Abstract:
Pre-Holocene paleoclimate reconstructions face a correlation conundrum because complications inherent in the stratigraphic record impede the development of synchronous reconstruction. The Pliocene Research, Interpretation and Synoptic Mapping (PRISM) paleoenvironmental reconstructions have carefully balanced temporal resolution and paleoclimate proxy data to achieve a useful and reliable product and are the most comprehensive pre-Pleistocene data sets available for analysis of warmer-than-present climate and for climate modeling experiments. This paper documents the stratigraphic framework for the mid-Pliocene sea surface temperature (SST) reconstruction of the North Atlantic and explores the relationship between stratigraphic/temporal resolution and various paleoceanographic estimates of SST. The magnetobiostratigraphic framework for the PRISM North Atlantic region is constructed from planktic foraminifer, calcareous nannofossil and paleomagnetic reversal events recorded in deep-sea cores and calibrated to age. Planktic foraminifer census data from multiple samples within the mid-Pliocene yield multiple SST estimates for each site. Extracting a single SST value at each site from multiple estimates, given the limitations of the material and stratigraphic resolution, is problematic but necessary for climate model experiments. The PRISM reconstruction, unprecedented in its integration of many different types of data at a focused stratigraphic interval, utilizes a time slab approach and is based on warm peak average temperatures. A greater understanding of the dynamics of the climate system and significant advances in models now mandate more precise, globally distributed yet temporally synchronous SST estimates than are available through averaging techniques. Regardless of the precision used to correlate between sequences within the mid-Pliocene, a truly synoptic reconstruction in the temporal sense is unlikely. SST estimates from multiple proxies promise to further refine paleoclimate reconstructions but must consider the complications associated with each method, what each proxy actually records, and how these different proxies compare in time-averaged samples.