Introduction
The value of pollen records for vegetation and climate reconstructions is now well-established. Above all, the West Mediterranean region is highly documented, especially for what concerns the Lower Pliocene (Suc et al., 1995a) and offers a favourite Weld for climatic quantifications (Fauquette et al., 1998a, 1998b and in press). The next step is to construct vegetation maps in order to (1) emphasize new problems for a better understanding of past and modern ecosystems, (2) to estimate vegetation changes on a large geographic scale, and (3) to provide data (on the continental carbon mass for example) to climate modellers (Chandler et al., 1994; Sloan et al., 1996). The first aim of this paper is to lay the foundations of such a project for the Mediterranean and Western European regions. Then, a review of vegetation changes throughout the whole Neogene is presented concerning a region rich in data (Southwestern Europe). Most of the pollen localities used in this paper are chronologically well-calibrated according to foraminifers and/or nannoplankton and, for some of them, palaeomagnetism and rodents (for more details, see Bessedik, 1984; Suc, 1989; Suc et al., 1995a).
Vegetation maps
Validation of pollen records
Modern pollen spectra (lacustrine, lagoonal or marine surface samples) are compared to present-day vegetation on a map which also takes into account cultivated surfaces (Fig. 18.1). We now make some comments on the value of pollen data. Warm-temperate deciduous trees are mostly predominant in the Northwestern Mediterranean area and northward. Pines and non-identified Pinaceae (poorly preserved pollen grains) are often over-represented because of their prolific production and the advantage they have in air and water transport.