The middle Pliocene has been described as the last period of geological time that displayed greater sustained warmth and reduced climatic variability over a time period greater than any Quaternary interglacial (Dowsett and Poore 1991; Dowsett et al. 1996). Geological proxy evidence supporting this assertion includes sea surface temperatures (SST) reconstructed from planktonic foraminifera (Dowsett and Poore 1991; Dowsett et al. 1992; Dowsett et al. 1996; Raymo 1994), ostracods (Cronin 1991; Forester 1991), siliceous microfossil records (Morley and Dworestzky, 1991; Barron 1992, 1996), terrestrial vegetation records (Thompson 1991; Heusser and Morley 1996; Thompson and Fleming 1996) and numerous records of higher-than-present sea levels (Dowsett and Cronin 1990; Brigham-Grette 1994; Zubakov and Borzenkova 1990). As a result, the middle Pliocene represents a unique time period in which to examine the behaviour of the Earth’s climate system under a regime of greater-than-present warmth.

A relatively precise stratigraphy, wide geographical distribution of sample sites, and numerous extant climatic boundary conditions all combine to make the middle Pliocene an ideal time period for palaeoclimatic modelling studies using general circulation models (GCM; Dowsett et al. 1999). With anthropogenic greenhouse gas emissions forcing greater warmth (IPCC 1995), significant potential benefit exists in the careful investigation of past warm periods in Earth’s history, such as the middle Pliocene. Furthermore, palaeoclimatic modelling investigations of this interval provide us with an opportunity to assess the validity and robustness of predictions produced by GCM that are being used to model past, present, and future climates alike (Chandler et al. 1994; Sloan et al. 1996; Haywood et al. 1999, 2000a, 2000b).

This paper describes the results of a new palaeoclimatic modelling study, focused on the USA, for the middle Pliocene using a state-of-the-art GCM. Previous GCM studies of the middle Pliocene have been carried out by Chandler et al. (1994), Sloan et al. (1996), and Haywood et al. (1999, 2000a, 2000b). However, this is the first to examine, in detail, the nature and dynamics of the palaeoclimate and the validity of model predictions on a regional scale.

The geological record of the USA during the middle Pliocene is diverse and well suited for the regional assessment of GCM predictions (Figure 1). A wealth of geological data are available, derived from tectonic basins that contain long sedimentary records of palaeoenvironmental change. In addition, numerous localities display evidence for Pliocene lacustrine, fluvial, and shallow marine environments (Thompson 1991; Thompson and Fleming 1996).