Logo: to the web site of Uppsala University

Ice cores from the relatively low-lying ice caps in Svalbard have not been widely exploited in climatic studies owing to uncertainties about the effect of meltwater percolation. However, results from two new Svalbard ice cores, at Lomonosovfonna and Austfonna, have shown that with careful site selection, high-resolution sampling and multiple chemical analyses it is possible to recover ice cores from which part of the annual signals are preserved, despite the considerable meltwater percolation. The new Svalbard ice cores are positioned in different parts of Svalbard and cover the past 800 years. In this paper we focus on the last 400 years. The 6180 signals from the cores are qualitatively similar over most of the twentieth century, suggesting that they record the same atmospheric signal. Prior to AD 1920, the Austfonna ice core exhibits more negative 6180 values than Lomonosovfonna, although there are intermittent decadal-scale periods throughout the record with similar values. We suggest that the differences reflect the effect of the inversion layer during the winter. The pattern in the ₆18₀ records is similar to the Longyearbyen airtemperature record, but on an annual level the correlation is low. The Austfonna record correlates well with the temperature record from the more distant and southwesterly located Jan Mayen. A comparison of the ice-core and sea-ice records from this period suggests that sea-ice extent and Austfonna 6180 are related over the past 400 years. This may reflect the position of the storm tracks and their direct influence on the relatively low-altitude Austfonna. Lomonosovfonna may be less sensitive to such changes and primarily record free atmospheric changes instead of variations in sea-ice extent, the latter is probably a result of its higher elevation.