lake (desiccation effect) and/or to would still continue to dissolve if pore waters remained decreased sedimentary burial rates which would cause undersaturated with regard to silica (Flower, 1993). valves to remain exposed to water for longer periods. Therefore, high pore-water concentrations of silica (biogenic silica production is high enough to saturate 8.2.2. The new diatom species and its affinity pore waters) during periods of high accumulation rates would prevent sedimented diatom valves from The Cyclotella genus is the most taxonomically, dissolving. In addition, well-preserved valves occur in ecologically, and morphologically diverse genus of the diatomaceous sequence, which corresponds to deep freshwater planktonic diatoms. This genus contains water conditions (peaks in planktonic taxa), indicating more than 300 species, characterized by extensive the role of limnological conditions in the valve polymorphism and biogeographical variability. Several preservation. Anoxic and calm water environment have morphological groups, each consisting of similar taxa, been linked to enhanced valve preservation, as have been divided within the Cyclotella genus. Species- bioturbation/mixing is being reduced (Flower and specific character variability has been linked to life- Ryves, 2009). On the other hand, lower water level, cycle variation, environmental conditions and genetic increased turbidity and high energy environments can differentiation (Mann, 1999; Kociolek and Stoermer, affect the quality of valves through physical breakage 2010). The average cell size of a population of centric which in turn can speed up the dissolution rate (Flower, species decreases during its life cycle as vegetative cell 1993; Reed, 1998; Ryves et al., 2006). Other factors, division leads to the production of daughter valves such as temperature and salinity, might also have which are smaller than the parent valves (Jewson, 1992; played a role (increased dissolution under high Beszteri et al., 2005). This size diminution can be salinity/temperature; Ryves et al., 2006). However, the accompanied by changes in valve shape, structure and Agios Floros fossil flora does not indicate substantial density of striae, position of marginal fultoportulae and changes in salinity. In addition to environmental shape of rimoportulae (Håkansson and Chepurnov, factors, valve preservation depends on the size and 1999, Meyer et al., 2001, Prasad and Nienow, 2006). degree of silicification of different species (Stoermer Cell wall variability can also be influenced by shifts in and Smol, 2010). Robust and more silicified species are the environmental conditions such as changes in more resistant to dissolution and also to physical salinity, nutrient availability and degree of silification. breakage compared to finer taxa (Ryves et al., 2006). These factors have been shown to affect the valve The variable valve preservation in the Agios Floros diameter, the density and shape of striation and costae, sequence might affect the representativeness of fossil the density of marginal fultoportulae and shape of assemblage. Large and heavily silicified species (e.g., satellite pores (Håkansson and Korhola, 1998; Roubeix Cyclotella paradistinguenda and Ellerbeckia arenaria) and Lancelot, 2008; Shirokawa et al., 2012). replace relatively smaller forms (e.g. Cocconeis and Understanding the mechanisms behind morphological Denticula). This can be associated with changes in the diversities is thus crucial to identify morphoboundaries diatom preservation (Fig. 18). During the transition to between different taxa and to establish new species. more robust forms (mainly above 350 cm), the overall In the Agios Floros core, species-specific character preservation seems to deteriorate, and consequently the variability between the two types of Cyclotella cells is more fragile taxa may be underrepresented. This change probably not linked to life-cycle variation of the same

24 species as larger and smaller valves occur in each The latter is evident from the two older lake stages population. In addition, there is no evidence for characterized by diatomaceous sediments deprived of heterovalvate frustules, although the number of intact plant remains. For the lake stage, at ca 4600 cal BP, the cells encountered in the core was low. The occurrence low C/N ratio (ca 12) points out that the proportion of of heterovalvy is attributed to the effect of changing allochthonous input is small with possibly higher environment or life-cycle variation resulting in contribution from algae, which would give a more morphological differences between the two valves of a positive imprint in the δ13C values. This suggests that single frustule (Theriot, 1987; Cox, 2014). In this case, Cyclotella paradistinguenda lacks the isotopic signature valves of the same frustule formed at different time and specific for planktonic species, which supports its possibly under different conditions could show distinct tychoplanktonic/benthic rather than planktonic affinity. ornamentation pattern, for example, by changing focus Cyclotella paradistinguenda is only known from level under LM. The situation related to heterovalvy the Agios Floros site without any extant populations. also involves morphological variation within a taxon This indicates that it is likely an endemic species, that is not necessarily continuous, and can alternate representing the Holocene morphological evolution of with a change in environmental conditions or cell cycle the Cyclotella genus in the Peloponnese. Cyclotella (Kociolek and Stoermer, 2010). In the Agios Floros paradistinguenda was likely favored by its capacity to core, the two type of cells have distinct and constant grow in the environmental settings, which prevailed in character differences and can easily be distinguished the Agios Floros fen during the lake event at 4600 cal even if found side by side. Their specific distribution BP (alkaline/shallow water/low nutrient availability). throughout the sequence also indicates that their ecological separation is important to produce a more 8.2.3. Origins of n-alkanes accurate palaeoenvironmental reconstruction. Therefore, the consilience between the morphological Sediment samples contain n-alkanes that have been and palaeo-ecological data suggests that the two deposited over years from multiple source plants and populations should be treated as distinct, and seems to vegetation. Therefore, no single n-alkane homologue provide sufficient evidence to recognize a new species. (as extracted from a sediment sample) is expected to Cyclotella paradistinguenda is currently classified represent an isolated plant group (Norström et al., as tychoplanktonic/benthic taxon indicating lower water 2018). For instance, long-chain n-alkanes can some- depth, and highly alkaline and/or even oligotrophic times originate from aquatic sources or a mixture of conditions compared to Cyclotella distinguenda. This aquatic and terrestrial sources (Aichner, 2009). This interpretation is based on the correlation between the relative proportion of contributing plant functional distinct distribution of these species and changes in the types can bias the interpretation of n-alkane δD signal. litho-stratigraphy, as well as the accompanying diatom In addition, δD records can be influenced by various flora. As discussed, Cyclotella paradistinguenda mainly environmental factors including salinity and vegetation occurs underneath peat, along with Ellerbeckia changes. Therefore, distinction between n-alkanes of arenaria. In this depth, the diatom abundance is low various genetic origins is crucial for a more accurate with high amount of dissolved vales than in the depths interpretation of the δD profiles of individual n-alkanes. where Cyclotella distinguenda is the dominant species. One way to trace the origin of fossil n-alkane is to The tychoplanktonic/benthic affinity of Cyclotella perform n-alkanes analysis on the present vegetation. In paradistinguenda is further supported by the bulk and particular, Norström et al., 2017 analyzed the n-alkane n-alkane carbon isotope data (Fig. 18). The δ13C values concentrations and chain-length distributions of the of planktonic species are generally lower (−32 ± 3 ‰) most common species of the modern macchia and than those of benthic taxa (−26 ± 3 ‰) (France, 1995; phrygana vegetation in SW Peloponnese (Agios Floros Doi et al., 2010). This difference is attributed to the area). The results indicate that the n-alkane production boundary layer (a film of water that sticks to the of drought adapted phrygana herbs and shrubs are six surface) effect. Benthic species grow in still water and times higher than the common local reed. In addition, have thicker boundary layers resulting in a great the reed vegetation is mainly dominated by the n-C27– diffusion resistance and in entrapment of otherwise C29 alkanes, while phrygana vegetation and Olea 13 discriminated C (Wang et al., 2013). On the other europaea primarily produce n-C31–C35 (Norström et al., hand, planktonic algae are exposed to a greater water 2017). This vegetation also dominates the area around turbulence which reduces the boundary layer thickness the Gialova Lagoon and the Lerna site suggesting causing more severe 13C depletion (France, 1995). significant contribution of these plants to the fossil The Agios Floros δ13C values (bulk and the n- sediment n-alkane pool. However, vegetation in these alkane molecules) are more negative at ca 5700 and at areas has changed during the Holocene and this might ca 5300 cal BP where Cyclotella distinguenda is the have also caused changes in the n-alkane sources. The dominant species compared to the lake stage at ca 4600 interpretation about the origin of n-alkanes can further cal BP dominated by Cyclotella paradistinguenda (Fig. be strengthened by combining the analysis on the 18). This difference is likely related to the specific present vegetation with various methods including the carbon isotope signature of different species assuming down-core n-alkane abundances, their δ13C values and much of the sedimentary OM is derived from algae. analysis of plant remains in the sediment cores. In the

Gialova study, the combined results from the n-alkanes hydroclimate signal i.e. the two proxies support each isotope analysis and the plant remains analysis indicate other. In the Lerna record, there is a good variability of different sources of the individual n-alkanes; n-C17 is all n-alkane δD profiles with the Paq and ACL indexes. derived from bacteria/algae, n-C23-C25 from aquatic These indexes reflect changes in the abundance of plants, n-C27-C29 aquatic and terrestrial vegetation aquatic macrophytes and drought sensitive vegetation, (local fen) and n-C31 mainly comes from terrestrial respectively, which can be attributed to climate factors. vegetation (Phrygana and Olea europaea) (Katrantsiotis Previous studies have indicated that the Gialova et al., 2018). In the Lerna core, no carbon isotope Lagoon and the Lake Lerna were isolated from the sea analysis was performed and macrofossil remains were during the studied time-intervals (Zangger, 1991; examined at a lower resolution. However, changes in Emmanouilidis et al., 2018). Therefore, the δD signal in the carbon chain length distribution clearly show two water has largely been determined by the balance distinct sources: n-C23-C25 likely from aquatic plants between direct precipitation amount and evaporation (submerged and/or floating plants) and n-C27-C31 from rate. The water supply in the studied sites mainly relies terrestrial vegetation (Katrantsiotis et al., 2019). on meteoric water deriving either directly from precipitation or via rivers and spring flows. Groundwater δD values from aquifer springs in the 8.2.4. Factors behind δD variability Peloponnese show strong correlation with δDprec The development of compound specific δD analysis has confirmed by low (high) spring flows during dry (wet) a great potential for hydroclimatic reconstructions. seasons (Dotsika et al., 2018; Norström et al., 2018). However, a number of environmental factors influence In the eastern Mediterranean region, the water 18 the δD values of individual n-alkanes, which can in turn isotope composition in precipitation (δ Oprec, δDprec) is bias the paleoclimate reconstruction. This holds mainly influenced by the temperature with cooler especially true for the Peloponnese peninsula, where the conditions leading to lower δDprecip (Fig. 19a) (Argiriou landscape development has been influenced by various and Lykoudis, 2006; IAEA, 2014; Bolin Centre environmental processes including climate changes, Database, Mouzaki, 2017). The amount effect also tectonic movements, sea level changes as well as plays an important role; high rainfall amounts lead to human impacts. The Peloponnesian sites studied in this more negative isotope values (Fig. 19b) (IAEA, 2014; thesis are especially characterized by a long history of Bolin Centre Database, Mouzaki, 2017). On an annual human activities which had a major influence on the basis, this relationship is valid for rainfall amount ≤100 vegetation (Jahns, 1993; Zangger et al., 1997; mm while on a seasonal basis the amount effect is Papazisimou et al., 2005). Humans have also influenced stronger during summer rains (Argiriou and Lykoudis, the hydrology of these ecosystems through diverting 2006). Other factors affecting δDprecip are changes in air rivers or draining lakes during different periods of the mass trajectories, associated with the relative Holocene (Zangger, 1991; Zangger et al., 1997). In dominance of westerly-northwesterly derived D- addition, the Gialova and Lerna sites are located close depleted air masses and southwesterly-derived D- to the shore and thus sea level changes might further enriched air masses (Argiriou and Lykoudis, 2006; impact the n-alkane δD records (e.g. lower δD values Dotsika et al., 2010). The isotopic signal of these air during periods of high salinity). Therefore, variation in masses may further be modified by the water vapor sedimentary δD records can be influenced by both produced in the Mediterranean (Dotsika et al., 2010). climate and non-climate factors and this must be The δ18O composition in laminated speleothems understood to improve climate reconstructions. One from the central-eastern Mediterranean region is mainly way to optimize the interpretative accuracy is to apply a governed by the amount effect (e.g. Bar-Matthews and multiproxy approach derived from independent proxies Ayalon, 2004; Zanchetta et al., 2014; Psomiadis et al., (terrestrial and aquatic). Multiproxy studies have the 2018). The Peloponnesian δD records show close advantage of providing both complimentary and resemblance to speleothem δ18O profiles in this region comparative information from an area with a complex (Katrantsiotis et al., 2018, 2019) indicating that the landscape development. This can assist in identifying amount effect has played an important role in the δD synchronous changes between different proxies that variability. The effects of air temperatures are likely could be attributed to a common climate mechanism. superimposed on the amount effect process and can be In the present study, the n-alkane δD records regarded as a response to the large temperature generally covary with each other indicating that they all variation between summer and winter. Higher summer reflect changes in δD composition of source water temperatures would enhance the δD signal though the modulated by a common climate factor. This is further intense evaporation of source water and leaf water. evident from the n-C17 alkane in the Gialova record and Nevertheless, on both monthly and annual base, there is the n-C23 alkane in the Agios Floros which have a stable an anticorrelation between the mean air temperature and source organism (relatively stable δ13C profile) and, total precipitation amount in the Peloponnese (Fig. 19c therefore, the δD variability of these n-alkanes can be and 19d; r=-0.57 and r=-0.69 for the period 2002-2015) considered as a result of shifts in δD of water. Τhe (IAEA Patras, 2014). This indicates that over longer Gialova δD profiles also covary with the XRF data time scales, periods of lower precipitation were likely indicating that both proxies respond to a common accompanied by higher temperatures and vice versa.

input. Therefore, higher δD values in the records indicate drier (and warmer) conditions, whereas lower δD values represent wetter (and colder) conditions. In summary, the above data suggest that a common climate signal override the vegetation impact changes at least over longer time-scales while any effect of increased fractionation under high salinity appears to a have a minimum impact on the δD records (cf. Aichner et al., 2017). However, abrupt vegetation changes and tectonic activities might have affected the δD signals over shorter periods. Hence, the climate reconstruction presented in this thesis is based on the long-term trends avoiding interpreting single peaks that might represent changes in non-climate (e.g. tectonics) processes. Apart from the application of multiproxy analysis, another way to optimize the interpretative accuracy is the comparison of proxy records from SW and NE Peloponnese and with other regional paleoclimatic studies. This may give further evidence of synchronous induced climate changes. Features that appear only in b individual records most likely represent non-climate variability. On the other hand, changes that occur across regional records likely represent variations in climate.

8.3. The role of tectonics in abrupt hydrological changes in the Agios Floros fen

The diatom and isotope data in the Agios Floros core reveal climate induced long-term trends which resemble other paleoclimatic signals from the Peloponnese and beyond. In addition, both proxy data infer abrupt and c short-term hydrological changes mainly at ca 5700 and ca 5300 cal BP. These two events are likely related to local environmental factors as they do not appear in other records from the central-eastern Mediterranean. The Agios Floros basin is bounded to the east by an active fault system, i.e. the Agios Floros-Pidima fault segment, which is a part of the Eastern Messinia Fault Zone (Fig. 20) (Kassaras et al., 2018). A significant number of historical and instrumental earthquakes have been assigned to this specific fault segment. Most recently, in the period 2011 June–October, a tectonic d swarm of 1222 earthquakes occurred in the hanging wall of this normal fault zone migrating from NNW (upper Messinia Plain) towards SSE (Agios Floros) (Fig. 21) (Kyriakopoulos et al., 2013). In addition, the Agios Floros-Pidima fault segment is connected to a Fig. 19. Data from IAEA Patras, NW Peloponnese, for the karst spring system on the foot of the Taygetos period 2001-2015. The correlation between (A) mean Mountains. The hydrology of these springs is largely monthly temperature and δD of precipitation (B) monthly depending on the rainfall amount but also other tectonic amount of precipitation and δD of precipitation, (C) mean monthly temperature and monthly amount of precipitation, processes. Increase in the flow rate of springs from and (D) mean annual temperature and annual precipitation. other areas with similar settings have been attributed to The latter is based on data for the period 2002-2015. anomalous behaviors of aquifers, due to seismic stresses before or after earthquakes (Kissin and Taking these factors into account, the variability of Grinevsky, 1990; Muir-Wood and King, 1993; Esposito δD n-alkanes from the three studied sites is regarded a et al., 2001). Taking into account the high seismicity combined response to: (i) δDprec, linked to rainfall and location of the studied site next to the springs, the amounts, (ii) evaporation of source water and leaf abrupt and short–term water level changes observed in waters, coupled to air temperature and (iii) δDlake (lagoon), the Agios Floros record can largely be attributed to determined by evaporation and precipitation/freshwater local tectonic processes and the associated earthquakes.

trigger water level rise that would be high enough (possibly ≤10 m) to promote the dominance of planktonic species i.e. Cyclotella distinguenda, and the formation of laminated sediments. In addition, these events are accompanied by high accumulation rates that can largely be associated with the landscape instability after earthquakes leading to the shallowing of the lakes. There is additional evidence for abrupt hydrological changes in the Messenian plain between 6000-5000 cal BP. Papazisimou et al., 2005 reported similar lake level changes in the central Messenia plain based on a pollen study from the Agios Floros fen. Kraft et al., (1975) found evidence for a sharp rise of 9 m in the relative sea level that flooded the lower Messenian plain between 5700 and 5300 cal BP. Engel et al., (2009) showed that a maximum landward shore displacement occurred at ca 5000 cal BP. Further to the south, ca 50 km from the site of Agios Floros, archaeological evidence indicate that the Alepotrypa cave entrance collapsed due to an earthquake at ca 5200-5000 cal BP (Boyd, 2015). In the western Greece, multiple regional tectonic events were detected at ca 6000-4500 cal BP, based on relative sea level (RSL) curves for seven coastal areas in NW Fig. 20. Topographical map of the Messenian plain and the Greece and the Peloponnese (Vött, 2007). These surrounding area showing the Eastern Messinia Fault Zone regional tectonic events were generally characterized by (Black lines). The Agios Floros area with the springs and the changes in the uplift/ subsidence rates and/or by the Agios Floros-Pidima fault segment are depicted in the box. redirection of local tectonic movements. In addition, a sequence of coseismic uplift movements were reconstructed at the Perachora Peninsula at the eastern shore of the Gulf of Corinth, NE Peloponnese at ca 6000-4500 cal BP (Pirazzoli et al., 2004). Therefore, it can be concluded that the tectonic events identified in the Agios Floros fen seem to coincide with a period of intense tectonic activity on a more regional scale (a period of high frequency–high magnitude earthquakes). In addition to tectonic activities, climate changes associated with enhanced precipitation might have contributed to the Agios Floros lake stages. The period between 6000-5000 cal BP is generally characterized by wetter conditions in the wider region of the Eastern Mediterranean (Finné et al., 2011). Therefore, basin subsidence would be accompanied by high rainfall events, which would also enhance the spring discharge. There is a lack of archaeological studies in the Agios Floros area that could allow us to speculate about human influence in the landscape changes between Fig. 21. Map of the Agios Floros and the surrounding area 6000 and 5000 cal BP. Nevertheless, the formation of (inset box from Fig. 20) showing the magnitudes and the lakes at ca 5700 and 5300 cal BP are more likely caused epicenters of earthquakes occurred in June–October 2011. by a natural mechanism, which is evident from the The map also highlights the seismicity migration towards deepening of the basin as described above and the south (reproduced from Kyriakopoulos et al., 2013). undisturbed laminated sequence. Evidence for vegetation changes associated with human activities, The Agios Floros basin has a maximum depth of 2 however, have been found in the lower Messenian plain m with an outlet to the south (Norström et al., 2018). and the slopes of Taygetos after 6000 cal BP (Engel et Therefore, the occurrence of high-water events in this al., 2009). This might add the possibility that the shallow basin at ca 5700 and ca 5300 cal BP would infilling of the lakes after these events would have been require deepening (subsidence) of the ground surface caused by high erosion related to human activities on possibly associated with two earthquakes. The the nearby slopes of the Taygetos Mountain range. combination of increase in the flow rate of springs The lake event at ca 4600 cal BP is different from before the earthquakes with the basin subsidence would the events at 5700 and 5300 cal BP. The former; 1) is

28 dominated by deposits consisting of gyttja clay rather plotted with selected Mediterranean records on a map to than diatomaceous laminated clay, 2) is characterized visualize the spatial distribution (Table 1, Fig. 22). by a diatom flora dominated by the tychoplanktonic 5000-4500 cal BP. Wetter conditions dominate the /benthic species Cyclotella paradistinguenda rather Peloponnese until 4600 cal BP when the NE part of the than the planktonic species Cyclotella distinguenda peninsula becomes drier between 4600-4500 cal BP. (Katrantsiotis et al., 2016), 3) contains a higher amount There is evidence for wetter and warmer conditions in of aquatic biomarkers, and 4) shows less abrupt isotopic the Aegean Sea during this period (Triantaphyllou et proxy response and is more extended in time (Norström al., 2009; Psomiadis et al., 2018). In the wider et al., 2018). These factors suggest the dominance of Mediterranean region, high humidity is reported from more shallow water, during the lake event at 4600 cal southern Italy (Magny et al., 2011) and also the Levant BP compared to the previous events. In addition, shifts region (Bar-Matthews and Ayalon, 2004; Cheng et al., in the n-alkane δD values are observed in the Lake 2015). Conversely, increased aridity is inferred from the Lerna at ca 4600-4500 cal BP, and thus these changes Anatolian region (Göktürk et al., 2011; Kuzucuoğlu et can be attributed to a regional climate mechanism. al., 2011), northern Italy (Magny et al., 2007) and the western Mediterranean, mainly from the Iberian Peninsula (Frigola et al., 2007; Jalut et al. 2009). 9. Regional climate reconstruction 4500-4100 and 4100-3900 cal BP. High humidity is evident in the Peloponnese until ca 4200 cal BP The climate reconstruction presented in this thesis is followed by an aridification trend at ca 4100-3900 cal based on the correlation of the Peloponnesian n-alkane BP. The wet conditions in the SE Aegean Sea terminate δD records revealing the dominance of sometimes at ca 4300 cal BP (Triantaphyllou et al., 2009). In the similar and sometimes opposing climate signals northern Aegean Sea, there is evidence for drier between NE and SW Peloponnese. This reconstruction conditions after ca 4100 cal BP (Psomiadis et al., 2018) is further supported by other paleoclimatic studies in while similar conditions are reported from the western the peninsula. The sedimentary Asea valley record in Balkans (Zanchetta et al., 2012). A shift towards high the central Peloponnese shows a similar pattern to Lake humidity is reported from the Anatolian region Lerna δD23, mainly before 2000 cal BP (Unkel et al., (Göktürk et al., 2011; Kuzucuoğlu et al., 2011) and the 18 2014). The Kapsia cave δ O record in the central- northern Italy with more unstable conditions at ca 4100- eastern Peloponnese covering the period between 2950 3900 cal BP (Magny et al., 2007). An aridification trend and 1170 cal BP exhibits similar long-term trends as the superimposed by more unstable conditions is reported Lake Lerna δD23 signal (Finné et al., 2014). In SW from the Levant region (Bar-Matthews and Ayalon, 18 Peloponnese, the Alepotrypa Cave δ O record bears a 2004; Cheng et al., 2015). Drier conditions are evident close resemblance to Agios Floros δD23, and especially in southern Italy (Magny et al., 2011) and also the to the Gialova Lagoon record (Boyd, 2015; Iberian Peninsula throughout these time-intervals 18 Katrantsiotis et al., 2019). The Mavri Trypa Cave δ O (Martín-Puertas et al., 2008; Jalut et al. 2009). The signal from this area, however, shows a hiatus between picture of the time at ca 4200 cal BP in the Eastern 2950 and 2200 cal BP which has been attributed to both Mediterranean as a pronounced dry period is supported drier conditions and to the rerouting of water through in records from across the region and also coincides the bedrock or human intervention in or near the cave with the decline of the Akkadian Empire in northern (Finné et al., 2017). In addition, this record shows Mesopotamia and the Old Kingdom of Egypt (Weiss et wetter conditions between 1800 - 1650 cal BP followed al., 1993; Cullen et al., 2000 Stanley et al., 2003). by an aridification trend when the Gialova and Agios Floros signals also point to a period of enhanced aridity. 3900-3000 cal BP. This period is characterized by an The Peloponnesian δD records are divided into opposite climate signal in the Peloponnese; wetter different climate zones representing periods of humidity conditions are evident in NE and drier conditions in SW (and temperature) changes. These signals are further with a reversal at ca 3300-3200 cal BP. More unstable integrated with other paleoclimate studies from the conditions dominate the northern Aegean region with Mediterranean region to provide an updated assessment higher humidity at ca 3700-3500 and after ca 3400 cal of the spatial climate variability during the last 5000 BP (Psomiadis et al., 2018). Alternate wet/dry episodes years. The comparison reveals large intra-regional are reported from the central Anatolian (Kuzucuoğlu et variability. The Agios Floros-Gialova records appear al., 2011) and the southern Levant (Bar-Matthews and sometimes synchronous with the western-central Ayalon, 2004). A wetter climate regime is inferred from Mediterranean studies (e.g. 3900-3000, 3000-2600, the northern Levant (Cheng et al., 2015) and NNW 700-300 cal BP) while the Lake Lerna signal sometimes Turkey (Göktürk et al., 2011). An aridification trend is matches other records from the eastern Mediterranean reported from the western Balkans (Zanchetta et al., (e.g. 4500-4200, 2600-1800 cal BP). However, similar 2012), in line with other records in southern (Magny et signals between SW and NE Peloponnese can also be al., 2007; Piccarreta et al., 2011) and northern Italy, observed during periods of large-scale events (4.2 and which indicate lower lake levels interrupted by a wet 2.8 ka events, MCA, LIA). The periods described spell at ca 3300 cal BP (Magny et al., 2007, 2011). In below are based on the Peloponnesian signals, also the Iberian Peninsula, most of the records report

29 enhanced aridity during this period (Martín-Puertas et Matthews and Ayalon, 2004) in contrast to the northern al., 2008 and references therein; Jalut et al., 2009). Levant (Cheng et al., 2105). At the same time, a trend

towards high humidity is also inferred from southern 3000-2600 cal BP. The Peloponnesian δD records Italy (Piccarreta et al., 2011; Grauel et al., 2013; Sadori generally show the dominance of wetter conditions in et al., 2016). On the other hand, enhanced aridity is the peninsula with more abrupt hydrological shifts in reported from the western Balkans (Zanchetta et al., SW Peloponnese. High humidity is inferred from the 2012) and northern Italy (Magny et al., 2007). Further northern Aegean region and the central Greece until ca west, a trend towards drier conditions is revealed in the 2600-2400 cal BP (Pavlopoulos et al., 2006; Psomiadis Iberian Peninsula mainly after ca 1600 cal BP (Martín- et al., 2018). More arid conditions are mainly reported Puertas et al., 2008; Nieto-Moreno et al., 2011). from NNW Turkey (Göktürk et al., 2011) and the northern Levant (Cheng et al., 2015) compared to the 1300-700 cal BP. Medieval Climate Anomaly (MCA). central Anatolian (Kuzucuoğlu et al., 2011) and There is a trend towards an abrupt wetness at ca 1300- southern Levant region (Bar-Matthews and Ayalon, 1200 cal BP followed by more arid conditions in SW 2004). In general, there is a trend towards enhanced Peloponnese. On the other hand an aridification trend is aridity in the Eastern Mediterranean after ca 3200 cal evident in NE Peloponnese throughout this period. BP (Roberts et al., 2001; Finné et al., 2011; Kaniewski More unstable conditions are reported from the North et al., 2013). Conversely, wetter or more unstable Aegean Sea with an arid and warmer period at ca 1000- conditions compared to the previous period are reported 800 cal BP (Gogou et al., 2016). Similar to SW from the western Balkans (Zanchetta et al., 2012), and Peloponnese, a shift from wetter to drier conditions is southern and northern Italy (Magny et al., 2007, 2011). inferred from the central Anatolian (Kuzucuoğlu et al., In the western Mediterranean (southern Iberia) there is 2011; Dean et al., 2015) whereas an aridity signal a shift from a prolonged period of drought to wetter through this period is evident from NNW Turkey and conditions from ca 3000-2800 cal BP (Martín-Puertas et the southern Levant region (Bar-Matthews and Ayalon, al., 2008 and references therein). This coincides with an 2004). More arid conditions are also inferred from the increased wetness in many European bogs (Borgmark western Balkans (Zanchetta et al., 2012). In general, and Wastegård, 2008; Engels and van Geel, 2012). records from the central Mediterranean are in

agreement with the Gialova record showing the 2600-1800 cal BP. Roman Warm Period (RWP). This dominance of enhanced aridity in the second half of the period is characterized in the Peloponnese by a SW-NE MCA (Magny et al., 2007; Piccarreta et al., 2011; opposite climate regime with intense aridity prevailing Grauel et al., 2013; Sadori et al., 2016). Conversely, in the NE part. There is evidence for drier conditions in drier condition are reported from the western the north Aegean Sea during this period (Psomiadis et Mediterranean, mainly from the southern part of the al., 2018). An aridification trend is also reported from Iberian Peninsula throughout the MCA period (Martín- NW Turkey (Göktürk et al., 2011) and the Levant Puertas et al., 2008; Nieto-Moreno et al., 2011). region (Bar-Matthews and Ayalon, 2004; Cheng et al., 2015), whereas records from central Anatolia show 700-200 cal BP. Little Ice Age (LIA). The records in both wet (Kuzucuoğlu et al., 2011) and dry conditions the Peloponnese generally infer the dominance of (Dean et al., 2015). In the western Balkans and Italy, wetter conditions, which are more intense in the SW studies provide no coherent picture. However, it seems parts. Increased wetness is reported from the North that relatively drier conditions prevail in the northern Aegean Sea after ca 600 cal BP, accompanied by Italy (Magny et al., 2007) and Sicily (Magny et al., fluctuations in SSTs (Gogou et al., 2016). The records 2011) while in the Gulf of Taranto region (Piccarreta et from the Eastern Mediterranean do not provide a al., 2011; Grauel et al., 2013) and the western Balkans coherent picture. More unstable conditions are reported (Zanchetta et al., 2012) a wet period is followed by an from NNW Turkey (Göktürk et al., 2011) while a shift aridification trend after ca 2000 cal BP. In the Iberia from wet to dry conditions is reported the southern peninsula, paleoclimatic records indicate the gradual Levant (Bar-Matthews and Ayalon, 2004). Both wet dominance of wetter and warmer conditions (Martín- and dry conditions are indicated from the central Puertas et al., 2008; Nieto-Moreno et al., 2011). Anatolian (Kuzucuoğlu et al., 2011; Dean et al., 2015). Enhanced aridity is inferred from the western Balkans 1800-1300 cal BP. Dark Ages Cold Period (DACP). (Zanchetta et al., 2012). In the wider region of the The Peloponnesian records infer a reversal of the central Mediterranean, there is evidence for the previous W-E climate signal, with more pronounced development of wetter and colder conditions during the aridity now being evident in SW Peloponnese. Similar LIA (Piccarreta et al., 2011; Grauel et al., 2013; Sadori to NE Peloponnese, high humidity is reported from the et al., 2016). In the western-central Europe this period northern Aegean after 2000 cal BP (Psomiadis et al., starts with the alpine glacier advance and increase in 2018) with colder conditions between ca 1500 and ca lake levels at ca 700 cal BP (Frisia et al., 2005; 1150 cal BP (Gogou et al., 2016). A shift from drier to Holzhauser et al., 2005). This is in agreement with wetter conditions at ca 1500 cal BP is inferred from other western Mediterranean records which indicate a Anatolia (Göktürk et al., 2011; Kuzucuoğlu et al., 2011; gradual shift to a wetter period (Martín-Puertas et al., Dean et al., 2015) and the southern Levant (Bar- 2008; Nieto-Moreno et al., 2011; Roberts et al., 2012).

Table 1. A compilation of paleoclimate records from the Peloponnese and the Mediterranean region with full references. Codes correspond to Fig. 22. Information regarding the location and the type of sites, the proxies and the periods covered in each study.

Code Reference Location Site type Proxy Time interval (cal BP) P1 Norström SW Peloponnese, Agios Floros Fen- Sediment core, δD23 Last ∼5000 et al., 2018 Greece Paleolake P1 Katrantsiotis SW Peloponnese, Gialova Lagoon Sediment core, δD31 Last ∼3600 et al., 2018 Greece P2 Katrantsiotis NE Peloponnese, Ancient Lake Sediment core, δD23 Last ∼5000 et al., 2019 Greece Lerna-Wetland A1 Triantaphyllou SE Aegean Sea Marine Sediment core, Last ∼13000 et al., 2009 Greece Biological and geochemical records A2 Psomiadis Thassos Island, Skala Marion Speleothem δ18Ο Last ∼5000 et al., 2018 Greece Cave A3 Gogou North Aegean Sea Athos basin Sediment core, Last ∼1500 et al., 2016 Marine Biological and geochemical records T1 Göktürk NNW Turkey Sofular Cave Speleothem δ13C Last ∼14000 et al., 2011 /South Black Sea T2 Kuzucuoğlu Central Anatolian Tecer Lake Sediment core, Last ∼6000 et al., 2011 plateaux mineral content and grain-size T3 Dean Central Anatolian Nar Gölü- Sediment core, δ18Ο Last ∼4500 et al., 2015 plateaux Brackish maar carbonate lake M1 Cheng Northern Levant Jeita Cave Speleothem δ18Ο, Last ∼20000 et al., 2015 δ13C, Sr/Ca M2 Bar-Matthews and Southern Levant Soreq cave Speleothem δ18Ο, δ13C Last ∼14000 Ayalon, 2004 B1 Zanchetta Western Balkans Lake Shkodra Sediment core, δ18Ο Last ∼4500 et al., 2012 Albania/ carbonate Montenegro C1 Magny Northern Italy Lake Accesa Sedimentological Last ∼11500 et al., 2007 (Tuscany, Italy) techniques C2 Magny Lago Preola (Sicily, Lake Preola Sedimentological Last ∼11000 et al., 2011 southernItaly) techniques C3 Piccarreta Basilicata region, floodplain Sedimentological/ Last ∼8000 et al., 2011 southern Italy, morphological techniques C4 Grauel Southern Italy Gulf Marine Sediment cores, Last ∼2500 et al., 2013 of Taranto δ18O and δ13C of Globigerinoides ruber C5 Sadori Sicily Lago di Pergusa Sediment cores, Last ∼2000 et al., 2016 (Lake) pollen and δ18O and δ13C from carbonates W1 Jalut et al., 2009 Southern-central Synthesis of Sediment cores, pollen Last ∼10000 and Carrión Iberian Peninsula terrestrial records et al., 2001 W2 Martín-Puertas Southern Iberian Zoñar Lake Sediment core, Last ∼4000 et al., 2008 Peninsula biological records W3 Nieto-Moreno Western Marine Sediment core, Last ∼4000 et al., 2011 Mediterranean Sea geochemical records

Fig. 22. Maps illustrating the pattern of climate variability in the Mediterranean region over the last 5000 years based on the comparison of selected climatic proxies with the Peloponnesian δD records. Colored dots indicate climate interpretation in the proxy records. Site numbers and letters correspond to the numbers and letters (code) in Table 1. Spatial variability between the maps reveals a regional drying trend from 4100 to 3900 cal BP and from 1300 to 700 cal BP (MCA). Changes to wetter conditions seem to have occurred at 3000-2600 cal BP and between 700 and 200 cal BP (LIA).

10. Drivers of Holocene climate variability and SW Peloponnese show similar and sometimes opposing climate signals attributed to the effects of The Holocene (ca last 11000 years) is characterized by different atmospheric modes (NAO, NCP). The similar low-frequency millennial-scale climatic oscillations; an signals at ca 4500-3900, ca 3000-2600 and after ca early-middle Holocene warm period, followed by a 1000 cal BP coincide with shifts in the reconstructed cooling trend after ca 5000 cal BP that ended with the NAO (Olsen et al., 2012; Katrantsiotis et al., 2019). present global warming (Renssen et al., 2012; Marcott This indicates that the NAO was exerted the main et al., 2013). This variability has been attributed to control with positive (negative) NAO phases creating orbitally driven changes in the incoming solar radiation drier (wetter) conditions. In addition, cooling events in at the top of the atmosphere (insolation) during boreal the North Atlantic (peaks in IRD curve) coincide with summer (Lorenz et al., 2006; Wanner et al., 2014). trends to arid episodes in the Lerna record before ca An increasing number of proxies around the world 3300 cal BP and thereafter there is a coincidence with indicate that the Holocene low-frequency climatic wetter time-intervals (Bond et al., 2001; Katrantsiotis et oscillations are superimposed by a series of century-to- al., 2019). In the eastern Mediterranean region, the millennial scale changes (Mayewski et al., 2004; Bond events have been associated with enhanced Wanner et al., 2011). In particular, repeated climate aridity. This has been attributed to cold-water input into shifts at periodicities of 2500 and 1000 years have been Mediterranean Sea as a result of the collapse of the associated with changes in the solar influx (Debret et North Atlantic Deep Water circulation leading to a al., 2007). In addition to this variability, there is reduction of evaporation and thereby less precipitation evidence for climate cycles of 1500 years attributed to (Bartov et al., 2003). In the western Mediterranean, a oceanic circulation changes (Bond and Lotti, 1995; noticeable positive correlation between winter rain Mayewski et al., 1997; Debret et al., 2007). These cold minima and Bond events was found during the early events (Bond cycles) were discovered as horizons of ice Holocene and an opposite pattern during the late rafted debris (IRD) in sediment cores from the North Holocene (Zielhofer et al., 2018). Periods of increase in Atlantic (Heinrich, 1988; Bond et al., 1997, 2001). storm activity in the French Mediterranean coast were Subsequent studies have shown the impact of the Bond also correlated with peaks in the IRD record attributed cycles at lower latitudes (e.g. Sierro et al., 2005; Smith to the southward displacement of storm track as a result et al., 2016). However, evidence of these events is not of complex oceanic-atmospheric interactions (stronger always found in paleoclimate records as each region or meridional temperature gradient due to the southward proxy have likely responded differently (Wanner et al., extension of the sea ice) (Raible et al., 2007; Sabatier et 2011; Moossen et al., 2015). The mechanisms causing al., 2012). Overall, these data indicate that there is a these high frequency climate events are not well- link between the climate variability in the North understood. It has been suggested that during the early Atlantic and the Mediterranean climate. However, it Holocene these changes were caused by the slowdown remains unclear which mechanism causes sometimes of the Atlantic overturning circulation due to meltwater dry and sometimes wet conditions in the Peloponnese in fluxes in the North Atlantic (Wanner et al., 2014). Late response to cooling events in the North Atlantic region. Holocene events have also been attributed to volcanic Although NAO is one of the dominant modes of and solar forcing and the complex interactions between variability over the central-eastern Mediterranean, other ocean and atmosphere (Breitenmoser et al., 2012; atmospheric modes have also played an important role. PAGES 2k Consortium, 2013; Wanner et al., 2014). In particular, the opposite signal observed between NE In the Mediterranean region, climatic conditions and SW Peloponnese during some periods can be have varied both spatially and temporally during the explained by the dominance of the NCP over the Holocene due to the complex landscape and the impacts peninsula, which also accounts for the present day of different atmospheric modes (Dormoy et al., 2009; climate variability (Kutiel et al., 2002). At ca 3900- Magny et al., 2012; Roberts et al., 2012). In particular, 3300, ca 3200-3000 and ca 1800–1300 cal BP, positive several processes have been put forward as potential NCP phases would lead to the intrusion of N-NE cold drivers of the Mediterranean climatic variability such as air masses over the Aegean Sea creating wetter and the North Atlantic Bond cycles (Bond et al., 2001; Dean colder conditions in NE Peloponnese. Conversely, at ca et al., 2015) and the large variations in the high-latitude 4600-4500, ca 3200, ca 2600–1800, and 1200-1000 cal atmospheric patterns (NAO, NCP) (Cullen et al., 2002; BP, more negative NCP phases would result in a Kutiel et al., 2002). In addition to these modes, the southwesterly anomalous circulation causing wetter African and Asian monsoons have been considered as conditions in SW Peloponnese and relatively drier in one of the key factors of the Holocene Mediterranean NE Peloponnese as this area is located on the leeward hydroclimate (Jones et al., 2006; Dean et al., 2015). side of the mountain in relation to the westerly winds.

10.1. High latitude atmosphere patterns 10.1.1. Physical Mechanism The Peloponnesian studies offer the opportunity to Overall, the similar and sometimes opposing climate investigate the impact of large-scale teleconnections in signals between NE and SW Peloponnese can mainly the Mediterranean. Variations in δD records from NE be explained by shifts in the dominance between the

NAO and the NCP over the area of the central-eastern the eastern Peloponnese (Fig. 23b and Fig. 25a). Mediterranean. The mechanism, which is responsible However, these conditions are mainly dependent on the for this variation, is not well-known. However, since position and extension of blocking system over Europe. both of the NCP and the NAO occur in the Northern An eastward extension of high pressure system would Hemisphere, it is expected that they would interact with result in drier and warmer conditions even in the areas each other. In particular, it has been shown that the East exposed to northerly winds. On the other hand, a less Atlantic/West Russia (EA/WR) pattern, which is the well-organized EA/WR-like blocking pattern has been more regional expression of NCP, is sensitive to the simulated when the NAO is characterized by more NAO phases (Lim, 2015). In general, EA/WR-like negative phases (Lim, 2015). This results in weaker blocking pattern has been found to be better organized pressure patterns and associated blocking over Europe and strengthened over the Atlantic and Western Europe as the low pressure anomalies are situated further west when the NAO is also characterized by more positive over Greenland due to the weakened upper-level phases (Lim, 2015). During these periods, the eastward westerlies (Wanner et al., 2001). During these periods, propagation of the EA/WR pattern is likely facilitated the axis of the Atlantic Jet is shifted southward, and by the stronger upper level westerly jet (Atlantic Jet) at therefore the south-western Europe receives more 40–60°N, whose exit points towards Scandinavia (Fig. precipitation (Fig. 24a) (Ionita, 2014). In the eastern 23a). The stronger pressure patterns and atmospheric Mediterranean region, low pressure systems originating blocking over Europe favor cold air advection towards from the Atlantic region and the western Mediterranean the Eastern Mediterranean leading to wetter and colder produce higher precipitation amount mainly in the areas conditions in areas exposed to northerly winds, such as exposed to westerly winds (Fig. 24b and Fig. 25b).

a b

Fig. 23 (a) Atmospheric circulation pattern over Europe during positive EA/WR (NCP) and NAO modes (b) Predictive distributions of daily accumulated precipitation in Greece for 8 January 2019. EA/WR (NCP) index was in a positive phase. Map with precipitation data were obtained from the numerical model simieteo.gr

a b

Fig. 24 (a) Atmospheric circulation pattern over Europe during negative EA/WR (NCP) and NAO phases (b) Predictive distributions of daily accumulated precipitation in Greece for 22 January 2019. EA/WR (NCP) index was in a negative mode. Map with precipitation data were obtained from the numerical model simieteo.gr

periods, more frequent cold air advections over the Aegean Sea associated with this blocking pattern would cause wetter and colder conditions in NE Peloponnese (Fig. 23). On the other hand, EA/WR-like blocking patterns should have been less frequent at ca 4600- 4500, ca 3200, ca 2600–1800, and ca 1200-1000 cal BP. This would favor the development of cyclone activity over the western-central Mediterranean creating wetter conations in SW Peloponnese and relatively drier and warmer in NE Peloponnese (Fig. 24). The similar signals during the rest periods are possibly associated with large scale atmospheric changes as discussed above (10.1). Aridity signals from both SW and NE Peloponnese at 4100-3900 (likely 4.2 ka event) and 1000-700 cal BP (MCA), are possibly associated with a more eastward extension of high pressure system over the European mainland (positive NAO phases). The a wetter conditions at ca 4500-4200, 3000-2600 (2.8 ka event), 700-200 cal BP (LIA) are likely related to large- scale cold air outbreaks from high-latitudes combined with increased cyclonic activities during more negative NAO phases (Weninger et al., 2009; Olsen et al., 2012).

10.2. Effects of African and Asian monsoons The Eastern Mediterranean is indirectly exposed to the influence of the African and Asian monsoonal systems. Paleoclimatic records from this region have previously confirmed the impact of the low-latitude atmospheric patterns on the Mediterranean hydroclimate changes during the late Holocene (Jones et al., 2006; Dean et al., 2015). Therefore, in addition to high-latitude atmospheric patterns, shifts in the Peloponnesian δD records are likely associated with changes in the intensity of monsoons and/or shifts in the relative dominance of the African monsoons and the Asian b monsoonal systems over the Eastern Mediterranean. There is evidence for wetter conditions in the Peloponnese between ca 5000-4600 cal BP when the Fig. 25 (a) Predictive distributions of accumulated total precipitation in Greece during a period (26-31/12/2016) with NAO was more positive. Similar climate signals are positive EA/WR (NCP) index. (b) Predictive distributions of also observed in other studies from the southern accumulated precipitation in Greece during a period (21- Mediterranean (e.g. Aegean Sea-Triantaphyllou et al., 30/01/2019) with negative EA/WR (NCP) phase. Data were 2009; southern Italy-Magny et al., 2011). The Mid- obtained from Global Forecast System and meteociel.fr. Holocene represents a period of redistribution of solar energy due to orbital forcing. This resulted in a Throughout the late Holocene, shifts in the NAO progressive decrease in Northern Hemisphere summer and EA/WR modes have affected the precipitation insolation and southward shift of the ITCZ pattern over the Mediterranean in a similar way. It has accompanied by a pronounced weakening of the been shown that there was a trend towards drier monsoonal systems (Wanner et al., 2008). The conditions in the central Mediterranean and Turkey and southward migration of low latitude atmospheric wetter in Israel between 2700 and 1800 cal BP, when patterns would lead to a decrease in the Hadley the North Atlantic was relatively warmer, and the jet circulation and the accompanied North Atlantic stream was located northward (stronger Polar Vortex; anticyclone (Tzedakis et al., 2009). This would unblock NAO+), (Dermody et al., 2012). The initiation of a possible intrusion of western humid airflow towards the period of cooler North Atlantic SSTs (ca 1800 cal BP) Mediterranean resulting in wetter climate conditions. coincided with an opposite pattern of precipitation. After ca 4500-4000 cal BP, the Hadley cell The results of this thesis indicate the dominance of circulation that has displaced southwards and weakened more persistent EA/WR-like blocking patterns (positive cannot be used indiscriminately to explain the summer NCP) over the western-central Europe at ca 3900–3300, climate variability. Instead, during this period, the ca 3200-3000 and ca 1800–1300 cal BP. During these convection has been enhanced over India and has been

35 very limited over North Africa. This implies that the different climatic zones of the peninsula has contributed Asian monsoonal system has played a more important to the development of a dense network of paleoclimatic on the summer conditions of the Eastern Mediterranean records. The integration of data allows us to reconstruct after 4500-4000 cal BP though modulating the strength changes in the E-W climate gradient, to stress local of the large-scale subsidence and the intensity of the climate differences and provide a more accurate Etesians (Ziv et al., 2004; Rizou et al., 2018). The assessment of the regional climatic signal over the last action centers of these two atmospheric patterns are 6000 years. Overall, the results of this thesis provide located over the wider region of Crete and the Aegean evidence for long-term climatic changes in the Sea, respectively (Tyrlis et al., 2013). The interplay Peloponnese, associated with the climate variability in between these two atmospheric patterns, along with the the Eastern Mediterranean. This variability is attributed complex topography in the Peloponnese, leads to a W-E to changes in the large-scale atmospheric circulation a see-saw pattern in summer temperatures over the patterns from the high-latitude and low-latitude regions. peninsula. In this respect, the opposite δD trends from The main conclusions can be summarized as follows: SW and NE Peloponnese might be related to some extent to the influence of the Asian monsoons. In  The diatom and n-alkane δD records from the Agios particular, there is a consistency between the Floros fen provide evidence for abrupt water level Peloponnesian δD signals with monsoonal records from changes at ca 5700 and ca 5300 cal BP. These the Arabian Sea and South-East Asia; intense monsoons changes are largely attributed to two periods of correspond to time-intervals of colder conditions earthquakes, which caused basin-subsidence and (lower δD values) in NE Peloponnese and warmer increased water flow from the nearby karst springs. conditions in SW Peloponnese (Katrantsiotis et al., These two events also seem to coincide with an 2018, 2019). The site of Lerna is exposed to summer intense tectonic activity on a more regional scale. northerly winds (Etesians) leading to a rapid decrease in  Another period of low to intermediate water level water temperatures and lower evaporation (Giadrossich was identified at ca 4600 cal BP. This period is et al., 2015; Katrantsiotis et al., 2019). SW Peloponnese dominated by the new fossil diatom species is located on the lee side of the mountain and thus is Cyclotella paradistinguenda described in this thesis. most affected by adiabatic warming associated with the  This species likely represents the Holocene large scale subsidence compared to NE Peloponnese. morphological evolution of the Cyclotella genus in It can be concluded that the African monsoonal the Agios Floros site. Cyclotella paradistinguenda system had major impact mainly affecting the summer was favored by its capacity to grow in the specific precipitation before 4500 cal BP. The Asian monsoonal environmental settings, which prevailed at that time system had a major influence on summer temperature (alkaline/shallow water/low nutrient availability). changes in the Peloponnese mainly after ca 4000 cal  The Agios Floros fen gradually began to develop BP. These effects of monsoonal systems are likely after ca 4500 cal BP likely due to a combination of superimposed on the precipitation signal associated human activities and also drier conditions. There is with the impacts of high-latitude atmospheric patterns evidence for return to wetter conditions with marked enhancing the δD variability. In addition, other major rainfall seasonality from ca 2800 cal BP. climate events such as the MCA and the LIA might  The n-alkane δD profiles and the XRF data analyzed have larger impact in our records outbalancing the in the Gialova core show similar patterns of influences of the Asian monsoons during these periods. variability. This reveals a common climate signal in SW Peloponnese during the last 3600 years, which also resembles the Agios Floros record. 11. Conclusions  The n-alkane δ13C values from the same core indicate high contribution of aquatic vegetation to This thesis presents multiproxy records of climatic and sedimentary OM with the expansion of submerged environmental changes in the Peloponnese peninsula, plants corresponding to wet and cold periods. SW Greece over the last 6000 years. The analyses were  The δD signals of macrophyte and leaf wax-derived performed on radiocarbon-dated sediment cores from n-alkanes from the Lake Lerna show a nearly the Agios Floros fen and the Gialova Lagoon in SW identical pattern to each other providing further Peloponnese and the Ancient Lake Lerna in NE evidence for precipitation and temperature changes Peloponnese. The combined dataset mainly consists of in NE Peloponnese over the last 5000 years. diatoms, distribution and compound-specific δ13C and  The synthesis of all proxies indicates sometimes δD composition of n-alkanes and XRF elemental similar and sometime opposing climate signal profiles. The high resolution multiproxy record from between NE and SW Peloponnese. This can be the Agios Floros fen enables identification of short-term attributed to the relative dominance of the low- events that can be attributed to local tectonic processes. latitude and high-latitude atmosphere patterns. In addition, the multiproxy approach from all sites has  The NAO likely exerted the main control with contributed to the identification of synchronous changes negative phases creating conditions of increased between the records that can be attributed to a common moisture in the Peloponnese at ca 4500-4100, ca climate mechanism. The location of the sites in two 3000-2600 and after ca 700 cal BP. Drier periods at

ca 4100-3900 and ca 1000-700 cal BP coincides method, dates obtained from pollen can be compared to with more positive NAO phases. Overall, these dry AMS dates derived from seeds of terrestrial plants. and wet periods correspond to large-scale climate Overall, the applicability of radiocarbon dating events associated with atmospheric circulation method in sediments with low organic carbon content changes in the North Atlantic region and beyond. and a lack of macrofossils is limited. Another way to  The opposing signal observed between NE and SW improve the chronology is to retrieve new cores from Peloponnese during some periods can be explained the same spots that can be dated by employing a by the dominance of the NCP over the peninsula. At combination of dating techniques. For instance, both ca 3900-3300, ca 3200-3000 and ca 1800-1300 cal radiocarbon and optically stimulated luminescence BP, positive NCP phases are likely associated with (OSL) dating methods can be applied to test their more persistent blocking systems over the central- suitability for establishing a chronology. The OSL southern Europe that could enhance north- method should be suitable for the studied sites as the easterly advection of cold air towards the Eastern sediment contain sufficient sand-sized quartz or Mediterranean. This would lead to the intrusion of feldspar grains in these depths where the organic the N-NE cold air masses over the Aegean Sea content is low. The establishment of new age-depth creating wetter and colder conditions in areas models from the same sites would further allow for exposed to northerly winds such as NE cross-correlation of ages between the different records. Peloponnese. On the other hand, at ca 4600-4500, ca Future studies should increase the sample 3200, ca 2600–1800, and 1200-1000 cal BP, more resolution, especially from the Agios Floros fen and persistent negative NCP phases are likely related to Gialova records, with emphasis on the period covering weaker pressure patterns and associated blocking the last 3500 years. This could enable a more accurate over the southern-central Europe. During this correlation with other regional records in NE period, the axis of the Atlantic Jet is shifted Peloponnese. Other proxies such as pollen could also be southward, and thus the south-western part of used to facilitate the interpretation, and to investigate Europe receives higher precipitation amount. This the impact of humans in vegetation changes. The southwesterly circulation would cause wetter comparison of pollen with the n-alkane δD data from conditions in SW Peloponnese and relatively drier the same cores would further allow investigating any in NE Peloponnese located on the leeward side of effects of vegetation changes on the isotope data. the mountain range in relation to the westerly winds. A quantitative reconstruction of past climatic  The regional climate signal detected in the present changes based on n-alkanes δD records requires a good study suggests that the biomarker isotope approach understanding of the H isotope fractionation between can provide valuable paleoclimatic information in a source water and biomass (e.g. Garcin et al., 2012). region with a complex landscape, which has evolved Thus, it is important to establish regional calibrations in response to various environmental factors. relating lipid δD values with lake and precipitation δD values. Future studies in the Peloponnese could measure the δD values of n-alkanes in surface sediments, and the δD composition of modern 12. Future perspectives vegetation and in water sources, including lake water, groundwater and precipitation. This approach would This PhD project contributes to a better understanding enhance our understanding of the links between the δD of palaeoclimate and paleoenvironmental changes in the values of sedimentary lipid biomarkers and their Peloponnese during the last ca 6000 years. However, respective hydrogen sources. Τhis is an important step additional studies could improve the understanding and for quantifying past hydroclimatic changes in this area. interpretation of available data from the peninsula. Quantitative reconstruction of lake-level changes Accurate chronological dating is essential for a based on diatoms has so far not been attempted in the more robust inter-regional correlation between different wider region of Greece due to the lack of modern palaeoclimate records. Although, the chronologies from diatom analogues. The reconstruction of maximum lake the Peloponnesian sites should be considered reliable, level changes in the Agios Floros fen was mainly based there is still possibility to improve the age depth models on the habitat preference of Cyclotella distinguenda for some parts of the records (i.e. Agios Floros 700-200 from the Alpine datasets. Therefore, modern diatom cm; Gialova 150-110 cm; Lerna 250-150 cm). Future samples from lakes and wetlands in the Peloponnese proxy studies on the same cores could perform AMS should be analyzed to establish a reliable transfer dating of pollen concentrate samples due to the absence function. Modern datasets in conjunction with lake- of macrofossil plant remains. In this case, it is also morphometric data could allow us to quantify the important to ensure that pollen grains in the samples are Holocene lake-level shifts in Agios Floros accurately. derived from terrestrial plants to avoid biases related to The causes behind the abrupt water level changes in reservoir effects. Although the use of pollen for dating the Agios Floros fen need to be investigated in close is relatively rare, previous studies have shown that detail. A number of factors indicate that these events pollen can provide reliable chronologies (Brown et al., are associated with local tectonic process, which needs 1989; Fletcher et al., 2017). To test the reliability of this to be confirmed. Firstly, successive coring could be

37 performed on a longitudinal section along the fen to 14. Acknowledgements establish the maximum depth and extension of lakes. Secondly, sediment cores can be retrieved from the First and foremost, I would like to thank my supervisor, lower and upper Messinia plain, which should cover the Elin Norström, and my co-supervisor, Karin Holmgren, last ca 6000 years. This is important to examine who introduced me to this project and gave me the whether the water level changes in Agios Floros are opportunity to contribute to this field of natural science related to local hydrological changes or represent in which I have a keen interest. Special thanks to Elin regional events. The latter is the most possible scenario Norström for constant support, inspiration and fruitful as indicated by other records in this plain. Thereafter, discussions. I would like to thank my supervisor, Stefan similar studies covering the period between 6000 and Wastegård, who provided me further support and trust. 5000 cal BP could be performed in the Gialova Lagoon I am grateful to my co-supervisor, Jan Risberg, for and the Lake Lerna to examine the potential relation of valuable support and fruitful discussions throughout the these events with the regional climate variability. course of my studies. I am thankful to my co- Although climate records are increasingly available supervisor, Rienk Smittenberg, for his help with organic from the Peloponnese, there are still large gaps in the geochemistry. I would like to thank Karin Helmens for spatial coverage. This holds especially true for the her support and fruitful discussions during my studies. eastern Peloponnese where the Lake Lerna record is the I am thankful to the 38th Ephorate of Prehistoric only available paleoclimatic study in this area. The and Classical Antiquities, Ministry of Culture and abundant availability of wetland and lakes on the Sports and the 26th Ephorate of Byzantine Antiquities, Peloponnese indicates that there is potential to find Ministry of Culture and Sports for issuing permissions additional places, to increase the spatial coverage and to to perform coring at the Gialova Lagoon and Agios achieve a better understanding of past climate Floros. I would also like to thank the Ephorate of variability. Some suitable sites for future investigations Antiquities of Argolida for kindly giving the permission have already identified within the context of this PhD to carry out coring at the Lake Lerna and the Swedish project. These sites have similar settings as the Gialova Institute at Athens for help with these preparations. Lagoon, Lake Lerna and the Agios Floros fen and are I am grateful to the present, and former, station located in different climatic zones of the peninsula; the managers of NEO station, Giorgos Maneas and Nikos Lake Taka, central Peloponnese, the Evrotas plain, SE Kalivitis, for assisting with the field work planning and Peloponnese and Lake Stymphalia, NE Peloponnese. performance in Agios Floros and Gialova Lagoon. I Paleoclimate proxies and climate modeling can would like to express my gratitude to Nikos Zacharias provide complementary sources of information on past for his hospitality in the laboratory of Archaeometry at climate changes. However, studies of paleoclimate the University of the Peloponnese. Special thanks to modeling have not been performed in the Peloponnese Pavlos Avramidis for his hospitality in the University of despite the availability of proxy records. Therefore, Patras and for our collaboration all these years. there is a great need to integrate the available records I would like to thank everyone who offered their with climate modeling to improve the understanding of time to help with analytical work. I am thankful to climate dynamics and to make better predictions of Malin Kylander, who helped me with the XRF analysis. future climate changes in the Eastern Mediterranean. I would like to acknowledge Klara Hajnal for assistance Finally, the results from this approach can be integrated with extracting n-alkanes from the Lerna samples and with archaeological findings, to improve our knowledge preparing the Lerna samples for bulk isotope analysis. of the impacts of past climate variability on former Special thanks to Anna Hägglund for assistance with human societies in this archaeologically rich region. operating the GC-IRMS, and Heike Siegmund for assistance with bulk isotope analysis. I am thankful to Marianne Ahlbom for helping out with the performance 13. Financial supports of the SEM analysis and also to Mats Regnell for helping out with the identification of macrofossils. This PhD project was financed by the Swedish I am grateful to the Domesticated Landscapes Research Council (grant number 621-2012-4344; P.I. group in Uppsala University, Erika Weiberg, Anton Karin Holmgren) and is a part of the Navarino Bonnier and Martin Finné for organizing all these Environmental Observatory (NEO) collaboration conferences and workshops and providing knowledge between Stockholm University, the Academy of Athens about human-environment interaction in the and TEMES S.A., Greece. Analytical work and travels Peloponnese. Special thanks to Martin Finné for were further financed by: the Bolin Centre for climate inspiring discussion and comments on the manuscripts. research and the project Domesticated Landscapes of I would also like to thank a number of former the Peloponnese (DoLP), Uppsala University with students and colleagues for their assistance in the lab funding from the Swedish Research Council (grant and field work. Thanks to Erika Modig and Helene number 421-2014-1181) and the Royal Swedish Sunmark for helping out with the performance of Academy of Letters, History and Antiquities. Funds for fieldwork in Lake Lerna. Thanks to Lucas Ageby for travels and conferences were further provided by the assisting with the construction of precipitation map of Scholarship Foundation "Penelope Katsea - Spandou". the Peloponnese and the Gialova topographical map.

Thanks to Nichola Strandberg for helping out with the Baas, M., Pancost, R., van Geel, B., Sinninghe-Damsté, J.S., examination of Gialova and Lerna cores for seeds. 2000. A comparative study of lipids in sphagnum species. I would like to express my gratitude to all technical Organic Geochemistry 31, 535–541. and administrative staff at the department of Physical Bar-Matthews, M., Ayalon, A., 2004. Speleothems as paleoclimate indicators, a case study from Soreq Cave Geography, Stockholm University, for their constant located in the Eastern Mediterranean region, Israel. In: support. Special thanks to all my fellow PhD Battarbee, R.W., Gasse, F., Stickley, C.E. (Eds.), Past candidates, former and current. It has been great and Climate Variability through Europe and Africa. very inspiring to share these years with you having Dordrecht: Kluwer Academic Publishers, pp. 363–391. great dinners and fun. Eva Rocha, Charlotta Högberg, Barker, P., 1992. Differential diatom dissolution in Late Simon Larsson, Lindsey Higgins, Anna Plikk, Sandra Quaternary sediments from Lake Manyara, Tanzania: An Sitoe. I am thankful to the Climate and Quaternary unit experimental approach. Journal of Paleolimnology 7, for the nice excursion and moments in the Peloponnese. 235–251. Last but not least, I would like to thank my family, my Bartov, Y., Goldstein, S.L., Stein, M., Enzel, Y., 2003. Catastrophic arid episodes in the Eastern Mediterranean wife and all my friends. Thanks for being supportive linked with the North Atlantic Heinrich events. Geology and for encouraging me during the years of my studies. 31, 439–442. Battarbee, R.W., 1986. Diatom analysis. In: Berglund, B.E. (Eds.), Handbook of Holocene Palaeoecology and 15. References Palaeohydrology. Chichester: John Wiley & Sons, pp. 527–570. Battarbee, R.W., Jones, V.J., Flower, R.J., Cameron, N.G., Aichner, B., 2009. Aquatic macrophyte-derived biomarkers as Bennion, H., Carvalho, L., Juggens, S. 2001. Diatoms. In: palaeolimnological proxies on the Tibetan Plateau. PhD Smol, J.P., Birks, H.J.B., Last, W.M. (Eds.), Tracking thesis. Universität Potsdam, Alfred Wegener Institute for Environmental Change Using Lake Sediments. Volume 3: Polar and Marine Research. Terrestrial, Algal, and Siliceous Indicators. Kluwer Aichner, B., Herzschuh, U., Wilkes, H., 2010. Influence of Academic Publishers, Dordrecht, pp. 155– 202. aquatic macrophytes on the stable carbon isotopic Bauska, T.K., Baggenstos, D., Brook, E.J., Mix, A.C., signatures of sedimentary organic matter in lakes on the Marcott, S.A., Petrenko, V.V., Schaefer, H., Tibetan Plateau. Organic Geochemistry 41, 706–718. Severinghaus, J.P., Lee, J.E., 2016. Carbon isotopes Aichner, B., Feakins, S.J., Lee, J.E., Herzschuh, U., Liu, X., characterize rapid changes in atmospheric carbon dioxide 2015. High resolution leaf wax carbon and hydrogen during the last deglaciation. Proceedings of the National isotopic record of late Holocene paleoclimate in arid Academy of Science 113, 3465–3470. Central Asia. Climate of the Past 11, 619–633. Bender, M.M., 1971. Variations in the 13C/12C ratios of plants Aichner, B., Hilt, S., Périllon, C., Gillefalk, M., Sachse, D., in relation to the pathway of photosynthetic carbon 2017. Biosynthetic hydrogen isotopic fractionation dioxide fixation. Phytochemistry 10, 1239–1244. factors during lipid synthesis in submerged aquatic Beszteri, B., Acs, E., Medlin, L., 2005. Conventional and macrophytes: effect of groundwater discharge and geometric morphometric studies of valve ultrastructure salinity. Organic Geochemistry 113, 10–16. variation in two closely related Cyclotella species Aichner, B., Ott, F., Słowiński, M., Noryśkiewicz, A.M., (Bacillariophyta). European Journal of Phycology 40, 89– Brauer, A., Sachse, D., 2018. Leaf wax n-alkane 103. distributions record ecological changes during the Birks, H.H., Birks, H.J.B., 2006. Multi-proxy studies in Younger Dryas at Trzechowskie paleolake (northern palaeolimnology. Vegetation History and Archaeobotany Poland) without temporal delay. Climate of the Past 14, 15, 235–251. 1607–1624. Björck, S., Wohlfarth, B., 2001. 14C chronostratigraphic Alcock, S.E., Cherry, J.F., 2004. Side-by-Side Survey. techniques in paleolimnology. In: Last, W.M., Smol, J.P. Comparative Regional Studies in the Mediterranean (Eds.), Tracking Environmental Change Using Lake World. Oxford: Oxbow Books. Sediments: Basin analysis, coring, and chronological Alpert, P., Neeman, B.U., Shay-El, Y., 1990. Intermonthly techniques. Volume 1. Dordrecht: Kluwer Academic variability of cyclone tracks in the Mediterranean. Journal Publishers, pp. 205–245. of Climate 3, 1474–1478. Blaauw, M., 2010. Methods and code for classical age- Argiriou, A.A., Lykoudis, S., 2006: Isotopic composition of modelling of radiocarbon sequences. Quaternary precipitation in Greece. Journal of Hydrology 327, 486– Geochronology 5, 512–518. 495. Bolin Centre Database, Mouzaki, 2017. Athanassas, C., Fountoulis, I., 2013. Quaternary neotectonic http://bolin.su.se/data/Norstrom-2017-1 configuration of the southwestern Peloponnese, Greece, Bond, G., Lotti, R., 1995. Iceberg discharges into the north based on luminescence ages of marine terraces. Journal of Atlantic on millennial time scales during the last Earth Science 24, 410–427. glaciation. Science 267, 1005–1010. Avramidis, P., Iliopoulos, G., Kontopoulos, N., Panagiotaras, Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., D., Barouchas, P., Nikolaou, K., Papadopoulou, P., 2015. deMenocal, P., Priore, P., Cullen, H., Hajdas, I., Bonani, Depositional environments, sediment characteristics, G., 1997. A pervasive millennial-scale cycle in North palaeoecological analysis and environmental assessment Atlantic Holocene and glacial climates. Science 278, of an internationally protected shallow Mediterranean 1257–1266. lagoon, Gialova Lagoon – Navarino Bay, Greece. Earth Bond, G.G., Kromer, B., Beer, J., Muscheler, R., Evans, M., and Environmental Science Transactions of the Royal Showers, W., Hoffmann, S., Lotti-Bond, R., Hajdas, I., Society of Edinburgh 105, 189–206. Bonani, G., 2001. Persistent solar influence on North

Atlantic climate during the Holocene. Science 294, 2130– Chikaraishi, Y., Naraoka, H., 2003. Compound-specific δD– 2136. δ13C analyses of n-alkanes extracted from terrestrial and Bony, S., Risi, C., Vimeux, F., 2008. Influence of convective aquatic plants. Phytochemistry 63, 361–371. processes on the isotopic composition (δ18O and δD) of Chikaraishi, Y., 2014. 13C/12C signatures in plants and algae. precipitation and water vapor in the tropics: 1. In: Holland, H., Turekian, K. (Eds.), Treatise on Radiative‐convective equilibrium and Tropical Ocean– geochemistry. Second Edition. Volume 12. Elsevier, Global Atmosphere–Coupled Ocean‐Atmosphere Amsterdam, Netherlands, pp. 95–123. Response Experiment (TOGA‐COARE) simulations. Cheng, H., Sinha, A., Verheyden, S., Nader, F.H., Li, X.L., Journal of Geophysical Research 113, D19305. Zhang, P.Z., Yin, J.J., Yi, L., Peng, Y.B., Rao, Z.G., Ning, Borgmark, A., Wastegård, S., 2008. Regional and local Y.F., Edwards, R.L., 2015. The climate variability in patterns of peat humification in three raised peat bogs in northern Levant over the past 20,000 years. Geophysical Värmland, south-central Sweden. GFF 130, 161–176. Research Letters 42, 8641–8650. Bouzos, D., Κontopoulos, Ν., Avramidis, P., 2002. Cline, E.H., 2010. The Oxford Handbook of the Bronze Age Oceanographic observations in the Lagoon of Gialova, Aegean (ca. 3000–1000 BC). Oxford University Press, (SW Peloponnese). In: Proceedings of 6th Pan-Hellenic Oxford. Geographical Conference of the Hellenic Geographical Collister, J.W., Rieley, G., Stern, B., Eglinton, G., Fry, B., Society, pp. 247–254. Thessaloniki, Greece. 1994. Compound-specific δ13C analyses of leaf lipids Bowen, G.J., 2008. Spatial analysis of the intra-annual from plants with differing carbon dioxide metabolisms. variation of precipitation isotope ratios and its Organic Geochemistry 21, 619–627. climatological corollaries. Journal of Geophysical Conroy, J.L., Cobb, K.M., Noone, D., 2013. Comparison of Research: Atmospheres 113, D05113. precipitation isotope variability across the tropical Pacific Boyd, M., 2015. Speleothems from warm climates: Holocene in observations and SWING2 model simulations. Journal records from the Caribbean and Mediterranean regions. of Geophysical Research: Atmospheres 118, 5867–5892. PhD dissertation. Department of Physical Geography, Conte, M., Giuffrida, S., Tedesco, S., 1989. The Stockholm University, Sweden. Mediterranean Oscillation: Impact on Precipitation and Bradley, R.S., 2014. Paleoclimatology: Reconstructing Hydrology in Italy. Proceedings of the Conference on Climate of the Quaternary. Third Edition. Academic Climate and Water 1. Academy of Finland, pp. 121–137. Press, Elsevier, Oxford. Cox, E.J., 2014. Diatom identification in the face of changing Brandimarte, L., Di Baldassarre, G., Bruni, G., D'Odorico, species concepts and evidence of phenotypic plasticity. P., Montanari, A., 2011. Relation between the north- Journal of Micropaleontology 33, 111–120. Atlantic oscillation and hydroclimatic conditions in Craig, H., 1961. Isotopic variations in meteoric waters. Mediterranean areas. Water Resources Management 25, Science 133, 1702–1703. 1269–1279. Cranwell, P.A., 1973. Chain-length distribution of n-alkanes Breitenmoser, P., Beer, J., Brönnimann, S., Frank, D., from lake sediments in relation to post-glacial Steinhilber, F., Wanner, H., 2012. Solar and volcanic environmental change. Freshwater Biology 3, 259–265. fingerprints in tree-ring chronologies over the past 2000 Cranwell, P.A., Eglinton, G., Robinson, N., 1987. Lipids of years. Palaeogeography, Palaeoclimatology, aquatic organisms as potential contributors to lacustrine Palaeoecology 313– 314, 127–139. sediments II. Organic Geochemistry 11, 513–527. Brocks, J.J., Summons, R.E., 2014. Sedimentary Cross, R.A., 1980. Distribution of sub-families of gramineae hydrocarbons, biomarkers for early life. In: Holland, H., in the old world. Kew Bulletin 35, 279–289. Turekian, K. (Eds.), Treatise on geochemistry. Second Croudace, I.W., Rindby, A., Rothwell, R.G., 2006. ITRAX: Edition. Volume 10. Elsevier, Amsterdam, Netherlands, description and evaluation of a new multi-function X-ray pp. 61–103. core scanner. In: Rothwell, R.G. (Ed.), New Ways of Broecker, W.S., Kulp, J.L., 1956. The radiocarbon method of Looking at Sediment Cores and Core Data. Geological age determination. American Antiquity 22, 1–11. Society of London, London, UK, pp. 51–63. Brown, T.A., Nelson, D.E., Mathewes, R.W., Vogel, J.S., Croudace, I.W., Rothwell, R.G., 2010. Micro-XRF sediment Southon, J.R., 1989. Radiocarbon dating of pollen by core scanners: important new tools for the environmental accelerator mass spectrometry. Quaternary Research 32, and earth sciences. Spectroscopy Europe, 22, 6–13. 205–212. Cullen, H.M., deMenocal, P.B., Hemming, S., Hemming, G., Bush, R.T., McInerney, F.A., 2013. Leaf wax n-alkane Brown, F.H., Guilderson, T., Sirocko, F., 2000. Climate distributions in and across modern plants: implications for change and the collapse of the Akkadian empire: paleoecology and chemotaxonomy. Geochimica et Evidence from the deep sea. Geology 28, 379–382. Cosmochimica Acta 117, 161–179. Cullen, H., Kaplan, A., Arkin, P.A., Demenocal, P.B., 2002. Carrión, J.S., Andrade, A., Bennett, K.D., Navarro, C., Impact of the North Atlantic Oscillation on Middle Munuera, M., 2001. Crossing forest thresholds: inertia Eastern climate and streamflow. Climatic Change 55, and collapse in a Holocene sequence from south-central 315–338. Spain. The Holocene 11, 635–653. Cundy, A.B., Sprague, D., Hopkinson, L., Maroukian, H., Chalié, F., Gasse, F., 2002. Late Glacial-Holocene diatom Gaki-Papanastassiou, K., Papanastassiou, D., Frogley, record of water chemistry and lake level change from the M.R., 2006. Geochemical and stratigraphic indicators of tropical East African Rift Lake Abiyata (Ethiopia). late Holocene coastal evolution in the Gythio area, Palaeogeography, Palaeoclimatology, Palaeoecology 187, southern Peloponnese Greece. Marine Geology 230, 161– 259–283. 177. Chikaraishi, Y., Naraoka, H., 2001. Organic hydrogen-carbon Currás, A., Zamora, L., Reed, J.M., García-Soto, E., Ferrero, isotope signatures of terrestrial higher plants during S., Armengol, X., Mezquita-Joanes, F., Marqués, M.A., biosynthesis for distinctive photosynthetic pathways. Riera, S., Julià, R., 2012. Climate change and human Geochemical Journal 35, 451–458. impact in central Spain during Roman times: high-

resolution multi-proxy analysis of a tufa lake record Duan, Y., Wu, B., Xu, L., He, J., Sun, T., 2011. (Somolinos, 1280 m a.s.l). Catena 89, 31–53. Characterisation of n-alkanes and their hydrogen isotopic Cvetkoska, A., Levkov, Z., Reed, J.M., Wagner, B., 2014. composition in sediments from Lake Qinghai, China. Late glacial to Holocene climate change and human Organic Geochemistry 42, 720–726. impact in the Mediterranean: the last ca. 17 ka diatom Eglinton, G., Hamilton, R.J., 1967. Leaf epicuticular waxes. record of Lake Prespa (Macedonia/Albania/Greece). Science 156, 1322–1335. Palaeogeography, Palaeoclimatology, Palaeoecology 406, Eglinton, T.I., Eglinton, G., 2008. Molecular proxies for 22–32. paleoclimatology. Earth and Planetary Science Letters Dansgaard, W., 1964. Stable isotopes in precipitation. Tellus 275, 1–16. 16, 436–68. Emmanouilidis, A., Katrantsiotis, C., Norström, E., Risberg, Darling, W.G., 2011. The isotope hydrology of quaternary J., Kylander, M., Sheik, A.T., Iliopoulos, G., Avramidis, climate change. Journal of Human Evolution 60, 417–427. P., 2018. Middle to late Holocene palaeoenvironmental Dayan, U., Ricaud, P., Zbinden, R, Dulac, F., 2017. study of Gialova lagoon, SW Peloponnese, Greece. Atmospheric pollution over the eastern Mediterranean Quaternary International 476, 46–62. during summer – a review. Atmospheric Chemistry and Engel, M., Knipping, M., Bruckner, H., Kiderlen, M. Kraft, Physics 17, 13233–13263. J.C., 2009. Reconstructing middle to late Holocene Dean, J.R., Jones, M.D., Leng, M.J., Noble, S.R., Metcalfe, palaeogeographies of the lower Messenian Plain S.E., Sloane, H.J., Sahy, D., Eastwood, W.J., Roberts, (southwestern Peloponnese, Greece); Coastline migration, C.N., 2015. Eastern Mediterranean hydroclimate over the vegetation history and sea level change. Palaeogeography, late glacial and Holocene, reconstructed from the Palaeoclimatology, Palaeoecology 284, 257–270. sediments of Nar lake, central Turkey, using stable Engels, S., van Geel, B., 2012. The effects of changing solar isotopes and carbonate mineralogy. Quaternary Science activity on climate: contributions from palaeo- Reviews 124, 162–174. climatological studies. Journal of Space Weather and Debret, M., Bout-Roumazeilles, V., Grousset, F., Desmet, M., Space Climate 2, A09. McManus, J.F., Massei, N., Sebag, D., Petit, J.R., Copard, Esposito, E., Pece, R., Porfido, S., Tranfaglia, G., 2001. Y., Trentesaux, A., 2007. The origin of the 1500-year Hydrological anomalies precursory of earthquakes in climate cycles in Holocene North-Atlantic records. Southern Apennines (Italy). Natural Hazards and Earth Climate of the Past 3, 569–575. System Sciences 1, 137–144. Delhon, C., Alexandre, A., Berger, J.F., Thiébault, S., Fairbanks, R.G., Mortlock, R.A., Chiu, T.-C., Cao, L., Brochier, J.L., Meunier, J.D., 2003. Phytolith assemblages Kaplan, A., Guilderson, T.P., Fairbanks, T.W., Bloom, as a promising tool for reconstructing Mediterranean A.L., Grootes, P.M., Nadeau, M.-J., 2005. Radiocarbon Holocene vegetation. Quaternary Research 59, 48–60. calibration curve spanning 0 to 50,000 years BP based on Dermody, B.J., de Boer, H.J., Bierkens, M.F.P., Weber, S.L., paired 230Th/234U/238U and 14C dates on pristine Wassen, M.J., Dekker, S.C., 2012. A seesaw in corals. Quaternary Science Review 24, 1781–1796. Mediterranean precipitation during the Roman Period Fang, J., Wu, F., Xiong, Y., Li, F., Du, X., An, D., Wang, L., linked to millennial-scale changes in the North Atlantic. 2014. Source characterization of sedimentary organic Climate of the Past 8, 637–651. matter using molecular and stable carbon isotopic Diefendorf, A.F., Freimuth, E.J., 2017. Extracting the most composition of n-alkanes and fatty acids in sediment core from terrestrial plant-derived n-alkyl lipids and their from Lake Dianchi, China. Science of the Total carbon isotopes from the sedimentary record: A review. Environment 473–474, 410–421. Organic Geochemistry 103, 1–21. Farquhar, G.D., Ehleringer, J.R., Hubick, K.T., 1989. Carbon Doi, H., Kikuchi, E., Shikano, S., Tagaki, S., 2010. isotope discrimination and photosynthesis. Annual Differences in nitrogen and carbon stable isotopes Review of Plant Physiology and Plant Molecular Biology, between planktonic and benthic microalgae. Limnology 40, 503–537. 11, 185–192. Feidas, H., Noulopoulou, Ch., Makrogiannis, T., Bora-Senta, Dore, M.H.I, 2005. Climate change and changes in global E., 2007. Trend analysis of precipitation time series in precipitation patterns: What do we know? Environment Greece and their relationship with circulation using International 31, 1167– 1181. surface and satellite data: 1955–2001. Theoretical and Dormoy, I., Peyron, O., Combourieu-Nebout, N., Goring, S., Applied Climatology 87, 155–177. Kotthoff, U., Magny, M., Pross, J., 2009. Terrestrial Feidas, H., 2017. Trend analysis of air temperature time series climate variability and seasonality changes in the in Greece and their relationship with circulation using Mediterranean region between 15000 and 4000 years BP surface and satellite data: Recent trends and an update to deduced from marine pollen records. Climate of the Past 2013. Theoretical and Applied Climatology 129, 1383– 5, 615–632. 1406. Dotsika, E., Lykoudis, S., Poutoukis, D., 2010. Spatial Feng H., Zhang, M., 2015. Global land moisture trends: drier distribution of the isotopic composition of precipitation in dry and wetter in wet over land. Scientific Reports 5, and spring water in Greece. Global and Planetary Change 18018. 71, 141–149. Fick, S.E, Hijmans, R.J., 2017. Worldclim 2: New 1-km Douglas, P.M.J., Pagani, M., Brenner, M., Hodell, D.A., spatial resolution climate surfaces for global land areas. Curtis, J.H., 2012. Aridity and vegetation composition are International Journal of Climatology 37, 4302–4315. important determinants of leaf-wax dD values in Ficken, K.J., Li, B., Swain, D.L., Eglinton, G., 2000. An n- southeastern Mexico and Central America. Geochimica et alkane proxy for the sedimentary input of submerged Cosmochimica Acta 97, 24–45. /floating freshwater aquatic macrophytes. Organic Duan, Y., He, J., 2011. Distribution and isotopic composition geochemistry 31, 745–749. of n-alkanes from grass, reed and tree leaves along a Finke, E.A.W., 1988. Landscape evolution of the Argive latitudinal gradient in China. Geochemical Journal 45, plain, Greece: Paleoecology, Holocene depositional 199–207.

history, and coastline changes. PhD Thesis. Stanford Garcin, Y., Schwab, V.F., Gleixner, G., Kahmen, A., Todou, University, University Microfilms, Ann Arbor, Michigan. G., Séné, O., Onana, J.-M., Achoundong, G., Sachse, D., Finné, M., Norström, E., Risberg, J., Scott, L., 2010. Siliceous 2012. Hydrogen isotope ratios of lacustrine sedimentary microfossils as late-Quaternary paleo-environmental n-alkanes as proxies of tropical African hydrology: indicators at Braamhoek wetland, South Africa. The insights from a calibration transect across Cameroon. Holocene 20, 747–760. Geochimica et Cosmochimica Acta 79, 106–126. Finné, M., Holmgren, K., Sundqvist, H.S., Weiberg, E., Gasse, F., 1987. Diatoms for reconstructing palaeo- Lindblom, M., 2011. Climate in the eastern environments and paleohydrology in tropical semi-arid Mediterranean, and adjacent regions, during the past 6000 zones - Example of some lakes from Niger since 12 000 years–A review. Journal of Archaeological Science 38, BP. Hydrobiologia 154, 127– 163. 3153–3173. Gasse, F., 1994. Lacustrine diatoms for reconstructing past Finné, M., Bar-Matthews, M., Holmgren, K., Sundqvist, H.S., hydrology and climate. In: Duplessy, J.C., Spyridakis, Liakopoulos, I., Zhang, Q., 2014. Speleothem evidence M.T. (Eds.), Long-Term climatic variations: Data and for late Holocene climate variability and floods in modelling. NATO ASI Series. Volume 122. Berlin: Southern Greece. Quaternary Research 81, 213–227. Springer-Verlag (NATO ASI series 22), pp. 335–369. Finné, M., Holmgren, K., Shen, C-C., Hu, H-M., Boyd, M., Gat, J.R., 1996. Oxygen and hydrogen isotopes in the Stocker, S., 2017. Late Bronze Age climate change and hydrologic cycle. Annual Review of Earth and Planetary the destruction of the Mycenaean Palace of Nestor at Sciences 24, 225–262. Pylos. Plos One 12, e0189447. Giadrossich, F., Niedda, M., Cohen, D., Pirastru, M., 2015. Fletcher, W.J., Zielhofer, C., Mischke, S., Bryant, C., Xu, X., Evaporation in a Mediterranean environment by energy Fink, D., 2017. AMS radiocarbon dating of pollen budget and Penman methods, Lake Baratz, Sardinia, Italy. concentrates in a karstic lake system. Quaternary Hydrology and Earth System Sciences 19, 2451–2468. Geochronology 39, 112–123. Giger, W., Schaffner, C., Wakeham, S.G., 1980. Aliphatic and Flocas, A.A., Giles, B.D., 1991. Distribution and intensity of olefinic hydrocarbons in recent sediments of Greifensee, frontal rainfall over Greece. International Journal of Switzerland. Geochimica et Cosmochimica Acta 44, 119– Climatology 11, 429–442. 129. Flocas, H.A., Simmonds, I., Kouroutzoglou, J., Keay, K., Giorgi, F., Lionello, P., 2008. Climate change projections for Hatzaki, M., Bricolas, V., Asimakopoulos, D., 2010. On the Mediterranean region. Global Planet Change 63, 90– cyclonic tracks over the Eastern Mediterranean. Journal of 104. Climate 23, 5243–5257. Gogou, A., Triantaphyllou, M., Xoplaki, E., Izdebski, A., Flower, R.J., 1993. Diatom preservation: Experiments and Parinos, C., Dimiza, M., Bouloubassi, I., Luterbacher, J., observations on dissolution and breakage in modern and Kouli K., Martrat, B., Toreti, A., Fleitmann, D., Rousakis, fossil material. Hydrobiologia 269–270, 473–484. G., Kaberi, H., Athanasiou, M., Lykousis, V., 2016. Flower, R.J., Ryves, D.B., 2009. Diatom preservation: Climate variability and socio-environmental changes in Differential preservation of sedimentary diatoms in two the northern Aegean (NE Mediterranean) during the last saline lakes. Acta Botanica Croatica 68, 381–399. 1500 years. Quaternary Science Reviews 136, 209–228. France, R.L., 1995. Carbon-13 enrichment in benthic Gonfiantini, R., 1986. Environmental isotopes in lake studies. compared to planktonic algae: Foodweb implications. In: Fritz, P., Fontes, J.Ch. (Eds.), Handbook of Marine Ecology Progress Series 124, 307–312. Environmental Isotope Geochemistry. The terrestrial Francus, P., Lamb, H., Nakagawa, T., Marshall, M., Brown, environment B. Volume 2. Elsevier, Amsterdam, E., Suigetsu 2006 project members, 2009. The potential of Netherlands, pp. 113–168. high-resolution X-ray fluorescence core scanning: Gourcy, L.L., Groening, M., Aggarwal, P.K., 2005. Stable applications in paleo-limnology. PAGES News 17, 93–95. oxygen and hydrogen isotopes in precipitation. In: Frigola, J., Moreno, A., Cacho, I., Sierro, F.J., Flores, J.A., Aggarwal, P.K., Gat, J.R., Froehlich, K. (Eds.), Isotopes Grimalt, J.O., Hodell, D.A., Curtis, J.H., 2007. Holocene in the water cycle. Springer, Dordrecht, pp. 39–51. climate variability in the western Mediterranean region Göktürk, O.M., Fleitmann, D., Badertscher, S., Cheng, H., from a deepwater sediment record. Paleoceanography 22, Edwards, R.L., Leuenberger, M., Fankhauser, A., Tüysüz, PA2209. O., Kramers, J., 2011. Climate on the southern Black Sea Frisia, S., Borsato, A., Spötl, C., Villa, I.M, Cucchi, F., 2005. coast during the Holocene: implications from the Sofular Climate variability in the SE Alps of Italy over the past 17 Cave record. Quaternary Science Reviews 30, 2433–2445. 000 years reconstructed from a stalagmite record. Boreas Grauel, A-L., Goudeau, M-L.S., de Lange, G.J., Bernasconi, 34, 445–455. S.M., 2013. Climate of the past 2500 years in the Gulf of Fountoulis, I., Mariolakos, I., 2008. Neotectonic folds in the Taranto, Central Mediterranean Sea: A high-resolution central-western Peloponnese, Greece. Zeitschrift der climate reconstruction based on δ18O and δ13C of Deutschen Gesellschaft für Geowissenschaften - German Globigerinoides ruber (white). The Holocene 23, 1440– Journal of Geology 159, 485–494. 1446. Fountoulis, I., Mavroulis, S.D., 2013. Application of the Grice, K., Eiserbeck, C., 2014. The analysis and application Environmental Seismic Intensity scale (ESI 2007) and the of biomarkers. In: Holland, H.D., Turekian, K.K. (Eds.), European Macroseismic Scale (EMS-98) to the Kalamata Treatise on geochemistry. Second Edition. Volume 12. (SW Peloponnese, Greece) earthquake (Ms = 6.2, Elsevier, Amsterdam, Netherlands, pp. 47–78. September 13, 1986) and correlation with neotectonic Grimm, E.C., 1987. CONISS: A Fortran 77 program for structures and active faults. Annals of Geophysics 56, stratigraphically constrained cluster analysis by the S0675. method of incremental sum of sequences. Computer and Fountoulis, I., Mariolakos, I., Ladas, I., 2014. Quaternary Geosciences 13, 13–35. basin sedimentation and geodynamics in SW Peloponnese Grimm, E.C., 1991. TILIA and TILIA-Graph 1.7.16. (Greece) and late stage uplift of Taygetos Mt. Bollettino Springfield, IL: Illinois State Museum, Research and di Geofisica Teorica ed Applicata 55, 303–324. Collection Center.

Grimm, E.C., 2004. TG View. Springfield, IL: Illinois State IPCC, 2014. Climate Change 2014: Synthesis Report. In: Museum, Research and Collection Center. Core Writing Team, Pachauri, R.K., Meyer, L.A. (Eds.), Guenther, F., Aichner, B., Siegwolf, R., Xu, B., Yao, T., Contribution of Working Groups I, II and III to the Fifth Gleixner, G., 2013. A synthesis of hydrogen isotope Assessment Report of the Intergovernmental Panel on variability and its hydrological significance at the Climate Change. IPCC, Geneva, Switzerland. Qinghai-Tibetan Plateau. Quaternary International, 313– Jahns, S., 1993. On the Holocene vegetation history of the 314, 3–16. Argive plain (Peloponnese, southern Greece). Vegetation Guiot, J., Cramer, W., 2016. Climate Change The 2015 Paris History and Archaeobotany 2, 187–203. Agreement Thresholds and Mediterranean Basin Jalut, G., Dedoubat, J.J., Fontugne, M., Otto, T., 2000. Ecosystems. Science 354, 465–468. Holocene circum-Mediterranean vegetation changes: Hassan, G.S., 2013. Diatom-based reconstruction of Middle to Climate forcing and human impact. Quaternary Late Holocene paleoenvironments in Lake Lonkoy, International 200, 4–18. southern Pampas, Argentina. Diatom Research 28, 473– Jansen, E., Overpeck, J., Briffa, K.R., Duplessy, J.-C., Joos, 486. F., Masson-Delmotte, V., Olago, D., Otto-Bliesner, B., Hayes, J.M., 1993. Factors controlling 13C contents of Peltier, W.R., Rahmstorf, S., Ramesh, R., Raynaud, D., sedimentary organic compounds: Principles and evidence. Rind, D., Solomina, O., Villalba, R., Zhang, D., 2007. Marine Geology 113, 111–125. Palaeoclimate. In: Solomon, S., Qin, D., Manning, M., Håkansson, H., Chepurnov, V., 1999. A study of variation in Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, valve morphology of the diatom Cyclotella meneghiniana H.L. (Eds.), Climate Change 2007: The physical science in monoclonal cultures: effect of auxospore formation and basis. Contribution of Working Group I to the Fourth different salinity conditions. Diatom Research 14, 251– Assessment Report of the Intergovernmental Panel on 272. Climate Change. Cambridge: Cambridge University Håkansson, H., Korhola, A., 1998. Phenotypic plasticity in the Press, pp. 433–497. diatom Cyclotella meneghiniana or a new species? Nova Jasper, J.P., Hayes, J.M., 1990. A carbon isotope record of Hedwigia 66, 187–196. CO2 levels during the late Quaternary. Nature 347, 462– He, D., Ladd, S.N., Sachs, J.P., Jaffé, R., 2017. Inverse 464. relationship between salinity and 2H/1H fractionation in Jewson, D.H., 1992. Size reduction, reproductive strategy and leaf wax n-alkanes from Florida mangroves. Organic the life cycle of a centric diatom. Philosophical Geochemistry 110, 1–12. Transactions of the Royal Society of London, B 336, 191– Heinrich, H., 1988. Origin and consequences of cyclic ice 213. rafting in the northeast Atlantic Ocean during the past Ji, F., Wu, Z., Huang, J., Chassignet, E., 2014. Evolution of 130,000 years. Quaternary Research 29, 143–152. land surface air temperature trend. Nature Climate Change Hellenic National Meteorological Service, 2018. 4, 462–466. http://www.hnms.gr/emy/el/climatology/climatology. Jones, M.D., Roberts, C.N., Leng, M.J., Türkeş, M., 2006. A Heymann, C., Nelle, O., Dörfler, W., Zagana, H., Nowaczyk, high-resolution late Holocene lake isotope record from N., Xue, J., Unkel, I., 2013. Late Glacial to mid-Holocene Turkey and links to North Atlantic and monsoon climate. palaeoclimate development of Southern Greece inferred Geology 34, 361–364. from the sediment sequence of Lake Stymphalia (NE- Jones, T.D, Lawson I.T., Reed J.M., Wilson G.P., Leng M. J., Peloponnese). Quaternary International 302, 42–60. Gierga, M., Bernasconi S.M., Smittenberg R.H., Hajdas, Holdsworth, G., Fogarasi, S., Krouse, H.R., 1991. Variation I., Bryant C.L. Tzedakis P.C., 2013. Diatom–inferred late of the stable isotopes of water with altitude in the Saint Pleistocene and Holocene palaeolimnological changes in Elias Mountains of Canada. Journal of Geophysical the Ioannina basin, northwest Greece. Journal of Research: Atmospheres 96, 7483–7494. Paleolimnology 49, 185–204. Holmgren, K., Risberg, J., Freudendahl, J., Achimo, M., Kahmen, A., Schefuss, E., Sachse, D., 2013. Leaf water Ekblom, A., Mugabe, J., Norström, E., Sitoe, S., 2012. deuterium enrichment shapes leaf wax n-alkane delta D Water level variations at Lake Nhauhache, Mozambique, values of angiosperm plants I: experimental evidence and during the last 2300 years. Journal of Paleolimnology 48, mechanistic insights. Geochimica et Cosmochimica Acta 311–322. 111, 39–49. Hou, J.Ζ., D'Andrea, W.J., MacDonald, D., Huang, Y., 2007. Kaniewski, D., Van Campo, E., Guiot, J., Le Burel, S., Hydrogen isotopic variability in leaf waxes among Thierry, O., Baeteman, C., 2013. Environmental roots of terrestrial and aquatic plants around Blood Pond, the Late Bronze Age crisis. Plos One 8, e71004. Massachusetts, USA. Organic Geochemistry 38, 977–984. Kassaras, I., Kazantzidou-Firtinidou, D., Ganas, A., Huang, Y., Shuman, B., Wang, Y., Webb, T., 2004. Hydrogen Kapetanidis, V., Tsimi, C., Valkaniotis, S., Sakellariou, isotope ratios of individual lipids in lake sediments as N., Mourloukos, S., 2018. Seismic risk and loss novel tracers of climatic and environmental change: a assessment for Kalamata (SW Peloponnese, Greece) from surface sediment test. Journal of Paleolimnology 31, 363– neighbouring shallow sources. Bolletino di Geofisica 375. Teorica e Applicata 59, 1–26. IAEA, International Atomic Energy Agency, 2014. Global Katrantsiotis, C., Norström, E., Holmgren, K., Risberg, J., Network of Isotopes in Precipitation Database. Skelton, A., 2015. High-resolution environmental http://nucleus.iaea.org reconstruction in SW Peloponnese, Greece, covering the IGME, Institute of Geology and Mineral Exploration, 1980. last c. 6000 years: Evidence from Agios Floros fen, Geological Map of Greece, 1:50000. Koroni-Pylos-Skhiza Messenian plain. The Holocene 26, 188–204. Sheet. Athens. Katrantsiotis, C., Risberg, J., Norström, E., Holmgren, K., Ionita, M., 2014. The Impact of the East Atlantic/Western 2016. Morphological study of Cyclotella distinguenda and Russia Pattern on the hydroclimatology of Europe from a description of a new fossil species Cyclotella mid-winter to late spring. Climate 2, 296–309. paradistinguenda sp. nov. from the Agios Floros fen, SW

Peloponnese, Greece in relation to other Cyclotella Krammer, K., Lange-Bertalot, H., 1991a. Bacillariophyceae 3. species. Diatom Research 31, 243–267. Teil: Centrales, Fragilariaceae, Eunotiaceae. In: Ettl, H., Katrantsiotis, C, Kylander, M., Smittenberg, R., Yamoah, Gerloff, J., Heynig, H., Mollenhauer, D. (Eds.), K.K.A., Hättestrand, M., Avramidis, P., Strandberg, N.A., Süßwasserflora von Mitteleuropa 2/3. Stuttgart, Jena: Norström, E., 2018. Eastern Mediterranean hydroclimate Gustav Fischer Verlag. reconstruction over the last 3600 years based on Krammer, K., Lange-Bertalot, H., 1991b. Bacillariophyceae sedimentary n-alkanes, their carbon and hydrogen isotope 4. Teil: Achnanthaceae. KritischeErgänzungenzu composition and XRF data from the Gialova Lagoon, SW Navicula (Lineolatae) und Gomphonema. In: Ettl, H., Greece. Quaternary Science Reviews 194, 77–93. Gärtner, G., Gerloff, J., Heynig, H., Mollenhauer, D. Katrantsiotis, C., Norström, E., Smittenberg, R., Finné, M., (Eds.), Süßwasserflora von Mitteleuropa 2/4. Stuttgart, Weiberg, W., Hättestrand, M., Avramidis, P., Wastegård, Jena: Gustav Fischer Verlag. S., 2019. Climate changes in the Eastern Mediterranean Kutiel, H., Benaroch, Y., 2002. North Sea-Caspian pattern over the last 5000 years and their links to the high-latitude (NCP)-an upper level atmospheric teleconnection atmospheric patterns and Asian monsoons. Global and affecting the Eastern Mediterranean: identification and Planetary Change 175, 36–51. definition. Theoretical and Applied Climatology 71, 17– Kendall, C., Caldwell, E.A., 1998. Fundamentals of isotope 28. geochemistry. In: Kendall, C., McDonnell, J.J. (Eds.), Kutiel, H., Maheras, P., Türkeş, M., Paz, S., 2002. North Sea– Isotope tracers in catchment hydrology. Elsevier, Caspian Pattern (NCP) — an upper level atmospheric Amsterdam, Netherlands, pp. 51 – 86. teleconnection affecting the eastern Mediterranean — Kendall, C., Silva, S.R., Kelly, V.J. 2001.Carbon and nitrogen implications on the regional climate. Theoretical and isotopic compositions of particulate organic matter in four Applied Climatology 72, 173–192. large river systems across the United States. Hydrological Küttel, M., Luterbacher, J., Wanner, H., 2011. Multidecadal Processes 15, 1301–1346. changes in winter circulation-climate relationship in Kissin, I.G., Grinevsky, A.O., 1990. Main features of hydro- Europe: Frequency variations, within-type modifications, geodynamic earthquake precursors. Tectonophysics 178, and long-term trends. Climate Dynamics 36, 957–972. 277–286. Kuzucuoğlu, C., Dörfler, W., Kunesch, S., Goupille, F., 2011. Knapp, A.B., Manning, S.W., 2016. Crisis in context: the end Mid- to late- Holocene climate change in central Turkey: of the late Bronze Age in the Eastern Mediterranean. The Tecer Lake record. The Holocene 21, 173–188. American Journal of Archaeology 120, 99–149. Kylander, M.E., Ampel, L., Wohlfarth, B., Veres, D., 2011. Kociolek, J.P., Stoermer, E.F., 2010. Variation and High resolution X-ray fluorescence core scanning analysis polymorphism in diatoms: The triple helix of of Les Echets (France) sedimentary sequence: new development, genetics and environment. A review of the insights from chemical proxies. Journal of Quaternary literature. Vie et Milieu, Life and Environment 60, 75– Science 26, 109–117. 87. Kyriakopoulos, C., Chini, M., Bignami, C., Stramondo, S., Korakis, G., 2006. Description of vegetation types in Pylos Ganas, A., Kolligri, M., Moshou, A., 2013. Monthly Lagoon and Sfacteria Island, Greece. In: Proceedings of migration of a tectonic seismic swarm detected by the International Conference “Sustainable Management DInSAR: southwest Peloponnese, Greece. Geophysical and Development of Mountainous and Island Areas”, pp. Journal International. 194, 1302–1309. 261–267. Island of Naxos, Greece. Labinger, J.A., Bercaw, J.E., 2002. Understanding and Koutsoubas, D., Dounas, C., Petihakis, G., Arvanitidis, C., exploiting C-H bond activation. Nature 417, 507–514. Triantafyllou, G., Dafnomili, E., Zivanovic, S., 1997. Ladd, S.N., Sachs, J.P., 2012. Inverse relationship between Dynamic changes of physicochemical parameters in salinity and n-alkane dD values in the mangrove Gialova Lagoon (SW Peloponessos, Ionian Sea). In: Avicennia marina. Organic Geochemistry 48, 25–36. Proceedings of the 5th Hellenic Conference on Lawrence, J.R., White, S.D., 1991. The elusive climate signal Oceanography and Fisheries, pp. 309-312. Kavala, Greece in the isotopic composition of precipitation. In: Taylor [in Greek with English Abstract]. H.P. Jr., O’Neil, J.R., Kaplan, I.R. (Eds.), Stable isotope Kraft, J.C., Rapp, G.Jr., Aschenbrenner, S.E., 1975. Late geochemistry: a tribute to Samuel Epstein. The Holocene paleogeography of the coastal plain of the Gulf Geochemical Society, Special Publication 3, St. Louis, of Messenia, Greece, and its relationships to Missouri, pp. 169–185. archaeological settings and coastal change. Geological Lazarova, M., Koutsios, A., Kontopoulos, N., 2012. Holocene Society of America Bulletin 86, 1191–1208. vegetation history of the Kotihi lagoon (northwest Kraft, J.C., Aschenbrenner, S.E., Rapp, G.Jr., 1977. Peloponnesus, Greece). Quaternary International 261, Paleogeographic reconstructions of coastal Aegean 138–145. archaeological sites. Science 195, 941–947. Lim, Y-K., 2015. The East Atlantic/West Russia (EA/WR) Kraft, J.C., Rapp, G.Jr., Aschenbrenner, S.E., 1980. Late teleconnection in the North Atlantic: climate impact and Holocene palaeogeomorphic reconstructions in the area of relation to Rossby wave propagation. Climate Dynamics, the Bay of Navarino: Sandy Pylos. Journal of 44, 3211–3222. Archaeological Science 7, 187–210. Lionello, P., Malanotte-Rizzoli, P., Boscolo, R., Alpert, P., Krammer, K., Lange-Bertalot, H., 1986. Bacillariophyceae Artale, V., Li, L., Luterbacher, J., May, W., Trigo, R., 1.Teil: Naviculaceae. In: Ettl, H., Gerloff, J., Heynig, H., Tsimplis, M., Ulbrich, U., Xoplaki, E., 2006. The Mollenhauer, D. (Eds.), Süßwasserflora von Mitteleuropa Mediterranean climate: an overview of the main 2/1. Stuttgart, New York: Gustav Fischer Verlag. characteristics and issues. In: Lionello, P., Malanotte- Krammer, K., Lange-Bertalot, H., 1988. Bacillariophyceae 2. Rizzoli, P., Boscolo, R. (Eds.), Mediterranean Climate Teil: Bacillariaceae, Epithemiaceae, Surirellaceae. In: Ettl, Variability. Amsterdam: Elsevier (Netherlands), pp. 1–26. H., Gerloff, J., Heynig, H., Mollenhauer, D. (Eds.), Lionello, P., Trigo, I.F., Gil, V., Liberato, M.L.R., Nissen, K., Süßwasserflora von Mitteleuropa 2/2. Stuttgart, New Pinto, J.G., Raible, C., Reale, M., Tanzarella, A., Trigo, York: Gustav Fischer Verlag. R.M., Ulbrich, S., Ulbrich, U., 2016. Objective

climatology of cyclones in the Mediterranean Region: a Marcott, S.A., Shakun, J.D., Clark, P.U., Mix, A.C., 2013. A consensus view among methods with different system reconstruction of regional and global temperature for the identification and tracking criteria. Tellus A: Dynamic past 11,300 years. Science 339, 1198–1201. Meteorology and Oceanography 68, 29391. Maroukian, H., Gaki-Papanastassiou, K., Piteros, Ch., 2004. Liu, W., Yang, H., Li, L., 2006. Hydrogen isotopic Geomorphological and archaeological study of the compositions of n-alkanes from terrestrial plants correlate broader area of the Mycenaean dam of Megalo Rema and with their ecological life forms. Oecologia 150, 330–338. ancient Tiryns, southeastern Argive Plain, Peloponnesus. Liu, W., Yang, H., Wang, H., An, Z., Wang, Z., Leng, Q., Bulletin of the Geological Society of Greece 36, 1154- 2015. Carbon isotope composition of long chain leaf wax 1163. n-alkanes in lake sediments: A dual indicator of Martín-Puertas, C., Valero-Garcés, B.L., Mata, M.P., paleoenvironment in the Qinghai-Tibet Plateau. Organic González-Sampériz, P., Bao, R., Moreno, A., Stefanova, Geochemistry 83–84, 190–201. V., 2008. Arid and humid phases in southern Spain during Loizidou, M., Giannakopoulos, C., Bindi, M., Moustakas, K., the last 4000 years: the Zoñar Lake record, Córdoba. The 2016. Climate change impacts and adaptation options in Holocene 18, 907–921. the Mediterranean basin. Regional Environmental Change May, S.M., Vött, A., Brückner, H., Grapmayer, R., Handl, M., 16, 1859–1861. Wennrich, V., 2012. The Lefkada barrier and beachrock Lorenz, S.J., Kim, J.-H., Rimbu, N., Schneider, R.R., system (NW Greece) - controls on coastal evolution and Lohmann, G., 2006. Orbitally driven insolation forcing on the significance of extreme events. Geomorphology 139– Holocene climate trends: evidence from alkenone data 140, 330–347. and climate modeling. Paleoceanography 21, PA1002. Mayewski, P.A., Meeker, L.D., Twickler, M.S., Whitlow, S., Lyras, G., 2011. River Pamisos, 7000 years of history. In: The Yang, Q., Lyons, W.B., Prentice, M., 1997. Major Valyra of Ithomi. Available at: features and forcing of high-latitude northern hemisphere http://efimeridavalira.blogspot.com/2011/10/7000.html atmospheric circulation using a 110,000-year long glacio- Magny, M., de Beaulieu, J.-L., Drescher-Schneider, R., chemical series. Journal of Geophysical Research 102, Vanniere, B., Walter-Simonnet, A.-V., Miras, Y., Millet, 26345–26366. L., Bossuet, G., Peyron, O., Brugiapaglia, E., Leroux, A., Mayewski, P.A., Rohling, E.E., Stager, J.C., Karlén, W., 2007. Holocene climate changes in the central Maasch, K.A., Meeker, L.D., Meyerson, E.A., Gasse, F., Mediterranean as recorded by lake-level fluctuations at van Kreveld, S., Holmgren, K., Lee-Thorp, J., Rosqvist, Lake Accesa (Tuscany, Italy). Quaternary Science G., Rack, F., Staubwasser, M., Schneider, R.R., Steig, Reviews 26, 1736–1758. E.J., 2004. Holocene climate variability. Quaternary Magny, M., Vannière, B., Calo, C., Millet, L., Leroux, A., Research 62, 243–255. Peyron, O., Zanchetta, G., La Mantia, T., Tinner, W., Mead, R., Xu, Y., Chong, J., Jaffé, R., 2005. Sediment and 2011. Holocene hydrological changes in south-western soil organic matter source assessment as revealed by the Mediterranean as recorded by lake-level fluctuations at molecular distribution and carbon isotopic composition of Lago Preola, a coastal lake in south-ern Sicily, Italy, n-alkanes. Organic Geochemistry 36, 363–370. Quaternary Science Reviews 30, 2459–2475. Meyer, B., Wulf, M., Håkansson, H., 2001. Phenotypic Magny, M., Peyron, O., Sadori, L., Ortu, E., Zanchetta, G., variation of life–cycle stages in clones of three similar Vannière, B., Tinner, W., 2012. Contrasting patterns of Cyclotella species after induced auxospore production. precipitation seasonality during the Holocene in the south- Diatom Research 16, 343–361. and north-central Mediterranean. Journal of Quaternary Meyers, R.A., Ishiwatari, R., 1993. Lacustrine organic Science 27, 290–296. matter–an overview of indicators of organic matter Magny, M., Combourieu-Nebout, N., de Beaulieu, J.L., Bout- sources and diagenesis in lake sediments. Organic Roumazeilles, V., Colombaroli, D., Desprat, S., Francke, Geochemistry 20, 867–900. A., Joannin, S., Ortu, E., Peyron, O., Revel, M., Sadori, Meyers, P.A., 1994. Preservation of elemental and isotopic L., Siani, G., Sicre, M.A., Samartin, S., Simonneau, A., source identification of sedimentary organic matter. Tinner, W., Vannière, B., Wagner, B., Zanchetta, G., Chemical Geology 1148, 289–302. Anselmetti, F., Brugiapaglia, E., Chapron, E., Debret, M., Meyers, P.A., 1997. Organic geochemical proxies of Desmet, M., Didier, J., Essallami, L., Galop, D., Gilli, A., paleoceanographic, paleolimnologic, and paleoclimatic Haas, J.N., Kallel, N., Millet, L., Stock, A., Turon, J.L., processes. Organic Geochemistry 27, 213–250. Wirth, S., 2013. North–south palaeohydrological contrasts Meyers, P.A., Teranes, J.L., 2001. Sediment organic matter. in the central Mediterranean during the Holocene: In: Last, W.M., Smol, J.P. (Eds.), Tracking environmental tentative synthesis and working hypotheses. Climate of change using lake sediments. Volume 2: Physical and the Past 9, 2043–2071. geochemical methods. Kluwer Academic Press, Maheras, P., Xoplaki, E., Davies, T., Martin-Vide, J., Dordrecht, pp. 239–269. Bariendos, M., Alcoforado, M.J., 1999. Warm and cold Meyers, P.A., 2003. Applications of organic geochemistry to monthly anomalies across the Mediterranean basin and paleolimnological reconstructions. A summary of their relationship with circulation; 1860–1990. examples from the Laurentian Great Lakes. Organic International Journal of Climatology 19, 1697–1715. Geochemistry 34, 261–289. Mann, D.G., 1999. The species concept in diatoms. Mook, W.G., 2000. Introduction, theory, methods, review. In: Phycologia 38, 437–495. Mook, W.G. (Eds.), Environmental isotopes in the Marcopoulou-Diacantoni A., Mirkou, M.R., Mariolakos, I., hydrological cycle. Principles and applications. Volume 1. Logos, E., Lozios, S., Fountoulis, I., 1989. Stratigraphic IHP-V Technical Documents in Hydrology. UNESCO - observations at the post-alpine of the Thouria-Ano IAEA, Paris, Vienna, pp. 1–280. Amphia (SW Peloponnese, Greece), and their neotectonic Moossen, H., Bendle, J., Seki, O., Quillmann, U., Kawamura, interpretation. Bulletin of the Geological Society of K., 2015. North Atlantic Holocene climate evolution Greece XXIII, 275–295 (in Greek with English abstract). recorded by high-resolution terrestrial and marine

biomarker records. Quaternary Science Reviews 129, area. Master thesis. Department of Biology, University of 111–127. Patras, Greece (In Greek with English abstract). Moreno, A., Giralt, S., Valero-Garcés, B., Sáez, A., Bao, R., Papazachos, B., Papazachou, C., 2003. The Earthquakes of Prego, R., Pueyo, J.J., González-Sampériz, P., Taberner, Greece. Thessaloniki: ZITI Editions. C., 2007. A 14 kyr record of the tropical Andes: the Lago Papazisimou, S, Christanis, K., Ioakim C., 2005. Holocene Chungará sequence (18°S, northern Chilean Altiplano). peat formation in Southern Peloponnese, Greece: The case Quaternary International 161, 4–21. of Aghios-Phloros fen. Revue de Paléobiologie 24, 657– Muir-Wood, R., King, G.C.P., 1993. Hydrological signatures 672. of earthquake strains. Journal of Geophysical Research Pariente, A., Touchais, G., 1998. Argos et l’Argolide: 98, 22035–22068. topographie et urbanisme. Actes de la Table Ronde Mügler, I., 2008. Paleoclimate reconstruction using internationale Athenes-Argos 28/4-1/5/1990: Argos kai compound – specific hydrogen isotope ratios of n-alkanes. Argolida: topographia kai poleodomia. Praktika diethnous PhD Thesis. Friedrich-Schiller-Universität, Jena. Synedriou Athēna-Argos 28/4–1/5/ 1990. Ecole Francais Mügler, I., Sachse, D., Werner, M., Xu, B., Wu, G., Yao, T., d'Athenes, Athenes. Gleixner, G., 2008. Effect of lake evaporation on δD Park, R., Epstein, S., 1961. Metabolic fractionation of C13 and values of lacustrine n-alkanes: a comparison of Nam Co C12 in plants. Plant Physiology 36, 133–138. (Tibetan Plateau) and Holzmaar (Germany). Organic Pavlopoulos, K., Karkanas, P., Triantaphyllou, M., Geochemistry 39, 711–729. Karympalis, E., Tsourou, Th., Palyvos, N., 2006. Nieto-Moreno, V., Martínez-Ruiz, F., Giralt, S., Jiménez- Paleoenvironmental evolution of the coastal plain of Espejo, F.J., Gallego-Torres, D., Rodrigo-Gámiz, M., Marathon, Greece, during the late Holocene: depositional García-Orellana, J., Ortega-Huertas, M., de Lange, G.J., environment, climate, and sea level changes. Journal of 2011. Tracking climate variability in the western Coastal Research 22, 424–438. Mediterranean during the Late Holocene: a multiproxy Perrier, R., 1975. Geological Map of Greece, Scale 1:5000. approach. Climate of the Past 7, 1395–1414. Meligalas Sheet. Athens: Institution of Geology and Norström, E., Katrantsiotis, C., Smittenberg, R.H., Kouli, K., Mineral Exploration. 2017. Chemotaxonomy in some Mediterranean plants and Peters, K.E., Walters, C.C., Moldowan, J.M., 2005. The implications for fossil biomarker records. Geochimica et biomarker guide, Volume 1: Biomarkers and isotopes in Cosmochimica Acta 219, 96–110. the environment and human history. 2nd Edition. Norström, E., Katrantsiotis, C., Finné, M., Risberg, J., Cambridge University Press, New York. Smittenberg, R., Bjursäter, S., 2018. Biomarker hydrogen Peyron, O., Magny, M., Goring, S., Joannin, S., de Beaulieu, isotope composition (δD) as proxy for Holocene hydro- J.-L., Brugiapaglia, E., Sadori, L., Garfi, G., Kouli, K., climatic change and seismic activity in SW Peloponnese, Ioakim, C., Combourieu-Nebout, N., 2013. Contrasting Greece. Journal of Quaternary Science 33, 563–574. patterns of climatic changes during the Holocene across O'Leary, M.H., 1988. Carbon isotopes in photosynthesis: the Italian Peninsula reconstructed from pollen data. fractionation techniques may reveal new aspects of carbon Climate of the Past 9, 1233–1252. dynamics in plants. BioScience 38, 328–335. Piccarreta, M., Caldara, M., Capolongo, D., Boenzi, F., 2011. Olsen, J., Anderson, N.J., Knudsen, M.F., 2012. Variability of Holocene geomorphic activity related to climatic change the North Atlantic Oscillation over the past 5,200 years. and human impact in Basilicata, Southern Italy. Nature Geoscience 5, 808–812. Geomorphology 128, 137–147. Olsson, I.U., 1983. Dating non-terrestrial materials. In: Mook, Pierantozzi, P., Torres, M., Bodoira, R., Maestri, D., 2013. W.G., Waterbolk, H.T. (Eds.), Proceedings of the 1st Water relations, biochemical – physiological and yield International Symposium 14C and Archaeology. ACT 8, responses of olive trees (Olea europaea L. cvs. Arbequina pp. 277–294. and Manzanilla) under drought stress during the pre- Olsson, I.U., 1991. Accuracy and precision in sediment flowering and flowering period. Agricultural Water chronology. Hydrobiologia 214: 25–34. Management 125, 13–25. O’Neil, J.R., 1986, Theoretical and experimental aspects of Piperno, D.R., 2006. A Comprehensive Guide for isotopic fractionation. In: Valley, J.W., Taylor, H.P., Archaeologists and Paleoecologists. AltaMira Press, O'Neil, J.R. (Eds.), Stable isotopes in high temperature Lanham, MD. geological processes. Reviews in Mineralogy. Volume 16. Pirazzoli, P.A., Stiros, S.C., Fontugne, M., Arnold, M., 2004. Mineralogical Society of America, Washington DC, pp. Holocene and Quaternary uplift in the central part of the 1–40. southern coast of the Corinth Gulf (Greece). Marine PAGES 2k Consortium, 2013. Continental-scale temperature Geology 212, 35–44. variability during the last two millennia. Nature Prasad, S., Kusumgar, S., Gupta, S.K., 1997. A mid to late Geoscience 6, 339–346. Holocene record of palaeoclimatic changes from Nal Pantazopoulou, K., 2015. Artificial enrichment of aqueous Sarovar: a palaeodesert margin lake in western India. formations of the Argive plain. Study of environmental Journal of Quaternary Science 12, 153–59. impacts. Bachelor thesis. Repository of EMT Institute of Prasad, A.K.S.K., Nienow, A., 2006. The centric diatom Technology "Theodoros Trivellas", Kavala, Greece (in genus Cyclotella (Stephanodiscaceae: Bacillariophyta) Greek). from Florida Bay, USA, with special reference to Papadopoulos, G.A., Baskoutas, I., Fokaefs, A., 2014. Cyclotella choctawhatcheeana and Cyclotella Historical seismicity of the Kyparissiakos Gulf, western desikacharyi, a new marine species related to the Peloponnese, Greece. Bollettino di Geofisica Teorica ed Cyclotella striata complex. Phycologica 45, 127–140. Applicata 55, 389–404. Prins, H.B.A., Elzenga, J.T.M., 1989. Bicarbonate utilization: Papakonstantinou, M., 2015. Side Scan Sonar bottom relief function and mechanism. Aquatic Botany 34, 59–83. interpretation and ecological evaluation of Gialova Psomiadis, D., Dotsika, E., Albanakis, K., Ghaleb, B., lagoon, Pref. of Messinia, Greece, as well as Land Cover Hillaire-Marcel, C., 2018. Speleothem record of climatic Land Use (LCLU) digital mapping in the wider protected changes in the northern Aegean region (Greece) from the

Bronze Age to the collapse of the Roman Empire. Round, F.E., Crawford, R.M., Mann, D.G., 1990. The Palaeogeography, Palaeoclimatology, Palaeoecology 489, Diatoms. Biology and Morphology of the Genera. 272–283. Cambridge University Press, Cambridge. Pu, Y., Zhang, H.C., Lei, G.L., Chang, F.Q., Yang, M.S., Rovner, I., 1971. Potential of Opal Phytoliths for Use in Zhang, W.X., Lei, Y.B., Yang, L.Q., Pang, Y.Z., 2010. paleoecological reconstruction. Quaternary Research 1, Climate variability recorded by n-alkanes of paleolake 343–359. sediment in Qaidam Basin on the northeast Tibetan Rozanski, K., Sonntag, C., Munnich, K.O., 1982. Factors Plateau in late MIS3. Science China Earth Sciences 53, controlling stable isotope composition of European 863–870. Precipitation. Tellus 34, 142–150. R Development Core Team, 2013. A language and Ryves, D.B., Juggins, S., Fritz, S.C., Battarbee, R.W., 2001. environment for statistical computing. R Foundation for Experimental diatom dissolution and the quantification of Statistical Computing, Vienna, Austria. microfossil preservation in sediments. Palaeogeography, Raible, C.C., Yoshimori, M., Stocker, T.F., Casty, C., 2007. Palaeoclimatology, Palaeoecology 172, 93–113. Extreme midlatitude cyclones and their implications for Ryves, D.B., Battarbee, R.W., Juggins, S.J., Fritz, S.C., precipitation and wind speed extremes in simulations of Anderson, N.J., 2006. Physical and chemical predictors of the Maunder Minimum versus present day conditions. diatom dissolution in freshwater and saline lake sediments Climate Dynamics 28, 409–423. in North America and West Greenland. Limnology and Reed, J.M., 1998. Diatom preservation in the recent sediment Oceanography 51, 1355–1368. record of Spanish saline lakes: Implications for Ryves, D.B., Anderson, N.J., Flower, R.J., Rippey, B., 2013 palaeoclimate study. Journal of Paleolimnology 19, 129– Diatom taphonomy and silica cycling in two freshwater 137. lakes and their implications for inferring past lake Rehfeld, K., Kurths, J., 2014. Similarity estimators for productivity. Journal of Paleolimnology 49, 411–430. irregular and age-uncertain time series. Climate of the Sabatier, P., Dezileau, L., Colin, C., Briqueu, L., Bouchette, Past 10, 107–122. F., Martinez, P., Siani, G., Raynal, O., Von Grafenstein, Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, U., 2012. 7000 years of paleostorm activity in the NW P.G., Bronk Ramsey, C., Buck, C.E., Cheng, H., Edwards, Mediterranean Sea in response to Holocene climate R.L., Freidrich, M., Grootes, P.M., Guilderson, T.P., events. Quaternary Research 77, 1–11. Haflidason, H., Hajdas, I., Hatté, C., Heaton, T.J., Sachse, D., Radke, J., Gleixner, G., 2004. Hydrogen isotope Hoffmann, D.L., Hogg, A.G., Hughen, K.A., Kaiser, K.F., ratios of recent lacustrine sedimentary n-alkanes record Kromer, B., Manning, S.W., Niu, M., Reimer, R.W., modern climate variability. Geochimica et Cosmochimica Richards, D.A., Scott, E.M., Southon, J.R., Staff, R.A., Acta 68, 4877–4889. Turney, C.S.M., van der Plicht, J., 2013. IntCal13 and Sachse, D., Radke, J., Gleixner, G., 2006. δD values of Marine13 Calibration Curves 0–50,000 Years BP. individual n-alkanes from terrestrial plants along a Radiocarbon 55, 1869–1887. climatic gradient – implications for the sedimentary Renssen, H., Seppä, H., Crosta, X., Goose, H., Roche, D.M., biomarker record. Organic Geochemistry 37, 469–483. 2012. Global characterization of the Holocene Thermal Sachse, D., Billault, I., Bowen, G.J., Chikaraishi, Y., Dawson, Maximum. Quaternary Science Reviews, 48 7– 19. T.E., Feakins, S.J., Freeman, K.H., Magill, C.R., Risi, C., Bony, S., Vimeux, F., 2008. Influence of convective McInerney, F.A., van der Meer, M.T.J., Polissar, P., processes on the isotopic composition (δ18O and δD) Robins, R.J., Sachs, J.P., Schmidt, H.-L., Sessions, A.L., of precipitation and water vapor in the tropics: 2. White, J.W.C., West, J.B., Kahmen, A., 2012. Molecular Physical interpretation of the amount effect. Journal of paleohydrology: interpreting the hydrogen isotopic Geophysical Research 113, D19306. composition of lipid biomarkers from photosynthesizing Rizou, D., Flocas, H.A., Hatzaki, М., Bartzokas, A., 2018. A organisms. Annual Review of Earth and Planetary statistical investigation of the impact of the Indian Sciences 40, 221–249. monsoon on the eastern Mediterranean circulation. Sadori, L., Giraudi, C., Masi, A., Magny, M., Ortu, E., Atmosphere 9, 1–27. Zanchetta, G., 2016. Climate, environment and society in Roberts, N., Reed, J.M., Leng, M.J., Kuzucuoğlu, C., southern Italy during the last 2000 years. A review of the Fontugne, M., Bertaux, J., Woldring, H., Bottema, S., environmental, historical and archaeological evidence. Black, E., 2001. The tempo of Holocene climatic change Quaternary Science Reviews 136, 173–188. in the eastern Mediterranean region: new high-resolution Scheffers, A., Kelletat, D., Vött, A., May, S.M., Scheffers, S., crater-lake sediment data from central Turkey. The 2008. Late Holocene tsunami traces on the western and Holocene 11, 721–736. southern coastlines of the Peloponnesus (Greece). Earth Roberts, N., Moreno, A., Valero-Garcés, B.L., Corella, J.P., and Planetary Science Letters 269, 271–279. Jones, M., Allcock, S., Woodbridge, J., Morellón, M., Schemmel, F., Niedermeyer, E.M., Schwab, V.F., Gleixner, Luterbacher, J., Xoplaki, E., Türkes, M., 2012. G., Pross, J., Mulch, A., 2016. Plant wax δD values record Palaeolimnological evidence for an east-west climate see- changing Eastern Mediterranean atmospheric circulation saw in the Mediterranean since AD 900. Global and patterns during the 8.2 kyr B.P. climatic event. Quaternary Planetary Change 84–85, 23–34. Science Reviews 133, 96–107. Rodwell, M.J, Hoskins, B.J., 1996. Monsoons and the Schmitt, J., Schneider, R., Elsig, J., Leuenberger, D., dynamics of deserts. Quarterly Journal of the Royal Lourantou, A., Chappellaz, J., Köhler, P., Joos, F., Meteorological Society 122, 1385–1404. Stocker, T.F., Leuenberger, M., Fischer, H., 2012. Carbon Roubeix, V., Lancelot, C., 2008. Effect of salinity on growth, isotope constraints on the deglacial CO₂ rise from ice cell size and silicification of an euryhaline freshwater cores. Science 336, 711–714. diatom: Cyclotella meneghiniana Kütz. Transitional Sessions, A.L., Burgoyne, T.W., Schimmelmann, A., Hayes, Waters Bulletin 2, 31–38. J.M., 1999. Fractionation of hydrogen isotopes in lipid biosynthesis. Organic Geochemistry 30, 1193–1200.

Sessions, L.A, 2016. Factors controlling the deuterium area, SW Peloponnisos, Greece. In: Marinos, P., Koukis, contents of sedimentary hydrocarbons. Organic G., Tsiambaos, G. Stournaras, G. (Eds.), Engineering Geochemistry 96, 43–64. Geology and the Environment. Balkema, Rotterdam. pp. Seyfarth, W., 1971. Ammianus Marcellinus. Römische 397–402. Geschichte. Vierter Teil. Buch 26-31. Darmstadt. Tipple, B.J., Pagani, M., 2007. The early origins of terrestrial Shirokawa, Y., Karino, K., Mayama, S., 2012. Development C4 photosynthesis. Annual Review of Earth and Planetary plasticity and genotype-environment interactions Sciences 35, 435–461. influence valve morphology in the Cyclotella Triantaphyllou, M.V., Ziveri, P., Gogou, A., Marino, G., meneghiniana species complex (Bacillariophyceae). Lykousis, V., Bouloubassi, I., Emeis, K-C., Kouli, K., European Journal of Phycology 47, 245–253. Dimiza, M., Rosell-Mele, A., Papanikolaou, M., Sierro, F.J., Hodell, D.A., Curtis, J.H., Flores, J.A., Reguera, Katsouras, G., Nunez, N., 2009. Late Glacial– Holocene I., Colmenero-Hidalgo, E., Bárcena, M.A., Grimalt, J.O., climate variability at the south-eastern margin of the Cacho, I., Frigola, J., Canals, M., 2005. Impact of iceberg Aegean Sea. Marine Geology 266, 182–197. melting on Mediterranean thermohaline circulation during Trigo, I.F., Bigg, G.R., Davies, T.D., 2002. Climatology of Heinrich events. Paleoceanography 20, PA2019. cyclogenesis mechanisms in the Mediterranean. Monthly Sifeddine, A., Wirrmann, D., Albuquerque, A.L.S., Weather Review 130, 549–569. Turcq, B., Cordeiro, R.C., Gurgel, M.H.C., Abrão, J.J., Trigo, R., Xoplaki, E., Zorita, E., Luterbacher, J., Krichak, 2004. Bulk composition of sedimentary organic matter S.O., Alpert, P., Jacobeit, J., Saenz, J., Fernandez, J., used in palaeoenvironmental reconstructions: examples Gonzalez-Rouco, F., Garcia-Herrera, R., Rodo, X., from the tropical belt of South America and Africa. Brunetti, M., Nanni, T., Maugeri, M., Turkes, M., Palaeogeography, Palaeoclimatology, Palaeoecology 214, Gimeno, L., Ribera, P., Brunet, M., Trigo, I.F., Crepon, 41–53. M., Mariotti, A., 2006. Relations between variability in Simpson, R.H., 2014. Mycenaean Messenia and the Kingdom the Mediterranean region and Mid-latitude variability. In: of Pylos. Prehistory Monographs 45. INSTAP Press, Lionello, P., Malanotte-Rizzoli, P., Boscolo, R. (Eds.), Philadelphia. Mediterranean Climate Variability. Elsevier, Amsterdam, Smith F.A., Freeman, K.H., 2006. Influence of physiology Netherlands, pp. 179–226. and climate on δD of leaf wax n-alkanes from C3 and C4 Tyrlis, E., Lelieveld, J., Steil, B., 2013. The summer grasses. Geochimica et Cosmochimica Acta 70, 1172– circulation over the eastern Mediterranean and the Middle 1187. East: influence of the South Asian monsoon. Climate Smith, A.C., Wynn, P.M., Barker, P.A., Leng, M.J., Noble, Dynamics 40, 1103–1123. S.R. Tych, W., 2016. North Atlantic forcing of moisture Tzedakis, P.C., Pälike, H., Roucoux, K.H., de Abreu, L., delivery to Europe throughout the Holocene. Scientific 2009. Atmospheric methane, southern European Reports 6, 24745. vegetation and low–mid latitude links on orbital and Snoeijs, P., Weckström, K., 2010. Diatoms and millennial timescales. Earth and Planetary Science Letters environmental change in large brackish-water ecosystems. 277, 307–317. In: Smol, P.J., Stoermer, F.E. (Eds.), The Diatoms: Ulbrich, U., Lionello, P., Belušić, D., Jacobeit, J., Knippertz, applications for the environmental and earth sciences. P., Kuglitsch, F.G., Leckebusch, G.C., Luterbacher, J., Cambridge University Press, Cambridge, pp. 287–308. Maugeri, M., Maheras, P., Nissen, K.M., Pavan, V., Pinto, Solak, C.N., Ács, É., 2011. Water quality monitoring in J.G., Saaroni, H., Seubert, S., Toreti, A., Xoplaki, E., Ziv, European and Turkish rivers using diatoms. Turkish B., 2012. Climate of the Mediterranean: synoptic patterns, Journal of Fisheries and Aquatic Sciences 11, 329–337. temperature, precipitation, winds and their extremes. In: Stanley, J.D., Krom, M.D., Cliff, R.A., Woodward, J., 2003. Lionello, P. (Eds.), The Climate of the Mediterranean Short contribution: Nile failure at the end of the Old region: from the past to the future. Elsevier, Amsterdam, Kingdom, Egypt: Strontium Isotopic and Petrologic Netherlands, pp. 301–346. Evidence. Geoarchaeology 18, 395–402. Unkel, I., Schimmelman, A., Shriner, C., Forsen, J., Stoermer, E.F., Smol, J.P., 2010. The diatom: applications for Heymann, C., Brückner, H., 2014. The environmental the environmental and earth sciences. Second edition. history of the last 6500 years in the Asea Valley Cambridge University Press, Cambridge. (Peloponnese, Greece) and its linkage to the local Tang, Y., Pang, H., Zhang, W., Li, Y., Wu, S., Hou, S., 2015. archaeological record. Zeitschrift für Geomorphologie 58, Effects of changes in moisture source and the upstream 89–107. rainout on stable isotopes in precipitation – a case study in Vassilopoulou, S., 2010. Morphotectonic analysis of southern Nanjing, eastern China. Hydrology and Earth System Argolis peninsula (Greece) based on ground and satellite Sciences 19, 4293–4306. data by GIS development. Bulletin of the Geological Tanner, B.R., Uhle, M.E., Mora, C.I., Kelley, J.T., Society of Greece 43, 516–526. Schuneman, P.J., Lane, C.S., Allen, E.S., 2010. van Andel, T.H., Zangger, E., Demitrack, A., 1990. Land use Comparison of bulk and compound-specific δ13C analyses and soil erosion in Prehistoric and Historical Greece. and determination of carbon sources to salt marsh Journal of Field Archaeology 17, 379–396. sediments using n-alkane distributions (Maine, USA). Vött, A., 2007. Relative sea level changes and regional Estuarine. Estuarine, Coastal and Shelf Science. 86, 283– tectonic evolution of seven coastal areas in NW Greece 291. since the mid-Holocene. Quaternary Science Review 26, Theriot, E., 1987. Principal component analysis and 894–919. taxonomic interpretation of environmentally related Wang, X.C., Chen, R.F., Berry, A., 2003. Sources and variation in silification in Stephanodiscus preservation of organic matter in Plum Island salt marsh (Bacillariophyceae). British Phycological Journal 22, sediments (MA, USA): long-chain n-alkanes and stable 359–373. carbon isotope compositions. Estuarine, Coastal and Shelf Tiniakos, L., Tiniakou, A., 1997. Geomorphological processes Science 58, 917–928. and environmental considerations in the Navarino Bay

Wang, L., Mackay, A.W., Leng, M.J., Rioual, P., Panizzo, paleoenvironmental indicators. Arctic, Antarctic, and V.N., Lu, H., Gu, Z., Chu, G., Han, J., Kendrick, C.P., Alpine Research 39, 318–331. 2013. Influence of the ratio of planktonic to benthic Wunsam, S., Schmidt, R., Klee, R., 1995. Cyclotella-taxa diatoms on lacustrine organic matter δ13C from (Bacillariophyceae) in lakes of the Alpine region and their Erlongwan maar lake, northeast China. Organic relationship to environmental variables. Aquatic Sciences Geochemistry 54, 62–68. 57, 360–386. Wanner, H., Brönnimann, S., Casty, C., Gyalistras, D., Xoplaki, E., González-Rouco, J.F., Gyalistras, D., Luterbacher, J., Schmutz, C., Stephenson, D.B., Xoplaki, Luterbacher, J., Rickli, R., Wanner, H., 2003. Interannual E., 2001. North Atlantic Oscillation–concepts and studies. summer air temperature variability over Greece and its Surveys in Geophysics 22, 321–381. connection to the large-scale atmospheric circulation and Wanner, H., Beer, J., Butikofer, J., Crowley, T.J., Cubasch, Mediterranean SSTs 1950–1999. Climate Dynamics 20, U., Fluckiger, J., Goosse, H., Grosjean, M., Joos, F., 537–554. Kaplan, J.., Küttel, M., Müller, S.A., Prentice, I.C., Yang, H., Liu, W., Leng, Q., Hren, M.T., Pagani, M., 2011. Solomina, O., Stocker, T.F., Tarasov, P., Wagner, M., Variation in n-alkane δD values from terrestrial plants at Widmann, M., 2008. Mid- to Late Holocene climate high latitude: Implications for paleoclimate change: an overview. Quaternary Science Reviews 27, reconstruction. Organic Geochemistry 42, 283–288. 1791–1828. Zanchetta, G., van Welden, A., Baneschi, I., Drysdale, R., Wanner, H., Solomina, O., Grosjean, M., Ritz, S.P., Jetel, M., Sadori, L., Roberts, N., Giardini, M., Beck, C., Pascucci, 2011. Structure and origin of Holocene cold events. V., Sulpizio R., 2012. Multiproxy record for the last 4500 Quaternary Science Reviews, 30, 3109– 3123. years from Lake Shkodra (Albania/Montenegro). Journal Wanner, H., Mercolli, L., Grosjean, M., Ritz, S.P., 2014. of Quaternary Science 27, 780–789. Holocene climate variability and change; a data-based Zanchetta, G., Bar-Matthews, M., Drysdale, R.N., Lionello, review. Journal of the Geological Society 172, 254–263. P., Ayalon, A., Hellstrom, J.C., Isola, I., Regattieri, E., Weiberg, E., Unkel, I., Kouli, K., Holmgren, K., Avramidis, 2014. Coeval dry events in the central and eastern P., Bonnier, A., Dibble, F., Finné, M., Izdebski, A., Mediterranean basin at 5.2 and 5.6 ka recorded in Corchia Katrantsiotis, C., Stocker, S.R., Andwinge, M., Baika, K., (Italy) and Soreq caves (Israel) speleothems. Global and Boyd, M., Heymann, C., 2016. The socio-environmental Planetary Changes 122, 130–139. history of the Peloponnese during the Holocene: Towards Zangger, E., 1991. Prehistoric coastal environments in an integrated understanding of the past. Quaternary Greece: the vanished landscapes of Dimini Bay and Lake Science Reviews 136, 40–65. Lerna. Journal of Field Archaeology 18, 1–15. Weiss, H., Courty, M.A., Wetterstrom, W., Guichard, F., Zangger, E., Timpson, M.E., Yazvenko, S.B., Kuhnke, F., Senior, L., Meadow, R., Curnow, A., 1993. The genesis Knauss, J., 1997. The Pylos Regional Archaeological and collapse of third millennium north mesopotamian Project, Part II: Landscape Evolution and Site civilization. Science 261, 995–1004. Preservation. Hesperia 66, 549–641. Weninger, B., Clare, L., Rohling, E.J., Bar-Yosef, O., Böhner, Zech, M., Zech, R., Buggle, B., Zöller, L., 2011. Novel U., Budja, M., Bundschuh, M., Feurdean, A., Gebel, G.- methodological approaches in loess research– H., Jöris, O., Linstädter, J., Mayewski, P., Mühlenbruch, interrogating biomarkers and compound specific stable T., Reingruber, A., Rollefson, G., Schyle, D., Thissen, L., isotopes. E&G Quaternary Science Journal 60, 170–187. Todorova, H., Zielhofer, C., 2009. The impact of rapid Zelilidis, A., Doutsos, T., 1992. An interference pattern of climate change on prehistoric societies during the neotectonic faults in the southwestern part of the Hellenic Holocene in the Eastern Mediterranean. Documenta Forearc Basin. Zeitschrift der Deutschen Gesellschaft 143, Praehistorica 36, 7–59. 95–105. Wiencke, M.H., 2010. Lerna. In: Cline, E.H. (Ed.), The Zelilidis, A., Kontopoulos, N., 1994. Pliocene-Pleistocene Oxford Handbook of the Bronze Age Aegean. Oxford fluvial/ wave-dominated deltaic sedimentation: The University Press, Oxford, pp. 660–670. Pamisos delta, southwest Peloponnesus, Greece. Wieser, M.E., Brand, W.A., 2017. Isotope ratio studies using Geological Magazine 131, 653–668. mass spectrometry. In: Lindon, J.C., Tranter, G.E., Zielhofer, C., Köhler, A., Mischke, S., Benkaddour, A., Koppenaal, D.W. (Eds.), Encyclopedia of spectroscopy Mikdad, A., Fletcher, W.J., 2018. Western Mediterranean and spectrometry. Third Edition. Elsevier, Amsterdam, hydro-climatic consequences of Holocene iceberg Netherlands, pp. 488– 500. advances (Bond events). Climate of the Past, Discussion, Willershäuser, T., Vött, A., Hadler, H., Ntageretzis, K., Emde, 1–20. K., Brückner, H., 2015. Holocene palaeotsunami imprints Ziller, S., Montesanto, B., 2004. Phytobenthos (diatoms) and in the stratigraphical record and the coastal water framework directive implementation: The case of geomorphology of the Gialova Lagoon near Pylos two Mediterranean rivers in Greece. Fresenius (southwestern Peloponnese, Greece). Zeitschrift für Environmental Bulletin 13, 128–138. Geomorphologie 59, 215–252. Ziv, B., Saaroni, H., Alpert, P., 2004. The factors governing Wilson, G.P., Reed, J.M., Lawson, I.T., Frogley, M.R., the summer regime of the eastern Mediterranean. Preece, R.C., Tzedakis, P.C., 2008. Diatom response to International Journal of Climatology 24, 1859–1871. the last glacial-interglacial transition in the Ioannina basin, northwest Greece: implications for Mediterranean palaeoclimate reconstruction. Quaternary Science Reviews 27, 428–440. Wooller, M.J., Zazula, G.D., Edwards, M., Froese, D.G., Boone, R.D., Parker, C., Bennett, B., 2007. Stable carbon isotope compositions of Eastern Beringian grasses and sedges: investigating their potential as