FAMOUS transient climate simulations compared to PRISM4 SST time-series between 3.3 and 3.2 Ma

Wednesday, 17 December 2014: 8:15 AM
Stephen J Hunter1, Harry J Dowsett2, Marci M Robinson3, Aisling M Dolan1, Alan M Haywood1 and Ulrich Salzmann4, (1)University of Leeds, School of Earth and Environment, Leeds, United Kingdom, (2)USGS, Baltimore, MD, United States, (3)USGS, Reston, VA, United States, (4)Northumbria University, Department of Geography, Newcastle-Upon-Tyne, United Kingdom
The Pliocene Model Intercomparison Project (PlioMIP1) identified the need to reduce uncertainty in both model boundary conditions and proxy based paleoenvironmental estimates. In addition, a number of regions and processes were identified where model-data discord was highest (e.g. tropical and mid-to-high latitude North Atlantic sea surface temperatures, and the model representation of polar amplification in general). One potential reason for data-model discrepancies may come from the inability of fixed boundary condition, equilibrium type climate model simulations to predict the full potential of climate variability expressed in the proxy records (e.g. due to changing orbital forcing). Therefore, transient models will be used to understand temporal and spatial variability between Marine Isotope Stage (MIS) M2 and KM5 (~3.3 Ma to 3.2 Ma), a key interval within the Pliocene. The Pliocene Research, Interpretation and Synoptic Mapping Project (PRISM4) has begun generation of high resolution multiple proxy time series to be used to verify simulated conditions in selected areas between MIS M2 and MIS KM5, including the PlioMIP2 KM5C time slice (3.205 Ma).

We use the full-complexity intermediate-resolution FAMOUS climate model driven by PRISM3 boundary conditions and run experiments with time varying orbital forcing, but without dynamic vegetation initially. We perform a preliminary data-model comparison by point-wise assessing the FAMOUS simulations against the new PRISM4 SST data focusing on the North Atlantic, a region where PlioMIP1 GCM simulations and paleoenvironmental estimates exhibited the greatest amount of discord.