Reconstruction of late Glacial and Early Holocene near surface temperature anomalies in Europe and their statistical interpretation

摘要

This paper presents the results of a variational analysis study based on the recently developed method by Gebhardt et al. to reconstruct paleo temperature fields on the typical scale of climate model output, providing the necessary basis for quantitative comparison of reconstructions and model results. Additionally, an extension of the approach takes into account the full statistics of the analysis error of the variational analysis to study the inherent uncertainties of the proxy data representing the imperfect paleo climate information and their impact on the field reconstruction. The method is applied to reconstruct realizations of monthly mean temperature anomaly fields in Europe for January and July in Europe as averages for several time slices in the Late Glacial and Holocene. Single realizations can differ substantially from the field of expectation values in terms of spatial patterns due to the inherent correlation of the spatial uncertainties of the paleo proxy data. Additionally, the fields of expectation values generally are much smoother than the realizations. This indicates that it is misleading to only consider expectation values. Further, the ensembles of realization are used to calculate robust error measures for area mean values, which is valuable for the quantitative comparison of the results of different time slices.