Reanalysing the deglaciation with models and data

Using data assimilation techniques for climate reanalysis should provide the best description of how and why our climate has changed through the past and up to the present. Due to both computational and data limitations, previous paleoclimate reanalyses of the authors and others have typically focussed either on time slices (such as the Last Glacial Maximum or mid-Pliocene Warm Period) or the relatively short transient of the Last Millennium, using a wide variety of methods. The forthcoming PMIP6 simulations of the last deglaciation, together with recently published compilations of core data, will provide us a new opportunity to reconstruct the fully global transient evolution of the full climate state over this period with more detail and accuracy than previously achieved. However, this problem is particularly challenging due to the possibility of errors in the timing and magnitude of forcing time series used to drive the models, and in the time scales of model responses. Building on the previous work, we will present some preliminary results which investigate strategies for accounting for these uncertainties.