JOURNAL OF QUATERNARY SCIENCE (2006) 21(8) 831–841 Copyright ß 2006 John Wiley & Sons, Ltd. Published online 30 May 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jqs.1004 Quantitative palaeotemperature records inferred from fossil pollen and chironomid assemblages from Lake Gilltja ¨rnen, northern central Sweden KARIN ANTONSSON, 1 * STEPHEN J. BROOKS, 2 HEIKKI SEPPA ¨ , 3 RICHARD J. TELFORD 4 and H. JOHN B. BIRKS 4,5,6 1 Department of Earth Sciences, Uppsala University, Uppsala, Sweden 2 Department of Entomology, Natural History Museum, London, England 3 Department of Geology, University of Helsinki, Helsinki, Finland 4 Bjerknes Centre for Climate Research, Bergen, Norway 5 Department of Biology, University of Bergen, Bergen, Norway 6 Environmental Change Research Centre, University College London, London, England Antonsson, K., Brooks, S. J., Seppa ¨, H., Telford, R. J. and Birks, H. J. B. 2006. Quantitative palaeotemperature records inferred from fossil pollen and chironomid assemblages from Lake Gilltja ¨rnen, northern central Sweden. J. Quaternary Sci., Vol. 21 pp. 831–841. ISSN 0267–8179. Received 1 September 2005; Revised 22 November 2005; Accepted 18 December 2005 ABSTRACT: Palaeotemperature reconstructions based on radiocarbon-dated fossil pollen and chironomid stratigraphies obtained from Lake Gilltja ¨rnen provide evidence of climate changes dur- ing the last 11 000 years in the boreal zone of northern central Sweden. The records show consistent trends during the early and mid-Holocene, indicating low temperatures at 11 000–10 000 cal. yr BP, followed by a rising trend and a period of maximum values from about 7000 to 4000 cal.yr BP. At 3000 cal. yr BP the chironomid-inferred temperature values rise abruptly, deviating from the late-Holocene cooling trend indicated by the pollen-based reconstruction and most of the other palaeotemperature records from central Scandinavia, probably as a result of local limnological changes in Lake Gilltja ¨rnen and its catchment. Comparison of the present results with a lake-level reconstruction from Lake Ljustja ¨rnen, ca. 100 km southwest of Lake Gilltja ¨rnen, shows that the low early-Holocene temperatures were associated with high lake-levels at 10 500–8500 cal. yr BP, whereas low lake-levels and dry conditions prevailed during the period of high temperatures at between 7500 and 5000 cal.yr BP, probably due to high summer evapotranspiration and lower precipitation. Copyright ß 2006 John Wiley & Sons, Ltd. KEYWORDS: sediments; pollen; chironomids; Holocene thermal maximum; dryness. Introduction Recent advances in quantitative environmental reconstruction techniques have led to the rapid expansion of fossil-based quantitative palaeoclimatological reconstructions. Such inves- tigations are usually based on mathematical transfer functions that transform fossil assemblages into quantitative estimates of the past climate (Birks, 1995, 2003; ter Braak, 1995). In north- ern Europe recent quantitative temperature estimates have been produced in the tree-line region of northern Scandinavia and Russia (Korhola et al., 1999; Brooks and Birks, 2001; Rose ´n et al., 2001, 2003; Seppa ¨ and Birks, 2001, 2002; Bigler et al., 2002; Hammarlund et al., 2002, 2004; Seppa ¨ et al., 2002; Birks and Birks, 2003; Bjune et al., 2004, 2005; Larocque and Hall, 2003; Solovieva et al., 2005; Velle et al., 2005a, b) and in the southern boreal and temperate zone of the Baltic countries and Scandinavia (Heikkila ¨ and Seppa ¨, 2003; Seppa ¨ and Poska, 2004; Veski et al., 2004; Seppa ¨ et al., 2005). In general, these reconstructions have shown consistent temperature patterns during the Holocene, characterised by a cooling during the last 4000 to 5000 yr and higher values during the early to mid- Holocene. However, these studies have also highlighted significant inconsistencies, especially in fine-scale variability between reconstructions based on different techniques or carried out at different sites (Rose ´n et al., 2001; Seppa ¨ and Birks, 2002). Such inconsistencies show that any proxy technique is subject to errors and biases that can result, for example, from the influ- ence of non-climatic factors on the distribution and abundance of the organisms used. A potential solution for assessing the reliability of the fine-scale variability is the combined use of several proxy techniques (Ammann et al. 2000; Lotter 2003). The assumption here is that in the multiproxy approaches such past reconstruction patterns that can be detected with several independent proxy techniques reflect more reliably past cli- matic variability. An additional value obtained from using * Correspondence to: K. Antonsson, Department of Earth Sciences, Uppsala University, Villava ¨gen 16, SE-75236, Uppsala, Sweden. E-mail: karin.antonsson@geo.uu.se