1Max-Planck-Institute for Meteorology, Bundesstr. 53, 20146 Hamburg, Germany
2University of Hamburg, KlimaCampus Hamburg, Germany
3Institute for Coastal Research, Helmholtz Centre Geesthacht, Germany
4Bert Bolin Centre for Climate Research, University of Stockholm, Sweden
Abstract. Are simulations and reconstructions of past climate and its variability comparable with each other? We assess if simulations and reconstructions are consistent under the paradigm of a statistically indistinguishable ensemble. Ensemble consistency is assessed for Northern Hemisphere mean temperature, Central European mean temperature and for global temperature fields for the climate of the last millennium. Reconstructions available for these regions are evaluated against the simulation data from the community simulations of the climate of the last millennium performed at the Max Planck Institute for Meteorology.
The distributions of ensemble simulated temperatures are generally too wide at most locations and on most time-scales relative to the employed reconstructions. Similarly, an ensemble of reconstructions is too wide when evaluated against the simulation ensemble mean.
Probabilistic and climatological ensemble consistency is limited to sub-domains and sub-periods. Only the ensemble simulated and reconstructed annual Central European mean temperatures for the second half of the last millennium demonstrates consistency.
The lack of consistency found in our analyses implies that, on the basis of the studied data sets, no status of truth can be assumed for climate evolutions on the considered spatial and temporal scales and, thus, assessing the accuracy of reconstructions and simulations is so far of limited feasibility in pre-instrumental periods.