GEOCHRONOMETRIA 42 (2015): 159–171 DOI 10.1515/geochr-2015-0019 Available online at http://www.degruyter.com/view/j/geochr COMPARING VARVE COUNTING AND 14C-AMS CHRONOLOGIES IN THE SEDIMENTS OF LAKE ŻABIŃSKIE, NORTHEASTERN POLAND: IMPLICATIONS FOR ACCURATE 14C DATING OF LAKE SEDIMENTS ALICJA BONK1, WOJCIECH TYLMANN1, TOMASZ GOSLAR2, 3, AGNIESZKA WACNIK4 and MARTIN GROSJEAN5 1Faculty of Oceanography and Geography, University of Gdańsk, Bażyńskiego 4, PL80952 Gdańsk, Poland 2Faculty of Physics, Adam Mickiewicz University, Umultowska 85, PL61614 Poznań, Poland 3Poznań Radiocarbon Laboratory, Foundation of the Adam Mickiewicz University, Rubież 46, PL61612 Poznań, Poland 4W. Szafer Institute of Botany, Polish Academy of Sciences, Lubicz 46, PL31512 Kraków, Poland 5Oeschger Centre for Climate Change Research and Institute of Geography, University of Bern, Erlachstrasse 9a, CH3012 Bern, Switzerland Received 28 May 2015 Accepted 7 September 2015 Abstract: Varved lake sediments from Lake Żabińskie (northeastern Poland) provide a high- resolution calendar-year chronology which allows validation of 14C dating results. Microscopic analy- sis of the varve microfacies revealed that laminations found in Lake Żabińskie were biogenic (calcite) varves. Three independent counts indicated a good preservation quality of laminae in the 348 cm long +12 sediment profile which contained 1000 /-24 varves. The varve chronology was validated with the 137Cs activity peaks, the tephra horizon from the Askja eruption at AD 1875 and with the timing of major land-use changes of known age inferred from pollen analysis. 32 AMS 14C dates of terrestrial macrofossils distributed along the profile were compared with the varve chronology. After identifica- tion of outliers, the free-shape model performed with 21 14C dates provided the best possible fit with the varve chronology. We observed almost ideal consistency between both chronologies from the pre- sent until AD 1250 while in the lower part (AD 1000–1250) the difference increases to ca. 25 years. We demonstrate that this offset can be explained by too old radiocarbon ages of plant remains trans- ported to the lake by the inflowing creek. Results of this study highlight that careful interpretation of radiocarbon age-depth models is necessary, especially in lakes where no annual laminations are ob- served and no independent method are used for cross-validation. | downloaded: 13.3.2017 Keywords: lake sediments, varve chronology, AMS 14C dating, radionuclides, age-depth model. 1. INTRODUCTION found in proglacial environments (Zolitschka, 2007). Further studies have shown that different processes may Varves were recognized and described for the first lead to seasonal changes in sediment composition and time by Swedish geologist Gerard De Geer who intro- accumulation of distinguishable sediment layers due to duced the term “varve” for clastic silt-clay sediments e.g. variations in biological activity, chemical processes and influx of terrestrial material (Zolitschka et al., 2015). Annually laminated (varved) sediments are preserved in Corresponding author: A. Bonk lakes under different climate conditions and environmen- http://boris.unibe.ch/73473/ e-mail: [email protected] tal settings. At present, varved lake sediments are regard- ed as one of the best archives to chronicle climatic and ISSN 1897-1695 (online), 1733-8387 (print) © 2015 A. Bonk et al. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License. source: Bereitgestellt von | Universitätsbibliothek Bern Angemeldet Heruntergeladen am | 27.11.15 11:11 1000-YEARS CHRONOLOGY FOR LAKE ŻABIŃSKIE, NORTHEASTERN POLAND environmental changes (Ojala, 2001; Snowball et al., This paper describes the establishment and validation 2002). They are also important recorders of hydrological of a 1000-yr long chronology for the varved sediment and limnological processes within lakes (Larsen et al., record from Lake Żabińskie, northeastern Poland. The 1998; Ojala et al., 2000; Zillén et al., 2003; Bonk et al., exceptional scientific value of this site is due to a very 2015). good preservation of varves in the sediments and high One of the major advantages of varved sediments is sedimentation rates which allow high-resolution (annual) their potential for developing accurate and high- analysis of different proxies. The lake already showed its resolution chronologies in calendar-year time scales for high potential for quantitative paleoenvironmental recon- relatively long periods of time (Kitagawa and van der structions (Amann et al., 2014; Hernández-Almeida et Plicht, 1998; Ralska-Jasiewiczowa et al., 1998; Brauer et al., 2015; Larocque-Tobler et al., 2015). Therefore, the al., 2001; Lücke et al., 2003; Stanton et al., 2010). Typi- best possible accuracy of the chronology is crucial for the cally, for well-preserved varved sequences it is possible reliability of the millennial-long climate reconstructions to obtain varve counting uncertainties as low as 1–2% derived from the sediments of this lake. To obtain this, (Kinder et al., 2013). Ideally, varve counting should be we employed two dating methods, i.e. continuous varve supported by experimental verification of the annual counting and 14C AMS dating of terrestrial macrofossils. nature of the observed laminations by sediment trap stud- Additionally, the uppermost part of the profile was vali- ies as well as microscopic analysis of varve structures and dated with the 137Cs activity peaks and the Askja AD compositions of individual laminae. Moreover, validation 1875 tephra horizon (Tylmann et al., submitted). In the of the varve chronology with additional independent age- lower part, where no other chronostratigraphic markers controls is necessary. This is usually made with chronos- were available, we used pollen data and historical infor- tratigraphic markers of known ages or application of mation to check the reliability of the varve- and radiocar- radiometric dating methods such as 137Cs, 210Pb and 14C bon-based chronologies against major land-use changes (Brauer et al., 2000; Ojala and Tiljander, 2003; Enters et of known ages in the catchment. In this respect, we will al., 2006; Stanton et al., 2010; Kinder et al., 2013; (1) provide a varve-based chronology for the sediment Tylmann et al., 2013). If available, volcanic ash (tephra) record of Lake Żabińskie, (2) present results of radiocar- horizons are excellent chronostratigraphic markers that bon dating of terrestrial macrofossils found in the sedi- indicate precisely the age of the deposited layers (Brauer ments, (3) show that analysis of 14C ages and indication et al., 1999; Lowe, 2001; Wulf et al., 2013). Also biolog- of outliers can help in developing a robust chronology, ical proxies may provide robust and independent valida- and (4) explain and discuss potential reasons of discrep- tion of the sediment ages (biostratigraphy). For example, ancies between varve- and radiocarbon-based chronolo- pollen data in combination with historical information gies. about human activities in the lake surroundings enable the reconstruction of environmental events and thus may 2. STUDY SITE help validate the varve chronology in certain time win- dows (Seppä, 2007). Lake Żabińskie is located in the Masurian Lake Dis- Comparison between varve-based chronologies and trict in northeastern Poland (Fig. 1). The landscape in this independent dating methods is not always straightforward area was formed during the Pomeranian Phase of the and, in some cases, it is difficult to obtain consistent Vistulian (Weichselian) glaciation (Szumański, 2000). results. Differences or offsets between varve chronologies Hence, the catchment area is dominated by glacial sedi- and calibrated radiocarbon ages are often reported in the ments with fluvial and biogenic sediments in river valleys literature (Stanton et al., 2010; Mellström et al., 2013). and waterlogged areas, respectively. The northwestern The problem is particularly pronounced when the time- and southwestern parts of the direct catchment are domi- scale of interest is relatively short, e.g. the last millenni- nated by forests that consist of pine, spruce and birch um, very high chronological accuracy is required and trees. Cultivated areas and wetlands dominate in the east- even small differences between both chronologies be- ern part. come crucial for accurate interpretation of proxy records, Lake Żabińskie is a small (41.6 ha) and deep lake e.g. as for detection/attribution studies that allow recogni- (44.4 m). The lake basin is slightly elongated in W-E tion of forced and unforced climate variability (Hegerl et direction and two parts can be distinguished: a shallow al., 2011). High-resolution dating of such records remains basin in the western part and the deep central basin in the a challenging task that requires multiple methods and middle of the lake. The lake has three inflows and one advanced algorithms for age-depth modeling. Excellent outflow (O1) that discharges water westward to the much preservation of varves allowing for highly accurate cal- larger Lake Gołdopiwo. The major inflow is from Lake endar-year chronologies provide a unique opportunity to Purwin (I1) in the north; two minor creeks feed Lake assess potential depositional lags or re-depositional ef- Żabińskie from the south (I2 and I3). The mouth of the I3 fects of individual samples (14C outliers) or to assess creek, which forms a small delta, is situated close to the chronologies with 14C age-depths models (e.g., Blaauw,