Following the Younger Dryas, the first evidence of glacial activity in southern Norway in response to climatic change occurred during the early Holocene, when terminal moraines from the Erdalen Event formed beyond the position of later ‘Little Ice Age’ moraines. In front of present outlet glaciers on both sides of the Jostedalsbreen ice cap, sets of pre-‘Little Ice Age’ marginal moraines have been morphostratigraphically corre lated with the Erdalen Event. The lithostratigraphy in a peat bog at Hellemyrane in Jostedalen, western Norway, provides firm evidence for a two-phase Erdalen Event associated with two glacier readvances of Nigardsbreen at the Preboreal/Boreal transition. Radiocarbon dates from the peat bog give age estimates for the two phases: the first Erdalen Event readvance took place between 10 100 and 10 050 cal. BP, while the second occurred close to 9700 cal. BP. The lowering of the equilibrium-line altitude (ELA) during these readvances is estimated using the ‘Little Ice Age’ ratio (LR) approach to be about 230 m and 215 m, respectively. The first readvance is attributed to an increase in winter precipitation, while the second readvance may have been primarily caused by a marked drop in summer temperature which reactivated already existing ice masses.

Much renewed research interest in Arctic regions stems from the increasing concentration of atmospheric greenhouse gases and the alleged climatic sensitivity of high latitude areas. Glacier and permafrost changes are among a number of proxies used for monitoring past and present Arctic climate change. Here we present observations on frozen in situ soil and vegetation, found below cold-based glacier Longyearbreen (78813?N), 2 km upstream from the present glacier terminus. Dating of the relict vegetation indicates that the glacier has increased in length from about 3 km to its present size of about 5 km during the last c. 1100 years. The meteorological setting of non-surging Longyearbreen suggests this example of late-Holocene glacier growth represents a widespread phenomenon in Svalbard and in adjoining Arctic regions. In addition, we use the subglacial permafrozen soil system to evaluate microbial survival capacity over considerable time periods, and we present evidence for microbes having survived more than 1100 years in a subglacial, permafrozen state.

Knowledge of the relationship between vegetation and modern pollen deposition is necessary to improve the interpretation of fossil pollen samples. We compared the cover of plant species and the modern pollen deposition in surface moss samples in 49 plots in rich-fen vegetation from the boreal vegetation zone in Norway in two areas (one oceanic and one more continental). These rich fens were used for traditional hay cutting until ca. 1950. Three management regimes are used: 1) biennial mowing 2) quadrennial mowing and 3) unmown. The unmown areas were traditional hay fens, with no management since the 1950s. Of the 89 plant taxa and corresponding 65 taxa of pollen and spores, 29 and 22, respectively, were sufficiently represented in both vegetation and as pollen, to allow direct comparisons of the two data sets. Most of the plant and pollen taxa were found in the plots from both study areas and in plots with different mowing frequency. Therefore, quantitative differences in vegetation cover and pollen percentages were the main focus in this investigation of plantpollen relationships in mown and unmown vegetation. The relative difference in plant cover was largest between biennially and unmown plots. 12 species showed a significant positive correlation with difference in plant cover between mown and unmown areas (p<0.05; Wilcoxon signed rank test), most pronounced for Molinia caerulea (most common in unmown plots), and Carex dioica and Thalictrum alpinum (most common in mown plots). A significant positive plant-pollen correlation (Spearman rank-correlation) was found for 10 pollen taxa. Thalictrum alpinum is a very good pollen indicator, and Cyperaceae, Eriophorum-type and Pedicularis-type are good pollen indicators of mowing. Poaceae/Molinia caerulea and Succisa pratensis are negatively correlated with mowing. The impact of mowing on species composition was similar in both study areas. Mown and unmown areas were differentiated in the vegetation data and could be identified by modern pollen data. These results will improve the interpretation of past land-use practices using pollen analysis of rich-fen vegetation.