Mid-Holocene climate change in Europe: a data-model comparison

Abstract

Abstract. We present here a comparison between the outputs of a set of 25 climate models run for the mid-Holocene period (6 ka BP) with a set of palaeo-climate reconstructions from over 400 fossil pollen sequences distributed across the European continent. Three climate parameters were available (moisture availability, temperature of the coldest month and growing degree days), which were then grouped together using cluster analysis to provide regions of homogenous climate change. Each model was then investigated to see if it reproduced 1) the same directions of change and 2) the correct location of these regions. A fuzzy logic distance was used to compare the output of the model with the data, which allowed uncertainties from both the model and data to be taken into account. The initial comparison showed that the models were only capable of reproducing regions of little climate change, as the data-based reconstructions have a much larger range of changes due to their local nature. A correction for the model standard deviation was then applied to allow the comparison to proceed, and this second test shows that the majority of models simulate all the observed patterns of climatic change, although most do not simulate the observed magnitude of change. The models were then compared by distance to data, and by the amount of correction required. The majority of the models are grouped together both in distance and correction, suggesting that they are becoming more consistent. A test against a set of zero anomalies (no climate change) shows that, whilst the models are unable to reproduce the exact patterns of change, they all produce the correct direction of change for the mid-Holocene.