From the Late Eocene to the Early Miocene (38-20 Ma), Southeastern France (Fig. 1) was a relatively homogenous region where transtensional to extensional basins developed (Séranne et al., 2021; Semmani et al., 2024). These basins were occupied by fluvial and lacustrine environments edged with prominent physiographic units. The latter entities are mainly mountains (Massif Central, Alps and Pyrenees) momentarily surrounded by sea intrusions or by the open sea (Fig. 2). Prior to the late Burdigalian invasion by the Perialpine Sea, the alluvial and lacustrine lowlands represented favourable sedimentary areas for vegetation and climate reconstructions based on pollen floras. The 38-20 Ma period was affected by global climate changes that can be summarized as (1) low amplitude short-term fluctuations within a warm temperature range during the Priabonian, (2) rapid abrupt cooling characterizing the Eocene-Oligocene Transition (EOT), (3) numerous secondary fluctuations of varying intensity during the Rupelian and most of the Chattian, (4) warming in the late Chattian (Fig. 3; Westerhold et al., 2020). The EOT may have started a short time (ca. 200 kyrs) before the Eocene-Oligocene boundary estimated at 33.91 Ma (Fig. 3; Galeotti et al., 2022).

Regional vegetation is recorded by a succession of megafloras showing that the lake shores were subject to a decrease in forest plant diversity to the benefit of open woodlands (e.g., de Saporta, 1891; Tanrattana et al., 2020; Charieau, 2022; Tosal et al., 2023). In contrast, the available pollen floras do not accurately reflect the changes in flora and vegetation between the lowlands and high elevations. Two major problems explain this knowledge gap: the inadequate botanical identification of pollen grains and/or the absence of detailed pollen counts (e.g.: Gruas-Cavagnetto, 1973; Schuler and Sittler, 1976; Châteauneuf, 1977; Châteauneuf and Nury, 1995). The absence of botanical identifications reduces our knowledge of plant diversity, the missing pollen counts make it impossible to quantify the changes in vegetation and, as a consequence, to accurately estimate the climate evolution. The widespread inability of palynologists to distinguish papillate pollen of the former 'Taxodiaceae' from the non-papillate pollen of the other Cupressaceae (Cupressus-Juniperus pollen-type) is another serious handicap.

Several issues therefore remain which palynological studies may help overcome: How high were the mountains in the region? Did marine incursions occur from the Alpine Sea, the North Sea, and/or the Mediterranean Sea? Are evaporites of marine and/or climatic origin? What was the amplitude of climate change during the EOT in terrestrial areas? In this study, we selected and compiled new pollen sites in Southeastern France ranging in age from the Priabonian to the Chattian (Fig. 1) in order to (1) decipher the nature of palaeoenvironmental changes between 38 and 20 Ma, including the EOT, and (2) quantify the amplitude of such climate changes. The new pollen floras were completed by previously studied Aquitanian sites (