Diatoms as temperature indicators: Comparison of Holocene inferences from seven subarctic lakes in northern Sweden
by
Christian Bigler
Climate Impacts Research Centre and Umeå University, Department of Ecology and Environmental Science
Coauthors: Peter Rosén, Ingemar Renberg

The study aims to assess the consistency of quantitative diatom-inferred July air temperature reconstructions from Holocene sediment cores from seven subarctic lakes in northern Sweden. Regional calibration-sets were developed covering broad climatic and ecological gradients with lakes situated in pine forest, birch forest and alpine areas. Transfer functions were developed using WA-PLS (weighted averaging partial least squares) regression and calibration based on surface sediments from the Laponia (n=60, r2 = 0.74, RMSEP = 0.88°C) and Abisko regions (n=100, r2 = 0.74, RMSEP = 0.96°C). The transfer functions were applied in the World Heritage site Laponia and around Abisko National Park. The areas are characterised by low human population density and low levels of atmospheric pollution deposition. The seven study lakes are relatively small (<0.7km2) and shallow (<18m), with an ice-covered period from autumn until early summer. The study clearly demonstrates that subfossil diatom assemblages contain climate-related information. The present July air temperature at each lake can be inferred accurately within the prediction error of the models. Though the inferred Holocene temperature changes were of the same magnitude as the prediction errors in the models, the diatom records in all seven lakes consistently infer a gradual negative temperature trend since c. 7, 000 cal BP until the present. However, the magnitude of the temperature decrease differs between the seven lakes. In Lakes Sjuodjijaure and Vuoskkujávri, the diatom-inferred temperature decreased gradually by c. 1.5°C since c. 7, 000 cal BP. In Lakes Niak, Jeknajaure and Njulla, a temperature decrease of c. 1.0°C since 7, 000 cal BP was recorded, whereas the changes recorded from Lakes Seukokjaure and 850 were minor. Before 7000 cal BP, substantial discrepancies in the inferred temperatures between the different lakes were observed. In some intervals, the diatom assemblages from the early Holocene showed a poor fit with the calibration-set samples (a so-called "no-modern-analogue" situation), and the reconstructions should therefore be interpreted with caution. It is likely that the reconstructions are influenced by local, lake-specific impacts such as glacier remains, permanent snowfields and geomorphological processes such as pedogenesis during the early Holocene. This is illustrated by non-synchronous loss-on-ignition curves for the lakes. These processes can have a large impact on the diatom assemblages and may override the temperature signal. This study clearly demonstrates the importance of a multi-site approach for interpretation of diatom-inferred climate reconstructions.

Date received: May 2, 2001

Copyright © 2001 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # cahi-71.