We compiled two recent high-resolution climate reconstructions (Büntgen et al. 2005a, b) estimating millennial-long temperature variations in the European Alps. These records show similar long-term behavior including the Medieval Warm Period (MWP), Little Ice Age (LIA), and recent warmth (Grove 1988, IPCC 2002, Lamb 1965), however, vary in their estimates of the absolute reconstructed temperature amplitude. Reasons for these amplitude ranges are manifold, including the utilization of differing tree-ring parameters (ring width and maximum latewood density), slightly varying instrumental targets, and differing calibration methods. This regional-scale finding is even more striking for the Northern Hemisphere, where uncertainty in reconstructed amplitudes is in the order of the total variance estimated over the last millennium (Esper et al. 2005a, b).
Herein, we more systematically study the effect of utilizing instrumental target data on the reconstructed Alpine temperature amplitude. We re-calibrate the two temperature reconstructions (Büntgen et al. 2005a, b) against two differing 20th century instrumental datasets representing the same area in the Alps, and test the effect of changing calibration periods and seasonal means on the obtained amplitude. We show that reconstructed amplitude ranges in the order of ~1-1.4°C solely result from scaling against different instrumental target datasets. Further amplitude variations are quantified and their relevance for the 'true' long-term course of Alpine temperature variability discussed.