Abstract
Abstract. Are simulations and reconstructions of past climate and its variability consistent with each other? We assess the consistency of simulations and reconstructions for the climate of the last millennium under the paradigm of a statistically indistinguishable ensemble. In this type of analysis, the null hypothesis is that reconstructions and simulations are statistically indistinguishable and, therefore, are exchangeable with each other. Ensemble consistency is assessed for Northern Hemisphere mean temperature, Central European mean temperature and for global temperature fields. Reconstructions available for these regions serve as verification data for a set of simulations of the climate of the last millennium performed at the Max Planck Institute for Meteorology. Consistency is generally limited to some sub-domains and some sub-periods. Only the ensemble simulated and reconstructed annual Central European mean temperatures for the second half of the last millennium demonstrates unambiguous consistency. Furthermore, we cannot exclude consistency of an ensemble of reconstructions of Northern Hemisphere temperature with the simulation ensemble mean. If we treat simulations and reconstructions as equitable hypotheses about past climate variability, the found general lack of their consistency weakens our confidence in inferences about past climate evolutions on the considered spatial and temporal scales. That is, our available estimates of past climate evolutions are on an equal footing but, as shown here, inconsistent with each other.
Highlights
We further evaluate the consistency of the temporal evolutions over the last millennium of the COSMOS-Mill simulation ensemble with reconstructions for Central European mean temperature (Dobrovolnyet al., 2010) and a temperature field reconstruction (Mann et al, 2009)
We first evaluate the consistency of the SIM ensemble relative to two reconstruction targets: the Central European temperature data by Dobrovolnyet al. (2010) and the ensemble mean of the Northern Hemisphere temperature Full ensemble reconstruction range (FRA) ensemble
If no reliable observable target can be identified, as is the case in periods and regions without www.clim-past.net/9/1089/2013/. Instrumental observations, such statistical analyses reduce the subjectivity in comparing simulation ensembles and statistical approximations from paleo-sensor data (Braconnot et al, 2012) under uncertainty and go beyond “wiggle matching”
Summary
We assess the consistency of simulations and reconstructions for the climate of the last millennium 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 Reconstructions available for these regions serve as verification data for a set of simulations of the climate of the last millennium performed at the Max Planck Institute for Meteorology. Probabilistic consistency is commonly evaluated by ranking the verification target data against the ensemble data (Anderson, 1996; Jolliffe and Primo, 2008; Annan and Hargreaves, 2010; Marzban et al, 2011; Hargreaves et al, 2011). Appendix A presents more details on the test and discusses the chosen approach
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
