FAMOUS transient climate simulations around the Pliocene Time Slice

In order to improve the potential for proxy-data and model comparisons to robustly identify areas of data model agreement and discord, a ';time slice' within the Late Pliocene has been chosen for high resolution data-based reconstruction (Haywood et al., 2013). This time slice is centered on the warm interglacial peak (MIS KM5c; 3.205 Ma) and represents a point within the Pliocene where there are minimum differences between the modern and Pliocene orbital configuration. Although, ideally the reconstruction of proxy-records will capture the peak of the MIS KM5c isotope event, given the pioneering nature of this effort, it is likely that data will span a larger time frame. Therefore, it is imperative that we understand the variations in climate around the Pliocene time slice. Here, we present the first transient simulations of the initial Pliocene time slice centered on the warm interglacial peak and spanning a period of 40 kyr. We use the full-complexity intermediate-resolution Famous climate model driven by PRISM3 boundary conditions (PlioMIP phase 2) and run experiments with changing orbit, with and without dynamic vegetation. In order to provide context for the simulations, model results are point-wise assessed against PRISM3 SST data focusing on the North Atlantic, a region where with conventional snap-shot (fixed boundary condition) GCM type experiments the proxy-model disparity is greatest. PRISM3 is currently the best available global reconstruction of Pliocene SSTs; however, PRISM3 spans the ';time slab' from 3.0 to 3.3 Ma. We distinguish geographic regions that consistently present the greatest differences between fixed-boundary condition and transient simulations. Quasi-periodic millennial-scale climatic oscillations throughout the southern ocean and regions of the North Atlantic have also been identified. Whilst this oscillatory behavior goes someway in reconciling proxy-model differences, the realism of this mechanism is questioned (i.e. is it a model artifact?).