bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Messinian vegetation and climate of the intermontane Florina-Ptolemais-Servia Basin, NW Greece inferred from palaeobotanical data: How well do plant fossils reflect past environments? Johannes M. Bouchal1*, Tuncay H. Güner2, Dimitrios Velitzelos3, Evangelos Velitzelos3, Thomas Denk1 1 Swedish Museum of Natural History, Department of Palaeobiology, Box 50007, 10405 Stockholm, Sweden 2 Faculty of Forestry, Department of Forest Botany, Istanbul University Cerrahpaşa, Istanbul, Turkey 3 National and Kapodistrian University of Athens, Faculty of Geology and Geoenvironment, Section of Historical Geology and Palaeontology, Greece bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 The late Miocene is marked by pronounced environmental changes and the appearance of strong temperature and precipitation seasonality. Although environmental heterogeneity is to be expected during this time, it is challenging to reconstruct palaeoenvironments using plant fossils. We investigated leaves and dispersed spores/pollen from 6.4–6 Ma strata in the intermontane Florina-Ptolemais-Servia Basin (FPS) of NW Greece. To assess how well plant fossils reflect the actual vegetation of the FPS, we assigned fossil-taxa to biomes providing a measure for environmental heterogeneity. Additionally, the palynological assemblage was compared to pollen spectra from modern lake sediments to assess biases in spore/pollen representation in the pollen record. We found a close match of the Vegora assemblage with modern Fagus–Abies forests of Turkey. Using taxonomic affinities of leaf fossils, we further established close similarities of the Vegora assemblage with modern laurophyllous oak forests of Afghanistan. Finally, using information from sedimentary environment and taphonomy, we distinguished local and distantly growing vegetation types. We then subjected the plant assemblage of Vegora to different methods of climate reconstruction and discussed their potentials and limitations. Leaf and spore/pollen records allow accurate reconstructions of palaeoenvironments in the FPS, whereas extra-regional vegetation from coastal lowlands is likely not captured. Keywords Biome reconstruction, proxy biases, climate reconstruction, plant macrofossils, dispersed pollen, light and scanning electron microscopy 1. Introduction The late Miocene (11.6–5.3 Ma) marks the time in the Neogene (23–2.58 Ma) with the largest shift from equable climate to strong latitudinal temperature gradients in both hemispheres (Herbert et al., 2016). This is well illustrated by the global rise of C4 dominated ecosystems (grasslands and savannahs in the tropics and subtropics; Cerling et al., 1997). In the Mediterranean region, vegetation changes did not happen synchronously with modern steppe and Mediterranean sclerophyllous woodlands replacing humid temperate forest vegetation at different times and places during the middle and late Miocene (Suc, 1984; Kovar-Eder, 2003; Fauquette et al., 2006; Kovar-Eder et al., 2014; Bouchal et al., 2018; Denk et al., 2018; Suc et al., 2018). During the latest Miocene (5.9–5.3 Ma) the desiccation of the Mediterranean Sea was caused by the isolation of the Mediterranean Sea from the Atlantic Ocean. Based on palynological studies across the Mediterranean region, this event did not have a strong effect on the existing vegetation. Open and dry environments existed in southern parts before, during, and after this so-called Messinian Salinity Crisis (MSC). In contrast, forested vegetation occurred in northern parts of Spain, Italy, and the Black Sea (Fauquette et al., 2006). Likewise, a vegetation gradient occurred from N and C Italy and Greece to Turkey, where humid temperate forests had disappeared by the early late Miocene (Denk et al., 2018). The Florina-Ptolemais-Servia Basin (FPS) of NW Greece is one of the best-understood intermontane basins of late Miocene age in the entire Mediterranean region. A great number of studies investigated the tectonic evolution, depositional history, and temporal constraints of basin fills (e.g. Steenbrink et al. 1999, 2000, 2006), plant fossils (e.g. Knobloch & Velitzelos, 1986a,b; Velitzelos & Gregor, 1985; Mai & Velitzelos, 1992; Kvaček et al., 2002; Velitzelos et al., 2014), and vertebrate fossils (van de Weerd, 1979; Koufos 1982, 2006; Koufos et al., 1991). 2 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 The Messinian flora of Vegora in the northern part of the FPS is dated at 6.4–6 Ma and represents the vegetation in this region just before the onset of the Messinian salinity crisis (the pre-evaporitic Messinian). This flora has been investigated since 1969 (Kvaček et al., 2002) and represents one of the richest late Miocene leaf floras in the eastern Mediterranean along with two other, slightly older, Messinian plant assemblages from the FPS, Likoudi/Drimos and Prosilio/Lava (4-6 in Fig. 1; Velitzelos et al., 2014). The focus of previous palaeobotanical studies in the FPS has been on macrofossils. In contrast, no comprehensive study of dispersed pollen and spores has been carried out in the FPS. While fruit and seed floras, to a great extent, and leaf floras, to a lesser extent, reflect local vegetation in an area, dispersed pollen and spores provide additional information about the regional vegetation. Therefore, a main focus of the present study is on spores and pollen of the Messinian plant assemblage of Vegora. We (i) investigated dispersed spores and pollen using a combined light and scanning electron microscopy approach (Daghlian, 1982; Zetter, 1989; Halbritter et al., 2018) that allows a more accurate determination of pollen and hence higher taxonomic resolution. We (ii) then compiled a complete list of plant taxa recorded for the site of Vegora including fruits and seeds, foliage, and spores and pollen. Based on the ecological properties of their modern analogue taxa, we assigned the fossil-taxa to functional types (vegetation units) and inferred palaeoenvironments of the FPS during the Messinian. Using leaf physiognomic characteristics, we (iii) conducted a Climate Leaf Analysis Multivariate Program (CLAMP) analysis (Spicer, 2008; Yang et al., 2011) to infer several climate parameters for the late Miocene of the FPS. We also (iv) used a modified “Co-existence approach” (Mosbrugger & Utescher, 1997) based on climatic requirements of modern analogue plant taxa to infer two climate parameters and (v) a Köppen signature analysis (Denk et al., 2013; Bouchal et al., 2018) based on the Köppen-Geiger climate types in which modern analogue taxa of the fossiltaxa occur. Finally, we (vi) discuss how well the translation of fossil plant assemblages into functional types (vegetation units, biomes) works for reconstructing past environments at local and regional scales. 2. Material and Methods 2.1. Geological setting The old open-pit lignite quarry of Vegora is located in western Macedonia, NW Greece, ca. 2 km E of the town of Amyntaio and is part of the Neogene Florina-Ptolemais-Servia intermontane basin (FPS). The FPS is part of the Pelagonian basin that extends to the north into North Macedonia (Fig. 1). The NNW–SSE trending FPS is ca. 120 km long and presently at elevations between 400 and 700 m a.s.l. and is flanked by mountain ranges to the east and the west. Main ranges include Baba Planina (2,601 m), Verno (2,128 m), and Askio (2,111 m) to the east of the basin and Voras (2,528 m), Vermio (2,065 m), Olympus (2,917 m) to the west (Fig. 1). These ranges are mainly comprised of Mesozoic limestones, Upper Carboniferous granites and Paleozoic schists. Continuous sedimentation since 8 Ma resulted in the accumulation of ca. 600 m of late Miocene to early Pleistocene lake sediments with intercalated lignites and alluvial deposits. The FPS basin formed in the late Miocene as a result of NE–SW extension in the Pelagonian Zone, the westernmost zone of the Internal Hellenides (Brunn, 1956; Pavlides & Mountrakis, 1987). A subsequent Pleistocene episode of NW–SE extension caused the fragmentation of the basin into several sub-basins (Steenbrink et al., 2006). 3 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 Basin fills overlay unconformably Paleozoic and Mesozoic rocks. Alpine and pre-Alpine basement of the area consists of Pelagonian metamorphic rocks (gneisses, amphibolites, mica schists, meta-granites and Permian to Triassic meta-sediments) and crystalline limestone of Triassic–Jurassic age (carbonate cover). Subpelagonian ophiolites and deep-sea sediments of Jurassic age, comprising the Vourinos ophiolitic complex, thrust over the Pelagonian carbonate rocks and are covered by Cretaceous strata (Brunn, 1956; Mavridis et al., 1977; Mountrakis, 1986; Roberts et al., 1988). The Vegora section belongs to the ca. 300 m thick Komnina Formation, which unconformably overlies pre-Neogene basement and is predominantly composed of alluvial sands and conglomerates, lacustrine (diatomaceous) marls and palustrine clays, with some intercalated (xylite-type) lignite seams (Steenbrink et al., 2006). The detailed description of the sequence at the Vegora quarry follows Kvaček et al. (2002) and Steenbrink et al. (2006) (Fig. 2). Since 2000, the lower part of the sequence was not accessible and the exposed sequence started with a ca. 10 m thick lignite seam (see fig. 3 in Kvaček et al., 2002 versus fig. 2, unit 1, in Steenbrink et al., 2006). The full Vegora section begins with hard marls (>15 m) followed by 10–15 m of clay sands and a white marl layer of 10 m. Then, a formation of clay sands follows with a total thickness of 15–20 m. This formation starts with lignitic marls followed by marls and clay sand intercalations. The sand is rich in mica. Above this, a lignite seam occurs with a thickness of 12–35 m. Within the seam, three xylitic layers with a total thickness of 10–12 m can be distinguished. The lower xylitic layer is about 3–4 m, the middle 0.5–1.5 m and the upper is 4–9 m in thickness (Kvaček et al., 2002). The upper lignite layer was the first visible layer in 2002 in the section (unit 1 in Steenbrink et al., 2006). Between the xylitic layers, sand layers of various thicknesses (0 to 15 m) occur. In general, the thickness of these sand layers is smaller towards the N and NE ends of the mine and becomes significantly larger towards the W and SW ends of the mine. The top of the upper xylitic layer is covered by 3–4 m thick shales, followed by a thick layer of light blue marls, 10–60 m thick (unit 2 in Steenbrink et al., 2006), and a layer of sandy marls, 10–40 m thick (units 3 and 4 in Steenbrink et al., 2006). Unit 3 is made up of greybrown lignitic clay at its base and multi-coloured mottled clays, silts and fine sands higher up. Unit 4 consists of cross-bedded conglomerates and coarse sands at the base overlain by mottled clays, silts and fine sands with calcareous nodules. Finally, the top of the section is made up of dark red, mottled sands, silts and clays (unit 5). Unit 2 of Steenbrink et al. (2006) corresponds to the main fossiliferous layers for plant fossils and diatoms (Gersonde & Velitzelos, 1978). The uppermost layer of the Neogene sediments in the area is a formation of marly limestones, of different thickness, which is not everywhere visible appearing only at the nearby villages of Neapoli and Lakia. All Neogene sediments of the area are inclined by a 10º slope towards NNW. The rocks on the top of the Neogene section are Quaternary alluvial deposits, conglomerates, sands, and gravels. This material, in general, has been supplied from erosion processes of the nearby metamorphic mountains. 2.2. Age In the upper part of unit 2, 4 cm-thick layers of tephra rich in biotite were found and used for 40 Ar/39Ar dating (Steenbrink et. al., 2006). The calculated age of 5.97 ±0.07 Ma corresponds to Messinian. Also, using palaeomagnetic data, the base of unit 2, just above the lignite seam, 4 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 can be correlated to the astronomical polarity timescale, indicating that its position corresponds to the end of chron C3An.2n. This would suggest an age of ca. 6.4 Ma for the beginning of unit 3. Therefore, the period of deposition of the light blue marls (unit 2) from Vegora can be narrowed down to ca. 400 ka. 2.3. Sample processing The palynological sample was taken from a slab piece (S115992) from a leaf layer in unit 2 of the Vegora mine. The sample was processed following standard protocols (20% HCl to dissolve carbonate, 40% HF to dissolve silica, 20% HCl to dissolve fluorspar; chlorination, acetolysis; see Halbritter et al., 2018) and the residue was transferred to glycerol. 2.4. Palynological investigation Light microscopy (LM) micrographs were taken with an Olympus BX51 microscope (Swedish Museum of Natural History [NRM], Stockholm, Sweden) equipped with an Olympus DP71 camera. The same grains were examined using LM and scanning electron microscopy (SEM; single grain method; Zetter, 1989; Halbritter et al, 2018). Specimens were sputter-coated with gold for SEM investigation. SEM micrographs were taken with an ESEM FEI Quanta FEG 650 SEM (Stockholm University). Residue and SEM stubs are stored at NRM under specimen numbers S11599201–S11599220. Terminology of palynomorphs follows Punt et al. (2007) and Halbritter et al. (2018). Size categories follow Halbritter et al. (2018). Palynomorphs were determined to family, genus or infrageneric level. In cases when no taxonomic affinity could be established, we used fossil form taxa which are not implying a particular systematic affiliation. The systematic palaeobotany section starts with algae, fern and fern allies, gymnosperms and is followed by angiosperms. Angiosperm classification and author names of orders and families follow APG IV (2016). 2.5. Inferring palaeoclimate estimates We employed three different (semi)quantitative methods to infer a range of climate parameters for the Messinian of NW Greece. CLAMP (climate leaf analysis multivariate program) is a physiognomy based, taxon free method of climate inference and makes use of the relationship between leaf architecture and climate. CLAMP uses calibration datasets of modern vegetation sites across the world to place a fossil leaf assemblage in physiognomic space, which then can be translated into numeric values for several climate parameters (Spicer, 2008; Yang et al., 2011). The Coexistence Approach (CA; Utescher et al., 2014) is a method of inferring palaeoclimate based on nearest living relatives (NLR) of fossil taxa. CA assumes that for a given climate parameter, the tolerances of all or nearly all taxa in a fossil assemblage will overlap to some degree; this overlap is called the climatic coexistence interval. Following best practices in applying the CA Utescher et al. (2014) provided several guidelines to apply the CA in a meaningful way. Among these guidelines, one is to exclude relict taxa (usually monotypic or comprising very few extant species) from the analysis, because of their likely unrepresentative modern distribution. Examples for such taxa are the East and SE Asian Craigia and Glyptostrobus. These taxa had a much wider distribution during parts of the Cenozoic including Arctic regions. For example, Budantsev & Golovneva (2009) described Craigia from the Eocene Renardodden Formation of Spitsbergen for which they inferred a mean annual temperature (MAT) of 8.4 °C and a coldest month mean temperature (CMMT) of –1 °C. In contrast, the two modern species occur in climates with MAT 13.2–21 °C and 5 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 CMMT 6.3–14.2 °C (Fang et al., 2009). Hence, it is assumed that for relict plants ecological niches may have changed considerably during the Cenozoic. For further assumptions of the CA and their critique, see Grimm & Potts (2016) and Grimm et al. (2016). Climate parameters for the NRL are given in the electronic supplementary material Table S1. Köppen signatures (Denk et al., 2015; Bouchal et al., 2018) is another approach to infer largescale climatic patterns for the Cenozoic that is based on NLR of fossil taxa. Modern distribution ranges are mapped on Köppen-Geiger climate maps (Kottek et al., 2006; Peel et al., 2007; Rubel et al, 2017; Global_1986-2010_KG_5m.kmz; definition of the used KöppenGeiger categories are given in the electronic supplementary material S6) and the Köppen climate types in which the modern taxa occur are taken as a proxy for the climate space in which the fossil taxa occurred. It is explicitly stated that climate niche evolution will negatively impact the reliability of the inferred palaeoclimate. To overcome this drawback, subgenera, sections, and genera are used as NLR, whereas single species are usually not considered for NLR. The representation of different climate types is first scored for each species within a genus as present (1)/absent (0) (electronic supplementary material Table S2). To summarize preferences for climate types of all modern analogues, an implicit weighting scheme is used to discriminate between modern analogues that are highly climatically constrained and those that occur in many climate zones. For each modern species, the sum of its Köppen signature is always 1. For example, if a species is present in two Köppen–Geiger climate types, Cfa and Cfb, both score 0.5. If a species is present in 10 Köppen–Geiger climate types, each of these climate types scores 0.1. The Köppen signature of a genus or section, the preferred NLR of a fossil taxon, is the sum of its species’ Köppen signatures for each climate type divided by the total number of scored species for this genus. By this, the percentage representation of each Köppen–Geiger climate type is determined for a genus/section (Bouchal et al., 2018). For pollen taxa of herbaceous and a few woody angiosperm groups that are resolved to family-level only, the distributions of extant members of the family were combined into a general family distribution range and the corresponding Köppen–Geiger climate types determined. 2.6. Characterisation of terrestrial biomes For convenience, we use the biome classification of Woodward et al. (2004) that recognizes five major tree biomes based on the physiognomy of the dominant species: Needleleaf evergreen (NLE), needleleaf deciduous (NLD), broadleaf evergreen (BLE), broadleaf cold deciduous and broadleaf drought deciduous (BLDcold, BLDdrought). These authors also observed that broadleaf drought deciduous vegetation grades substantially into broadleaf evergreen vegetation. Besides, shrublands are defined as lands with woody vegetation less than 2 m tall. Savannahs are defined as lands with herbaceous or other understorey systems, where woody savannahs have forest canopy cover between 30 to 60 %, and savannah has forest canopy cover between 10 and 30 % (Woodward et al., 2004). This very broad definition of savannah may be strongly oversimplified. Thus, for savannah-like vegetation, we make a distinction between steppe and forest-steppe of temperate regions with a continuous layer of C3 grasses and savannah and woody savannah of tropical regions with a continuous layer of C4 grasses (Ratnam et al., 2011). 3. Results 3.1. Pollen and spores: diversity and environmental signal We determined more than 50 palynomorph taxa from a leaf layer in unit 2 of the Vegora 6 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 section (Figs 3–5). A comprehensive taxonomic account including pollen morphological descriptions is provided in the electronic supplementary material S3. The fossil-taxa comprise two algae, five ferns, twelve herbaceous plants, one woody liana, and more than 30 woody trees and shrubs (Table 1). Besides the taxonomic evaluation, 430 palynomorphs were counted to assess the abundance of different taxonomic groups, life forms, and pollination syndromes. Roughly, half of the pollen taxa were present in very small amounts (1–3 grains in the counted sample or < 1%; electronic supplementary material Table S4). The presence and abundance of Spirogyra zygospores/ aplanospores indicate a lake with shallow lake margins (reed belt with Typha) and stagnant, oxygen-rich, open freshwaters. Spores of Osmunda (> 4%) and Leavigatosporites haardti (> 1.6%) are of moderate abundance suggesting that the producing pteridophytes grew close to the sedimentation area, the Messinian Vegora Lake. Among conifers and wind-pollinated trees, strong pollen producers such as Pinus (subgenus Strobus 11%, subgenus Pinus ~8%), Abies (8.4%), Cathaya (7%) and the Betulaceae Alnus (9.3%) and the Fagaceae Fagus (7%) are most abundant. Another group of wind-pollinated woody trees and shrubs was represented with abundances between 2% and 5%. Among these were both deciduous and evergreen oaks (Quercus) and conifers such as Cedrus and Tsuga and undifferentiated papillate Cupressaceae. Among the taxa that are represented by single or few pollen grains, a significant number belonged to insect-pollinated plants. Insect-pollinated trees, shrubs and lianas include Craigia, Platycarya, Castaneoideae and Parthenocissus; Hedera and Sassafras are further insect-pollinated woody taxa, recorded in the leaf fossil record. Herbaceous taxa comprise Apiaceae, Caryophyllaceae, Geranium, Succisa, and Cichorioideae. The ratio arboreal pollen (AP) to non-arboreal pollen (NAP) is 89.5% to 10.5% indicating a forest dominated (tree prevalent) local and regional vegetation according to the threshold values of Favre et al. (2008). Forest types (biomes of Woodward et al., 2004) represented by the pollen assemblage are needleleaf evergreen and deciduous forests (NLE, NLD), broadleaf deciduous forests (BLD), broadleaf evergreen forests (BLE) and mixed forests (MIXED). In addition, BLD and NLD either thrived on well-drained soils or in temporally or permanently inundated areas. A few taxa might also indicate the presence of closed or open shrublands and grasslands (herbaceous taxa including sparse Poaceae with affinity to Poa/Lolium, Chenopodiaceae, Apiaceae etc. and woody taxa including palms; see Table 1 for other woody taxa). These may have been associated with BLE woodlands (Q. mediterranea, Q. sosnowskyi) or with mesophytic evergreen forests of Q. drymeja (see below). Alternatively, they may have originated from an independent vegetation type (for example montane grasslands). Among needleleaf forest biomes, for some taxa, the attribution to a distinct forest type is not straightforward. For example, conifers such as Cathaya may have been part of the montane hinterland vegetation on well-drained soils but may also have been important elements of peat-forming vegetation (Schneider, 1992; Dolezych & Schneider, 2006, 2007, 2012). Based on pollen abundances (electronic supplementary material Table S4) local (close to the lake), regional (occurring in the FPS) and extra-regional (potentially occurring outside the FPS) vegetation can be inferred. Regional vegetation consisted of BLD forests subjected to flooding (Alnus) and NLD swamp forests (papillate Cupressaceae). Close to the lake, a mixed forest with Fagus, Abies and Cathaya thrived (using the modern Abant Gölü of northern Turkey as a reference for pollen rain vegetation relationships; van Zeist & Bottema, 1991). Deciduous oaks (mainly of sect. Cerris) also might have been part of local forest vegetation (BLD). Pinus and Cedrus NLE forests and evergreen oak forests (BLE) grew at some 7 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 distance from the lake (regional vegetation; using threshold abundances of Cedrus, 7%, and evergreen Quercus, 20%, as indicators of local source vegetation; Bell & Fletcher, 2016). 3.2. Leaf fossils, fruits, and seeds: diversity and environmental signal Leaf and fruit remains from the Vegora mine have been collected and described since half a century (Schneider & Gregor, 1973; Velitzelos & Gregor, 1985, 1990; Knobloch & Velitzelos, 1987, Mai & Velitzelos, 1997). The most recent reviews are those of Kvaček et al. (2002) and Velitzelos et al. (2014). Table 1 provides an updated taxon list. Fruits and seeds recovered from the lignite seam of the Vegora section represent aquatic and reed vegetation. From the lignite seam also trunks of tall trees (as Sequoioxylon) in upright position were recovered. The plant assemblage of the blue marls represents needleleaf evergreen and deciduous, and broadleaf evergreen and deciduous, as well as mixed forests. Among needleleaf forests, swamp forests are typically represented by Taxodioideae, while Sequoioideae, Keteleeria, Pinus and others may have been part of peat-forming vegetation (forests types NLE, NLD). Based on the great abundance of Alnus leaves, a local alder swamp may also be inferred (BLD). Fagus is among the most abundant taxa based on the number of recovered leaf remains suggesting that it was part of the mesic forest vegetation close to the lake. Also, deciduous foliage of Quercus sect. Cerris (Q. kubinyi, possibly Q. gigas) might have grown in the vicinity of the lake, either forming mixed stands with Fagus or oakdominated forests. Kvaček et al. (2002) referred to this vegetation as Fagetum gussonii/Quercetum mixtum. Evergreen oaks are abundant in the Vegora leaf assemblage but fairly rare in the pollen record (electronic supplementary material Table S4). This indicates that the leathery leaves of these taxa were transported to the area of sedimentation by slow-flowing streams and that the source vegetation was further away from the lake. Kvaček et al. (2002) referred to these evergreen forests as sclerophyllous (Quercetum mediterraneum). Denk et al. (2017) distinguished between sclerophyllous Mediterranean oak forest and laurophyllous Q. floribunda forest from Afghanistan which is a better analogue for the widespread western Eurasian fossil-taxon Q. drymeja. Hence we infer an ecological cline from mesic evergreen oak forests to sclerophyllous forest and shrublands in the Messinian of Vegora (cf. Freitag, 1982). Well-drained forests dominated by needleleaf taxa occurred in the montane vegetation belt (Abies) and on rocky substrates (Cedrus, Pinus). Only a few taxa are potentially representing open shrubland vegetation (Acer spp., Chamaerops). 3.3. Inferring past climate with CLAMP Forty-one dicot leaf morphotypes were scored for the CLAMP analysis. Given the distinctly temperate appearance of this flora, we used the calibration dataset Physg3arcAZ_GRIDMet3arAZ. Physg3arcAZ includes 173 sites, among them the 144 Physg3brcAZ sites plus 29 sites corresponding to the alpine nest (Wolfe, 1993). The alpine are the coldest sites known to have a different physiognomic behaviour (Wolfe, 1993); they are characterised by a WMMT lower than 16 °C and a CMMT lower than 3 °C (Wolfe, 1993). The reconstructed climate parameters are MAT 10–13.5 °C, WMMT 19.2–22.8 °C, CMMT 1–5 °C, GROWSEAS 6–8 months, GSP 700–1100 mm, MMGSP 110–160 mm, Three_WET 500–780 mm, Three_DRY 180–260 mm, and Three_WET to Three_DRY ratio < 4. In terms of the Köppen-Geiger climate classification, this translates into a temperate Cfb climate (Tcold > 0 and < 18 °C; without a dry season; warm summer Thot < 22 °C). 8 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 In addition, we used the calibration dataset PhysgAsia1_HiResGRIDMetAsia1 that adds 45 sites from China to the Physg3brcAZ dataset. Using this dataset, the reconstructed climate parameters are generally cooler and drier than the ones obtained from Physg3arcAZ. MAT 8.7–11.5 °C, WMMT 19–22.6 °C, CMMT -2.7–2.3 °C, GROWSEAS 5.5–7 months, GSP 460–1100 mm, MMGSP 100–160 mm, Three_WET 400–750 mm, Three_DRY 80–220 mm, and Three_WET to Three_DRY ratio < 5.5. In terms of the Köppen-Geiger climate classification, this translates into a temperate Cfb to cold Dfb climate (Tcold > 0 and < 18 °C versus Tcold < 0 °C; both without a dry season and warm summer Thot < 22 °C). Score sheets and full documentation of the CLAMP analyses are provided in electronic supplementary material S5. 3.4. Inferring past climate with CA Using the coexistence approach, we estimated coldest month mean temperature (CMMT) and mean annual temperature (MAT) coexistence intervals to see how CA behaves including and excluding relictual and monotypic taxa. Following Utescher et al. (2014) relict taxa with very limited modern distribution were excluded from the analysis. Excluded taxa are plotted to the left of the diagram in Fig. 6. For the monotypic genus Chamaerops, the tribus Trachycarpeae was used as NLR. For CMMT a lower boundary value of 1.2 °C is estimated based on the cold tolerance of Trachycarpeae. Chamaerops has a slightly warmer CMMT of 4 °C. For mean annual temperature (MAT) the lower boundary is defined by Zelkova (8.6 °C) and the upper boundary by Acer sect. Acer (21.2 °C). When only the 10–90% percentiles were considered, MAT low was defined by Zelkova as 9.9 °C and MAT high by Acer sect. Acer as 18.4 °C (Fig. 6). Inclusion of a priori excluded relict species with a limited distribution would greatly change the estimated climate values. CMMT low would be defined by Craigia (6.5 °C) and CMMT high by Sequoia (7.5 °C). Likewise, MAT low would be defined by the monotypic conifer Cathaya (13.4 °C) and MAT high again by Sequoia (15.3 °C; electronic supplementary material Table S1). Using only the 10–90% percentiles, MAT low would be defined by Craigia (14 °C) and MAT high (14.1 °C) by Sequoia. 3.5. Inferring past climate with Köppen signatures Based on 700 Köppen signatures of modern species (rarely sections and families) genus- to family-specific Köppen signatures were used to generate Köppen signatures for the Vegora assemblage of unit 2. Temperate C climates are by far the most common ones represented by modern analogues of the Vegora plant assemblage. Cfa/b and Cwa/b climates represent 50% of the occurrences of NLR taxa when pollen and spores are considered, and 54% when macrofossils are considered (Fig. 7). Csa/b climates are represented by 11–13%. Snow climates (CMMT < 0°C) are represented by 17% (Df, Dw) and 2–3% (Ds). Thus, C and D climates make up > 80% of all NLR occurrences. In contrast, equatorial climates are represented by 10% (spores and pollen) and 6.5% (macrofossils). Arid B climates are represented by < 10% in the spores/pollen and macrofossil assemblages. 4. Discussion 4.1. How well do plant fossils reflect past environments of the FPS? It has long been known that there is no exact relationship between fossil (and modern) assemblages of dispersed spores and pollen and the actual vegetation (e.g. van Zeist et al., 9 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 1975; Connor et al., 2005; Traverse, 2007; Bell & Fletcher, 2016; Marinova et al., 2018). Marinova et al. (2018) pointed out several problems when inferring vegetation from pollen diagrams. These included (i) pollen production biases which generally result in the overrepresentation of woody species and the under-representation of herbaceous species in the pollen assemblage, (ii) transport of tree pollen into non-forested areas resulting in poor delineation of ecotonal boundaries, (iii) upslope transport of pollen from lowland areas in upland areas resulting in poor delineation of altitudinal vegetation gradients and tree line. Furthermore, these authors found that samples from small basins (<1 km2) are more likely to be reconstructed accurately because they sample an appropriate pollen source area to reflect regional vegetation patterns in relatively heterogeneous landscapes. In contrast, large uncertainties were observed when inferring the local vegetation in large basins, e.g. the Black Sea. Here, large pollen source areas result in strongly mixed signals which do not well discriminate the vegetation belts around a specific site. We note that this caveat may in fact be beneficial when inferring the past vegetation in a larger area. The FPS is a basin that extends ca. 120 km x 30 km and is flanked by high mountains. Hence, a rich pollen assemblage with a strongly mixed signal is expected to reflect the actual vegetation types in the region although it may be challenging to correctly assign particular pollen types to vegetation units. For instance, Ivanov (2001) interpreted a pollen diagram from a Tortonian section in the Sandanski Graben (Bulgaria; 8 in Fig. 1) with a considerable amount of herbaceous pollen (including Poaceae, Chenopodiaceae, and Artemisia making up c. 5 to 20% of the pollen spectrum) to reflect extra-regional open vegetation on an elevated plateau in addition to swamp forests, riparian forests, and mixed mesophytic forests developed in a river valley and adjoined slopes. Here, downslope transportation of pollen from open landscapes blurred the local signal of the pollen record, but at the same time added regional and extra-regional vegetation information. A close relationship between the actual vegetation and the pollen spectrum from recent and Holocene sediment samples has also been reported for NW Turkey (van Zeist & Bottema, 1991). Modern surface-sample spectra accurately depicted the regional vegetation, although some taxa were underrepresented in the pollen spectra while others were overrepresented. For example, pollen spectra of Fagus-Abies dominated areas showed relatively low percentages of these two taxa (10.4% and 7.4%), while high amounts of Pinus (ca. 30%) derived from forests thriving at some distance from the pollen trap. Likewise, comparatively high amounts of Juniperus, Quercus, and Carpinus did not reflect the local vegetation but a regional signal. In combination with a weak herbaceous signal (Poaceae < 5%; Chenopodiaceae, Caryophyllaceae, Apiaceae represented by single pollen grains) the strong arboreal signal provided a fairly accurate picture of the forest communities at a regional scale (van Zeist & Bottema, 1991). In cases of bad pollen preservation (oxidised sedimentary rocks) it should be kept in mind that only pollen with durable exines (high sporopollenin content) will be preserved (e.g. Pinus, Chenopodiaceae/Amaranthaceae; Traverse, 2007) resulting in a biased signal. In contrast to dispersed spores and pollen, macrofossils (leaves) mainly reflect local and regional vegetation whereas extra-regional vegetation is usually not reflected. Leaf remains are mostly scattered isolated carbonised compression fossils, which are not concentrated abundantly (“Blätterton” layers or paper shales) in distinct fossiliferous layers (Kvaček et al., 2002); both small and larger leaves are usually not fragmented and hence there is no indication for long-distance transport in high-energy depositional settings. At the same time, low pollen abundances of evergreen oaks along with abundant leaf fossils representing 10 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 evergreen oaks might indicate that these leaves were transported by slow-flowing streams over relatively large distances. Rarely, large fruit bodies are encountered, mostly represented by conifer cones. Therefore, a combined wind and water transport from habitats bordering the lake can be assumed (Kvaček et al., 2002). Among woody plants, Fagaceae (Fagus, Quercus) are ecologically diverse and niche conserved at the genus/section level. Fagus is exclusively found on well-drained soils and hence was not an element of the swamp forest vegetation. However, under humid equable climates, lowland coastal and deltaic forests may contain Fagus and hardwood hammocks with rich broadleaf deciduous and evergreen forests may be present next to aquatic and hydric vegetation (Ferguson et al., 1999). In contrast, white oaks, sect. Quercus, may thrive in swamp forests, riparian forests, mesic forests of lowlands and uplands, or may form Mediterranean scrub. These different ecologies are well reflected in leaf morphology whereas pollen morphology at the sectional level does not discriminate different species/ecologies (Denk & Grimm, 2009). Since white oaks are represented by pollen only, no further conclusions can be drawn as to their ecologies. Other sections of Quercus (sects Cerris, Ilex) represented in the Vegora assemblage are highly niche conserved and exclusively found on well-drained soils. Based on differences in leaf morphology (leaf size, deciduousness), fossil-taxa such as Quercus gigas (leaf lamina up to 22 cm long; sect. Cerris) might indicate humid temperate conditions on northern slopes (Kvaček et al., 2002), while Quercus kubinyi (Cerris) might have been part of drier slopes. These fossil-species could have been accessory elements in Fagus-dominated or oak-dominated forests (see above, Results section). Section Ilex comprises evergreen species exclusively growing on welldrained soils. Closest relatives of the Messinian taxa are modern Mediterranean species (the fossil-species Q. sosnowskyi resembles the modern species Q. alnifolia, endemic to Cyprus, by leaf shape and leaf epidermal features; Kvaček et al., 2002) and Himalayan/East Asian species (e.g. Q. drymeja resembles the modern Q. floribunda, south of the Himalayas; Denk et al., 2017). Inferring the ecological properties of these fossil-taxa is not straight-forward: Morphologically they either resemble modern E Mediterranean taxa or temperate Himalayan taxa. At the same time, time-calibrated molecular phylogenies suggest that the modern Mediterranean members of sect. Ilex diverged from their Himalayan sister species during late Oligocene to early Miocene times, long before the deposition of the plant assemblage of Vegora (Jiang et al., 2019). Within the Mediterranean clade, the most mesic species Q. ilex also occurs in humid temperate forests of the Euxinian region (N Turkey, W Georgia) and diverged from the remaining species of western Eurasian sect. Ilex no later than 9 Ma (Jiang et al., 2019). Assuming that fully Mediterranean climate conditions, with precipitation minima during the summer, in the Mediterranean region did not establish prior to the early Pliocene (Suc, 1984; Velitzelos et al., 2014) we speculate that the Messinian members of sect. Ilex were chiefly temperate species that went extinct during the Pliocene (cf. Jiang et al., 2019). Specifically, Q. drymeja might have formed a forest belt above the Fagetum gussonii/Quercetum mixtum and below the needleleaf evergreen forest belt above. Other Quercus sect. Ilex such as Q. mediterranea and Q. sosnowskyi may have formed woody shrublands or forests on drier sites (edaphically or due to the aspect of the slope). Concerning the presence of grasslands or open woodlands the palaeobotanical data at hand cannot discriminate between different scenarios. For taxa that are known from the macrofossil record (Chamaerops, evergreen oaks) it is almost certain that they were part of the regional flora of the FPS. The woody genera Olea, Cotinus, and Pistacia, known only from the pollen record of unit 2, are typical elements of the present Mediterranean and submediterranean vegetation belt in S Europe. Bell & Fletcher (2016) found that soil samples in open vegetation 11 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 plots in N Morocco recorded 20% to 35% arboreal pollen (AP). Main contributors to this regional to extra-regional airborne pollen rain were Quercus types and Olea. In our sample, arboreal pollen makes up almost 90% of the total count. Quercus certainly was a major component of local to regional vegetation because it is the most prominent component with several deciduous and evergreen species in the leaf flora of Vegora. In contrast, Olea makes up less than 1% in the pollen count. No leaf and seed remains reminiscent of Olea are recorded from Vegora. This along with the known tendency of/ability for long-distance transport (Bell & Fletcher, 2016) might indicate the presence of Olea at a greater distance from the Vegora basin. The same can be assumed for Cotinus and Pistacia. The latter, however, do also occur in open-canopy pine forests. In case of herbaceous taxa represented by single or very few pollen grains in the palynological record, these may also reflect long-distance dispersal from high mountain or even from more distant coastal areas to the west of the FPS. They would then provide an extra-regional vegetation signal. Potential elements of open vegetation include Apiaceae, Chenopodiaceae, Poaceae, Geranium, Caryophyllaceae, and Cichorioideae. Except for Poaceae and Chenopodiaceae these taxa are predominantly insect-pollinated. For wind-pollinated taxa represented with 1–3 grains in the pollen count (Poaceae, Chenopodiaceae) we assume that this is indicative of a regional or extra-regional source vegetation. The insect-pollinated taxa, also represented by 1–3 grains in the pollen count, are difficult to assign to either local or regional/extra-regional vegetation. If these groups were local elements, they would have been quite rare, based on the low numbers of their pollen. They could have been part of the lakeshore vegetation, of open rocky places, of the understorey of forest vegetation or meadows above the tree line. Alternatively, these elements could have been brought in by long-distance dispersal (LDD) from coastal plains to the southeast and east of the FPS. In sum, the combined macrofossil and microfossil record offers an accurate picture of the different vegetation types present in the FPS during the Messinian. The fossil record suggests that the local and regional vegetation in the FPS comprised a range of ecologically different zonal and azonal forest types, while LDD of several herbaceous taxa may potentially have contributed to an extra-regional pollen signal. 4.2. Inferring Messinian pre-evaporitic vegetation of the FPS and adjacent areas Our multi-proxy palaeobotanical study of the Messinian assemblage of Vegora is based on information from fruits and seeds, leaves, and dispersed pollen and spores. For the main flora in unit 2 (blue marls) we used information from leaf fossils and dispersed spores and pollen. As discussed above, there is strong evidence for the presence of a wide range of forest and forest/shrubland types in the FSP. Furthermore, a small number of woody and herbaceous taxa could reflect open vegetation. The latter are represented by low numbers of pollen grains, which could be ascribed to LDD from remote areas including dry uplands or coastal plains. In order to evaluate the pre-evaporitic Messinian vegetation of the FPS, we compared our finds with previously published data on other plant fossil localities in the FPS and surrounding areas. In addition, a vertebrate locality from the Axios valley (Dytiko1, 2, 3; 3 in Fig. 1) is roughly coeval with the Messinian pre-evaporitic assemblages of the FPS (Koufos, 2006). The hypsodonty index of this fauna of 1.45-1.86 (NOW database, http://pantodon.science.helsinki.fi/now/locality.php?p=ecometrics) corresponds to the diet types “mixed-closed habitats”, “regular browsers”, and “selective browsers” according to Janis (1988) and hence provides an excellent match with the environments inferred for the FPS. 12 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 From Lava (Fig. 1), Steenbrink et al. (2000) investigated two sequences covering two sedimentary cycles each. Based on palaeomagnetic correlation these sequences are dated as c. 6.8–6.7 Ma and c. 6.3 Ma. The pollen assemblages are comparable to the Vegora assemblage but differ in some respects. First, the Lava sections have a continuous high amount of Pinus pollen (20 to >60%) suggesting that the fossil site was located closer to pine forests than was the Vegora lake. Second, Cedrus pollen is abundant with values between 10 and >30%. Third, Steenbrink et al. (2000) did not report evergreen oak pollen, although evergreen Quercus is known from Lava based on leaf fossils (Velitzelos et al., 2014). Steenbrink et al. (2000) inferred a humid temperate climate without dry season for the investigated sedimentary cycles. In addition, they suggested that expansions of Fagus accompanied by a decrease of Abies might reflect subtle increases in montane humidity. Overall, they suggested continuously wet and warm-temperate climate conditions for the investigated period for Lava. Velitzelos et al. (2014) provided revised taxon lists for the roughly coeval macrofossil (leaves and fruits/seeds) localities Prosilio and Lava (age based on palaeomagnetic correlation, 6.7– 6.4 Ma; Steenbrink et al., 2006). The macroflora is very similar to the one from Vegora in terms of composition. However, whereas Quercus sosnowskyi is among the most abundant elements in Vegora, only a few leaves represent this species in Prosilio; also Glyptostrobus is much less abundant. Pinus is represented by cones, leaf fascicles, and leafy branches; this is in accordance with the high amount of pine pollen documented in the palynological record. Fagus is a frequent element as well, while Abies is not recorded in the macroflora. Likoudi, 20 km S of Lava, is located in a small basin south of the main FPS (Fig. 1). The macroflora (leaves, fruits and seeds) is very rich (see revised and updated floral list in Velitzelos et al., 2014). The precise age of the Messinian diatomaceous marls is not clear (Knobloch & Velitzelos, 1986) although it unambiguously is pre-evaporitic. The flora is characterised by the high diversity of conifers (11 genera of Cupressaceae and Pinaceae, including Torreya – as Egeria sp. in Velitzelos et al., 2014). As in Vegora, Fagus is a dominating element. Other taxa (Cercis, Laria, cf. Nerium) are not known from other FBS floras. Well-preserved cones of Cedrus and cones and leafy twigs of Cathaya and Taiwania suggest that these genera were not growing at high elevations but nearby the area of deposition (lake). If coeval with the Lava deposits, this would explain the relatively high amounts of Cedrus pollen in the palynological section of Lava. Ivanov & Slavomirova (2002) investigated a 70 m succession of lacustrine sediments in the Bitola Basin (Northern Macedonia; borehole V-466; 1 in Fig. 1) about 10 km E of Bitola and 40 km NNW of Vegora. Based on a vertebrate fauna on top of these sediments the plantbearing sediments are assigned a late Miocene age (Dumurdžanov et al., 2002; OgnjanovaRumenova, 2005). From these sediments, abundant leaves of Quercus sosnowskyi have been reported (Dumurdžanov et al., 2002). The pollen assemblage is similar to the one from Vegora by its high amounts of woody taxa (both conifers and angiosperms) and the composition and very low abundance of herbaceous taxa. In contrast to the Vegora assemblage, few taxa are represented by high percentages (Pinus 20–40%, Fagus up to 10%, Quercus 10–20%, Taxodioideae and Sequoioideae 10–15%, and Alnus, up to 20%). In the Vegora sample, only Pinus reached more than 10% in the pollen count. For N and C Italy, Kovar-Eder et al. (2006) and Bertini & Martinetto (2008) reported prevalence of deciduous Fagus and Quercus in the northern parts whereas in addition sclerophyllous plants (e.g. Quercus mediterranea) were common in Messinian pre-evaporitic assemblages of C and S Italy (Gabbro I, Senigallia, Palena). 13 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 From the Serres Basin (7 in Fig. 1), Psilovikos & Karistineos (1986), Karistineos & Ioakim (1989) and Suc et al. (2015) reported pollen assemblages that reflect lowland swamp forests (NLD and BLD) and hinterland vegetation including broadleaf deciduous and needleleaf evergreen forests. The age of these localities is younger than Vegora and based on mammal data (Koufos, 2006) and geological data (Suc et al., 2015) might have been deposited during the post-evaporitic Messinian. It is remarkable that post-evaporitic and evaporitic Messinian floras from Italy (Bertini & Martinetto, 2008) and Greece (Velitzelos et al., 2014) mainly reflect moist conditions and persistence of forested environments. This is in strong contrast to conditions from Turkey, where steppe and steppe forest were established by the end of the middle Miocene (Denk et al., 2018). Main vegetation types recognised for the pre-evaporitic Messinian of the FPS and their modern analogues are summarised in Table 3. 4.3. Inferring Messinian pre-evaporitic climate of the FPS and adjacent areas Fruit/seed and leaf records mainly represent azonal vegetation and may not adequately reflect the zonal vegetation growing at some distance from the area of sedimentation. Zonal vegetation is mainly controlled by large-scale climate (regional, extra-regional). Azonal vegetation is controlled by edaphic conditions rather than large-scale climate. For these reasons, seed/fruit and leaf assemblages either used for physiognomic (CLAMP) or taxonomic (Köppen signatures, CA) methods may not be suitable for a meaningful (representative) climate reconstruction or they may reconstruct a local climate. However, as shown by Ferguson et al. (1999), under mild and humid climate conditions, the vegetation close to the area of sedimentation consists of a mosaic of vegetation types, some of which are composed of woody plants that also dominate the hinterland vegetation. For example, natural levee and hardwood hammock vegetation associated with azonal swamp, riparian and bog vegetation in SE North America contains taxa such as Fagus, Magnolia, Ilex spp., Carpinus, and Symplocos. In this case, the limitations outlined above are not valid. In drier, strongly seasonal climate settings, the azonal vegetation will not be sufficient to produce meaningful (representative) regional climate estimates. When the palynological record is taken into consideration, both the local and the regional (hinterland, vertical vegetation belts) vegetation is likely to be captured. This is reflected in the much greater diversity of conifers and herbaceous taxa in the pollen record as compared to the macrofossil record in the Vegora assemblage. In addition, long-distance dispersal may add information from the extra-regional vegetation (e.g. Olea; van Zeist & Bottema, 1991; Bell & Fletcher, 2016). The resulting differences in reconstructed climate/vegetation are seen in the Köppen signatures for the leaf fossil and pollen/spore records of Vegora (Fig. 7). Taxa extending into tropical climates are much better represented in the pollen record than in the leaf record. Therefore, Köppen signatures, although not providing exact values for different climate parameters, offer important qualitative information. This can be illustrated by comparing three floras from the E Mediterranean region, all dominated by Quercus spp. and Fagus and ranging in age from early Burdigalian to Messinian: The early Burdigalian flora of Güvem, Anatolia (MN3, 20–18 Ma; Denk et al., 2017) has a tropical signal of 16% in Köppen signatures from pollen and spore assemblages, whereas tropical signal from leaf fossils is 11.9%. The middle Miocene floras of the Yatağan basin (Langhian/Serravallian, MN6/MN7+8; 14.8–13.8 Ma; Bouchal et al., 2018) show a trend from the more humid MN6 14 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 zone to the more arid MN7+8 zone with tropical signal from pollen and spore assemblages decreasing from 11% in pollen zones 1 and 2 to 7% in the transitional pollen zone 2/3. The tropical signal in Köppen signatures from the macrofossil assemblage from pollen zone 2 is 9%. In the pre-evaporitic assemblages of Vegora, the tropical signal is 9% and 6% for the pollen/spore and leaf assemblages. In contrast, the coexistence approach (CA), in producing exact values for selected climate parameters, will be prone to produce hybrid climates if fossil assemblages represent a range of lowland azonal and hinterland vegetation and vegetation from different vertical belts. Such artificial climates may be randomly expressed as narrow coexistence intervals including different vertical vegetation belts, or in very broad coexistence intervals (cf. Table 2). Thus, Köppen signatures are considered a more dynamic way of climate reconstruction as they account for the possibility that different elements may derive from different vertical vegetation belts. CLAMP (Physg3arcAZ) reconstructed WMMT that corresponds with a temperate climate with warm summers, Cfb Köppen climate type, underscoring the temperate character of the Vegora assemblage. The calibration set Asia 1 reconstructed a cooler climate with CMMT below 0°C. This would appear to be in conflict with the cold tolerance of the palm Chamaerops that is recorded from the leaf record and palms in general as recorded in the pollen record. However, all three approaches to reconstruct the pre-evaporitic Messinian climate of the FPS suggest a temperate climate with weak seasonality (Table 2). In view of the relatively homogeneous vegetation signal from the FPS (both from dispersed spores/pollen and leaves), the inferred climate is highly plausible. 5. Conclusion The present study used palaeobotanical data to reconstruct palaeoenvironments in an intermontane basin of NW Greece shortly before the Messinian salinity crisis. For the period 6.4-6 Ma, leaf fossil data and dispersed spores and pollen indicate the presence of various types of forest including riparian and mesic forests, deciduous and evergreen forests, laurophyllous forests and sclerophyllous woodlands. Open landscapes dominated by herbaceous plants were not reconstructed for the Florina-Ptolemais-Servia Basin. Based on sparse pollen records resulting from potential long-distance dispersal it cannot be ruled out that herbaceous plants played a more important role in coastal lowlands to the W of the study area. However, contemporaneous vertebrate faunas from the Axios valley also suggest mixed forested vegetation. We found that the combined leaf and dispersed spore/pollen records allow fairly accurate reconstruction of local and regional vegetation. Leaf fossils offer more species-diagnostic features than pollen and a combination of leaf taphonomy and pollen frequencies allow discriminating local and regional vegetation. Furthermore, specific comparison with modern pollen spectra was made in order to understand biases in pollen abundances. This provided a transfer function for the interpretation of fossil pollen assemblages. The results of this study confirm previous findings of a N to S gradient of temperature and precipitation seasonality in the Mediterranean area. Our results also reinforce the notion that steppe and forest steppe environments evolved earlier in Turkey with deciduous oaks playing important roles in the woody flora. In contrast, laurophyllous evergreen oak forest persisted in Greece/Italy into the late Miocene and as relict into the Pliocene. Climate reconstructions utilizing three different approaches to climate reconstruction resulted in roughly similar values that translate into a cool temperate Cfb climate according to the Köppen-Geiger climate classification. When comparing CLAMP, 15 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 CA and Köppen signatures, we noticed that leaf assemblages may be biased towards autochthonous plant communities and by this the climate signal will be highly local. However, under humid mild climates plants from the hinterland vegetation may thrive from lowlands to high altitudes and interspersed in riparian landscapes (e.g. on hammocks, gallery forests etc.). Furthermore, the taphonomy of the fossil plant assemblage of Vegora suggests that slow-flowing streams had transported leaves from different vegetation types of the hinterland into the lake and thus the leaf taphocoenoses would be representative of both the local and regional vegetation. This causes a further problem for CLAMP and more so for CA: climate signals in the fossil plant assemblage may be highly mixed and may derive from different vertical vegetation belts. Köppen signatures, while not generating exact values for particular climate parameters, overcome the problem of hybrid climates as they collect climatic/environmental signal in a plant assemblage without averaging different signals into a single, possibly artificial, signal. Data accessibility. All data used in this article are made available in the Supplemental Material. Author’s contributions. T.D. designed the study and wrote the first draft. J.M.B. investigated dispersed pollen and spores and made the Köppen signatures and CA analyses. T.H.G. made the CLAMP analysis. T.D., J.M.B., T.H.G., D.V., and E.V. evaluated the fossil data and wrote the final paper. Competing interests. We declare we have no competing interests. Funding. This study was financed by a grant of the Swedish Research Council (VR; project no. 2015-03986) to T.D. T.H.G. was supported by The Scientific and Technological Research Council of Turkey (TÜB İTAK), 2219 Post-Doctoral Research Fellowship Program (2017/2), project no. 1059B191700382. Supplemental Material. The supplement related to this article is available online at: https://figshare.com/articles/JMB_THG_DV_EV_TD_Supplementary_Material/10327646 References Akhani H, Djamali M, Ghorbanalizadeh A, Ramezani E. 2010 Plant biodiversity of Hyrcanian relict forests, N Iran: an overview of the flora, vegetation, palaeoecology and conservation. Pak. J. Bot. 42, 231–258. Akkemik Ü. 2003 Tree rings of Cedrus libani at the northern boundary of its natural distribution. IAWA J. 24, 63–73. Angiosperm Phylogeny Group (APG). 2016 An updated of the angiosperm phylogeny group classification for the orders and families of flowering plants APG IV. Bot. J. Linn. Soc. 181, 1–20. Bell BA, Fletcher WJ. 2016 Modern surface pollen assemblages from the Middle and High Atlas, Morocco: insights into pollen representation and transport. Grana 55, 286–301. Bertini A, Martinetto E. 2008 Messinian to Zanclean vegetation and climate of Northern and Central Italy. Boll. Soc. Paleontol. Ital. 47, 105–121. Brunn JH. 1956 Contribution à l’étude géologique du Pinde septentrional et d’une partie de la Macédonie. Ann. Géol. Pays Hellèn. 7, 1–358. Bouchal JM, Güner TH, Denk T. 2018 Middle Miocene climate of southwestern Anatolia from multiple botanical proxies. Clim. Past 12, 1427–1440. Budantsev LY, Golovneva LB. 2009 Fossil Flora of Arctic II - Palaeogene Flora of Spitsbergen. St. Petersburg: Komarov Bot. Inst. Russ. Acad. Sci. 16 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 Cerling TE, Harris JM, MacFadden BJ, Leakey MG, Quade J, Eisenmann V, Ehleringer JR. 1997 Global change through the Miocene/Pliocene boundary. Nature 389, 153–158. Connor SE, Thomas I, Kvavadze EV, Arabuli GJ, Avakov GS, Antonio Sagona A. 2004 A survey of modern pollen and vegetation along an altitudinal transect in southern Georgia, Caucasus region. Rev. Palaeobot. Palyn. 129, 229–250. Daghlian CP. 1982 A simple method for combined light, scanning and transmission electron microscope observation of single pollen grains from dispersed pollen samples. Pollen et Spores 24, 537–545. Denk T, Grimm GW. 2009 Significance of pollen characteristics for infrageneric classification and phylogeny in Quercus (Fagaceae). Int. J. Plant Sci. 170, 926–940. Denk T, Velitzelos D. 2002 First evidence of epidermal structure of Ginkgo from the Mediterranean Tertiary. Rev. Palaeobot. Palynol. 120, 1–15. Denk T, Frotzler N, Davitashvili N. 2001 Vegetational patterns and distribution of relict taxa in humid temperate forests and wetlands of Georgia (Transcaucasia). Biol. J. Linn. Soc. 72, 287–332. Denk T, Grimm GW, Grímsson F, Zetter R. 2013 Evidence from “Köppen signatures” of fossil plant assemblages for effective heat transport of Gulf Stream to subarctic North Atlantic during Miocene cooling, Biogeosci. 10, 7927–7942. https://doi.org/10.5194/bg10-7927-2013. Denk T, Güner TH, Kvaček Z, Bouchal MJ. 2017 The early Miocene flora of Güvem (Central Anatolia, Turkey): a window into early Neogene vegetation and environments in the Eastern Mediterranean. Acta Palaeobot. 57, 237–338. Denk T, Velitzelos E, Güner HT, Bouchal MJ, Grímsson F, Grimm GW. 2017 Taxonomy and palaeoecology of two widespread western Eurasian Neogene sclerophyllous oak species: Quercus drymeja Unger and Q. mediterranea Unger. Rev. Palaeobot. Palynol. 241, 98–128. Denk T, Zohner CM, Grimm GW, Renner SS. 2018 Plant fossils reveal major biomes occupied by the late Miocene Old-World Pikermian fauna. Nature Ecol. Evol. 2, 1864– 1870. Denk T, Güner HT, Bouchal JM. 2019 Early Miocene climate and biomes of Turkey: Evidence from leaf fossils, dispersed pollen, and petrified wood. Palaeogeogr., Palaeoecol., Palaeoclimatol. 530, 236–248. Dolezych M, Schneider W. 2006 Inkohlte Hölzer und Cuticulae dispersae aus dem 2. Miozänen Flözhorizont im Tagebau Welzow (Lausitz) – Taxonomie und vergleichende feinstratigraphisch-fazielle Zuordnung. Z. geol. Wiss. 34, 165–259. Dolezych M, Schneider W. 2007 Taxonomie und Taphonomie von Koniferenhölzern und Cuticulae dispersae im 2. Lausitzer Flözhorizont (Miozän) des Senftenberger Reviers. Palaeontograph. B 276, 1–95. Dolezych M, Schneider W. 2012 Fossil conifers in peat bog environments – results from the Central European Neogene. Japan. J. Palynol. 58 (Special Issue), 48. Dumurdžanov N, Krstić N, Mihajlović D, Ognjanovo-Rumenova N, Petrov G. 2002 New data on stratigraphy of the Neogene and Pleistocene in Mariovo, Macedonia. Geol. Macedon. 17, 43–52. Erdős L, Ambarli D, Anenkhonov OA, Bátori Z, CserhalmiD, Kiss M, Kröel-Dulay G, Liu H, Magnes M, Molnár Z, et al. 2018 The edge of two worlds: A new synthesis on Eurasian forest-steppes. Appl. Vegetation Sci. 21, 345–362. 17 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 Favre E, Escarguel G, Suc J-P, Vidal G, Thévenod L. 2008 A contribution to deciphering the meaning of AP/NAP with respect to vegetation cover. Rev. Palaeobot. Palynol. 148, 13–35. Fang J, Wang Z, Tang Z. 2009 Atlas of Woody Plants in China. Volumes 1 to 3 and index. Beijing: Higher Education Press. Fauquette S, Suc J-P, Bertini A, Popescu S-M, Warny S, Bachiri Taoufiq N, Perez Villa, M-J, Chikhi H, Subally D, Feddi N, Clauzon G, Ferrier J. 2006 How much did climate force the Messinian salinity crisis? Quantified climatic conditions from pollen records in the Mediterranean region. Palaeogeogr., Palaeoclimatol.,Palaeoecol. 238, 281–301. Ferguson DK, Hofmann C-C, Denk T. 1999 Taphonomy: field techniques in modem environments. In: Jones TP, Rowe NP (eds) Fossil Plants and Spores: modern techniques. London: Geological Society, pp 210-213. Freitag, H. 1971 Die natürliche Vegetation Afghanistans. Vegetatio 22, 285–344. Gersonde R, Velitzelos E. 1978 Diatomeenpaläoökologie im Neogen-becken von Vegora NW Mazedonien (vorläufige Mitteilung). Ann. Gèol. Pays Hellèn. 30, 373–382. Grimm GW, Bouchal JM, Denk T, Potts A. 2016 Fables and foibles: A critical analysis of the Palaeoflora database and the Coexistence Approach for palaeoclimate reconstruction. Rev. Palaeobot. Palyn. 233, 611–622. Grimm GW, Potts A. 2016 Fallacies and fantasies: the theoretical underpinnings of the Coexistence Approach for palaeoclimate reconstruction. Clim. Past 12, 611–622. Halbritter H, Ulrich S, Grímsson F, Weber M, Zetter R, Hesse M, Buchner R, Svojtka M, Frosch-Radivo A. 2018 Illustrated Pollen Terminology, 2nd ed. Cham (Switzerland): Springer Nature. Herbert TD, Lawrence KT, Tzanova A, Peterson LC, Caballiero-Gill R, Kelly CS. 2016 Late Miocene global cooling and the rise of modern ecosystems. Nature Geosci. 9, 843–847. Ivanov DA. 2001 Palaeoecological interpretation of a pollen diagram from the Sandanski graben (Southwest Bulgaria). Comptes Rend. Acad. Bulg. Sci. 54, 65–68. Ivanov DA, Slavomirova E. 2002 Preliminary palynological data on Neogene flora from Bitola Basin (F.Y.R.O.M.). Comptes Rend. Acad. Bulg. Sci. 55, 81–86. Janis CM. 1986 An estimation of tooth volume and hypsodonty indices in ungulate mammals, and the correlation of these factors with dietary preference. In: DE Russell, JP Santoro, D Sigogneau-Russell (eds) Teeth revisited. Proc. 7th Int. Symp. Dental Morph. Mém. Mus. Nat. Hist. naturelle, Sér. C 53, 367–387. Jiang X-L, Hipp AL, Deng M, Su T, Zhou Z-K, Yan M-X. 2019 East Asian origins of European holly oaks (Quercus section Ilex Loudon) via the Tibet-Himalaya. J. Biogeogr. 46, 2188–2202. doi.org/10.1111/jbi.13654 Karistineos N, Ioakim C. 1989 Palaeoenvironmental and palaeoclimatic evolution of the Serres Basin (N. Greece) during the Miocene. Palaeogeogr., Palaeoclimatol., Palaeoecol. 70, 275–285. Knobloch E, Velitzelos E. 1986a Die obermiozäne Flora von Likudi bei Elassona (Thessalien, Griechenland). Doc. Nat. 29, 5–20. Knobloch E, Velitzelos E. 1986b Die obermiozäne Flora von Prosilion bei Kozani (SudMazedonien, Griechenland). Doc. Nat. 29, 29–33. Kottek M, Grieser J, Beck C, Rudolf B, Rubel F. 2006 World map of the Köppen-Geiger climate classification updated. Meteorol. Z. 15, 259–263. Koufos GD. 1982 Hipparion crassum Gervais, 1859 from the lignites of Ptolemais (Macedonia-Greece). Proc. Kon. Ned. Akad. Weten. B. 85, 229–239. 18 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 Koufos GD. 2006 The Neogene mammal localities of Greece: Faunas, chronology and biostratigraphy. Hellen. J. Geosci. 41, 183–214 Koufos GD, Kostopoulos DS, Koliadimou KK. 1991 Un nouveau gisement de mammiferes dans le Villafranchien de Macedoine occidentale (Gréce). C. R. Acad. Sci. Paris, ser. II 313, 831–836. Kovar-Eder J, Jechorek H, Kvaček Z, Parashiv V. 2008 The Integrated Plant Record: An essential tool for reconstructing Neogene zonal vegetation in Europe. Palaios 23, 97– 111. Kovar-Eder J, Kvaček Z, Martinetto E, Roiron P. 2006 Late Miocene to Early Pliocene vegetation of southern Europe (7-4 Ma) as reflected in the megafossil plant record. Palaeogeogr., Palaeoclimatol., Palaeoecol. 238, 321–339. Kozlowski G, Bétrisey S, Song Y. 2018 Wingnuts (Pterocarya) and walnut family. Relict trees: linking the past, present and future. Natural History Museum Fribourg: Switzerland. Kvaček Z, Velitzelos D, Velitzelos E. 2002 Late Miocene Flora of Vegora Macedonia N. Greece. Athens: Koralis. Maharramova E. 2015 Genetic diversity and population structure of the relict forest trees Zelkova carpinifolia (Ulmaceae) and Pterocarya fraxinifolia (Juglandaceae) in the South Caucasus. PhD Dissertation. Freie Universität Berlin. Mai HD. 1995 Tertiäre Vegetationsgeschichte Europas. Stuttgart: Gustav Fischer. Mai DH, Velitzelos E. 1992 Über fossile Pinaceen-Reste im Jungtertiär von Griechenland. Feddes Repert. 103, 1–18. Mai DH, Velitzelos E. 1997 Paläokarpologische Beiträge zur jungtertiären Flora von Vegora (Nordgriechenland). Feddes Repert. 108, 507–526. Marinova E, Harrison SP, Bragg F, Connor S, de Laet V, Leroy SAG, Mudie P, Atanassova J, Bozilova E, Caner H, Cordova C, Djamali M, Filipova-Marinova M, Gerasimenko N, Jahns S, Kouli K, Kotthoff U, Kvavadze E, Lazarova M, Novenko E, Ramezani E, Röpke A, Shumilovskikh L, Tanţău I, Tonkov S. 2018 Pollen-derived biomes in the Eastern Mediterranean–Black Sea–Caspian-Corridor. J. Biogeogr. 45, 484–499. Mayer H, Aksoy, H. 1986 Wälder der Türkei. Stuttgart: Gustav Fischer. Mosbrugger V, Utescher T. 1997 The coexistence approach — a method for quantitative reconstructions of Tertiary terrestrial palaeoclimate data using plant fossils. Palaeogeogr., Palaeoclimatol., Palaeoecol. 134, 61–86 Ognjanova-Rumenova N. 2005 Upper Neogene siliceous microfossils from Pelagonia Basin (Balkan Peninsula). Geol. Carpathica 56 (4), 347–358. Pavlides SB, Mountrakis DM. 1986 Neotectonics of the Florina–Vegoritis–Ptolemais Neogene Basin (NW Greece): an example of extensional tectonics of the greater Aegean area. Ann. Géol. Pays Hellen. 33, 311–327. Peel MC, Finlayson BL, McMahon TA. 2007 Updated world map of the Köppen-Geiger climate classification. Hydrol. Earth Syst. Sci., 11, 1633–1644. Psilovikos A, Karistineos N. 1986 A depositional sedimentary model for the Neogene uraniferous lignites of the Serres graben, Greece. Palaeogeogr., Palaeoclimatol., Palaeoecol. 56, 1–16. Punt W, Hoen PP, Blackmore S, Nilsson RH, Le Thomas A. 2007 Glossary of pollen and spore terminology. Rev. Palaeobot. Palynol. 143, 1–81. 19 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 Ratnam J, Bond WJ, Fensham RJ, Hoffmann WA, Archibald S, Lehmann CER, Anderson MT, Higgins SI, Sankaran M. 2011 When is a ‘forest’ a savanna, and why does it matter? Glob. Ecol. Biogeogr. 20, 653–660. Roberts S, Rassios A, Wright L, Vacondios I, Vrachatis G, GrivasE, Nesbitt RW, Neary CR, Moat T, Kostantopolou L. 1988 Structural controls on the location and form of the Vourinos chromite deposits. In: Boissonnas J, Omenetto P (eds) Mineral deposits within the European Community. Springer, Berlin Heidelberg New York, pp 249–266. Rubel F, Brugger K, Haslinger K, Auer I. 2017 The climate of the European Alps: Shift of very high resolution Köppen-Geiger climate zones 1800–2100. Meteorol. Z., 26, 115– 125. https://doi.org/10.1127/metz/2016/0816. Schneider W. 1992 Floral successions in the Miocene swamps and bogs of Central. Europe. Z. Geol.Wiss. 20, 555–570. Spicer RA. 2008 CLAMP. In: Gornitz V. Ed. Encyclopedia of Paleoclimatology and Ancient Environments. Dordrecht: Springer. Steenbrink J, Van Vugt N. Hilgen FJ, Wijbrans JR, Meulenkamp JE. 1999 Sedimentary cycles and volcanic ash beds in the lower Pliocene lacustrine succession of Ptolemais (NW Greece): discrepancy between 40Ar/39Ar and astronomical ages. Palaeogeogr., Palaeoclimatol., Palaeoecol. 152, 283–303. Steenbrink J, Van Vugt N, Kloosterboer-van Hoeve ML, Hilgen FJ. 2000 Refinement of the Messinian APTS from sedimentary cycle patterns in the lacustrine Lava section (Servia Basin, NW Greece). Earth Planet. Sci. Lett. 181, 161–173. Steenbrink J, Hilgen FJ, Krijgsman W, Wijbrans JR, Meulenkamp JE. 2006 Late Miocene to Early Pliocene depositional history of the intramontane Florina-Ptolemais-Servia Basin, NW Greece: Interplay between orbital forcing and tectonics. Palaeogeogr., Palaeoclimatol., Palaeoecol. 238, 151–178. Suc J-P. 1984 Origin and evolution of the Mediterranean vegetation and climate in Europe. Nature 307, 429–432. Suc J-P, Popescu S-M, Do Couto D, Clauzon G, Rubino J-L, Melinte-Dobrinescu MC, Quillévéré F, Brun J-P, Dumurdžanov N, Zagorchev I, Lesdić V, Tomić D, Sokoutis D, Meyer B, Macaleţ R, Jelen B, Rihelj H. 2015 Marine gateway vs. fluvial stream within the Balkans from 6 to 5 Ma. Mar. Pet. Geol. 66, 231–245. Suc J-P, Popescu SM, Fauquette S, Bessedik M, Jiménez-Moreno G, Taoufiq B, Zheng Z, Medail F, Klotz S. 2018 Reconstruction of Mediterranean flora, vegetation and climate for the last 23 million years based on an extensive pollen dataset. Ecol. mediterr. 44, 53–85. Traverse A. 2007 Paleopalynology. Topics in Geobiology 28. Dordrecht, Springer. Utescher T, Bruch AA, Erdei B, François I, Ivanov D, Jacques FMB, Kern AK, Liu Y-SC, Mosbrugger V, Spicer RA. 2014 The Coexistence Approach—Theoretical background and practical considerations of using plant fossils for climate quantification. Palaeogeogr., Palaeoclimatol., Palaeoecol. 410, 58–73. Van de Weerd A. 1979 Early Ruscinian rodents and lagomorphs (Mammalia) from the lignites near Ptolemais (Macedonia, Greece). Proc. Kon. Nederl. Akad. Wet., B. 82, 127–170. Van Zeist W, Bottema S. 1991 Late Quaternary vegetation of the Near East. Beih. Tübinger Atlas Vord. Orient, Reihe A (Naturwiss.) 18, 1–156. Van Zeist W, Woldring H, Stapert, D. 1975 Late Quaternary vegetation and climate of southwestern Turkey. Palaeohist. 14, 35–143. 20 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 Velitzelos D, Denk T. 2002 Leaf epidermal characteristics of late Tertiary conifers from Greece: taxonomic significance and limitations. 6th Europ. Paleobot. Palynol. Conf. Athens, Greece, Abstracts, 183–184. Velitzelos D, Bouchal JM, Denk T. 2014 Review of the Cenozoic floras of Greece. Rev. Palaeobot. Palynol. 204, 1–15. Velitzelos E, Gregor H-J. 1985 Neue paläofloristische Befunde im Neogen Griechenlands. Doc. Nat. 25, 1–4. Velitzelos E, Krach JE, Gregor H-J, Geissert F. 1983 Bolboschoenus vegorae – ein Vergleich fossiler und rezenter Rhizomknollen der Strandbinse. Doc. Nat. 5, 1–57. Velitzelos E, Kvaček Z. 1999 Review of the late Miocene flora of Vegora western Macedonia, Greece. Acta Palaeobotanica, Suppl. 2 (Proceed. 5th EPPC) 419–427. Velitzelos E, Petrescu I. 1981 Seltene pflanzliche Fossilien aus dem Braunkohlebecken von Vegora. Ann. géol. Pays hellén. 30, 767–777. Velitzelos E, Schneider HE. 1979 Jungtertiäre Pflanzenfunde aus dem Becken von Vegora in West-Mazedonien. 3. Mitteilung: Eine Fächerpalme (Chamaerops humulis L.). Ann. géol. Pays hellén. 29, 796–799. Woodward FI, Lomas MR, Kelly CK. 2004 Global climate and the distribution of plant biomes. Phil. Trans. R. Soc. Lond. B. 359, 1465–1476. Yang J, Spicer RA, Spicer TEV, Li C-S. 2011 'CLAMP Online': a new web-based palaeoclimate tool and its application to the terrestrial Paleogene and Neogene of North America. Palaeobiodiv. Palaeoenviron. 91, 163–183. Zetter R. 1989 Methodik und Bedeutung einer routinemäßigen kombinierten lichtmikroskopischen und rasterelektronenmikroskopischen Untersuchung fossiler Mikrofloren. Cour. Forschungsinst. Senck. 109, 41–50. Table and Figure Captions Table 1. Plant taxa recorded from unit 1 (lignite seam) and unit 2 (blue marls) of the Vegora section. Table 2. Estimated climate parameters for the pre-evaporitic Messinian of Vegora from two CLAMP calibration datasets and from CA. Table 3. Vegetation types recognised for the pre-evaporitic Messinian of the Florina– Ptolemais–Servia Basin. Figure 1. Fossil localities and lithological map of the Florina–Ptolemais–Servia Basin. Map redrawn after Steenbrink et al. (1999, 2006), Ognjanova-Rumenova (2005), Ivanov (2001) and Koufos (2006). Fossil localities: (1) Bitola Basin, Republic of North Macedonia, PF. (2) Vegora Basin, MF and PF (3) Dytiko, VF. (4) Prosilio, MF. (5) Lava, MF. (6) Likoudi, MF. (7) Serres Basin. (2–7) Greece. (8) Sandanski Graben, Bulgaria, PF. Abbreviations: Plant macrofossils (MF), palynoflora (PF), vertebrate fossils (VF). Figure 2. Lithology and polarity zones of the Vegora section (redrawn after Steenbrink et al., 2006). Position of fossil bearing strata following Velitzelos and Schneider (1979) and Kvaček et al. (2002). Figure 3. Light microscopy (LM) and scanning electron microscopy (SEM) micrographs of algae, fern and fern allies, and gymnosperm palynomorphs. (a) Botryococcus sp. cf. B. kurzii. (b) Spirogyra sp. 1/ Ovoidites elongatus. (c) Spirogyra sp. 2/Cycloovoidites cyclus. (d–e) Osmunda sp., (d) EV, (e) PV. (f) Cryptogramma vel Cheilanthes sp, PV. (g–h) Pteris sp., (g) PV, (h) DV. (i) Davalliaceae vel Polypodiaceae sp./ 21 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 Verrucatosporites alienus (R.Potonié) P.W.Thomson et Pflug, 1953, EV. (j) Leavigatosporites haardti, EV. (k–l) Inaperturopollenites hiatus. (m) Abies sp., EV. (n–o) Cathaya sp., (n) PV, (o) SEM detail, nanoechinate sculpturing of cappa (PRV). (p) Cedrus sp., EV. (q) Pinus subgenus Pinus sp., EV. (r) Pinus subgenus Strobus sp., EV. (s–t) Tsuga sp. 1, (s) PV, (t) monosaccus and corpus detail, PRV. (u–v) Tsuga sp. 2, (u) PV, (v) monosaccus and corpus detail, PRV. Abbreviations: equatorial view (EV), polar view (PV), distal view (DV), proximal view (PRV). Scale bars 10 µm (LM, h, t, v), 1 µm (o). Figure 4. LM and SEM micrographs of Poales, Vitales, Rosales, Fagales, Malpighiales, and Geraniales. (a) Typha sp, tetrad, PV. (b–c) Poaceae gen. indet., EV, (c) exine detail, PRV. (d–e) Monocotyledone indet., (d) PV, (e) PRV. (f–g) Parthenocissus sp., EV. (h) Ulmus vel Zelkova sp., PV. (I) Fagus sp., EV. (j–k) Quercus sect. Cerris sp., EV, (k) SEM detail, mesocolpium exine sculpturing. (l–m) Quercus sect. Ilex sp., EV, (m) SEM detail, mesocolpium exine sculpturing. (n–o) Quercus sect. Quercus sp., PV, (o) SEM detail, apocolpium exine sculpturing. (p–q) Castanoideae gen. indet. sp., EV, (q) SEM detail, mesocolpium exine sculpturing. (r) Carya sp., PV. (s) Platycarya sp., PV. (t) Engehardioideae gen. indet., PV. (u) Alnus sp., PV. (v) Betula sp., PV. (w) Carpinus sp., PV. (x) Corylus sp., PV. (y) Salix sp., EV. (z–aa) Geranium sp., (z) PV, (aa) clavae detail. Abbreviations: equatorial view (EV), polar view (PV), proximal view (PRV). Scale bars 10 µm (LM, e, g), 1 µm (c, k, m, o, q, aa). Figure 5. LM and SEM micrographs of Sapindales, Malvales, Caryophyllales, Cornales, Asterales, Dipsacales, and Apiales. (a–b) Cotinus sp, EV. (c–d) Pistacia sp., (c) PV, (d) exine SEM detail. (e–f) Acer sp. 1, (e) PV, (f) mesocolpium SEM detail. (g–h) Acer sp. 2, (g) PV, (h) mesocolpium SEM detail. (i–j) Craigia sp., (i) PV, (j) apocolpium SEM detail. (k) Amaranthaceae gen. indet. sp. 1. (l) Amaranthaceae gen. indet. sp. 2. (m–n) Caryophyllaceae gen. indet. sp. (o–p) Nyssa sp., (o) PV, (p) exine sculpturing and aperture SEM detail. (q–r) Fraxinus sp., (q) EV, (r) mesocolpium SEM detail. (s–t) Olea sp., EV. (u) Cichorioideae gen. indet. sp., PV. (v) Asteroideae gen indet. sp. 1, PV. (w) Asteroideae gen indet. sp. 2, PV. (x–z) Valeria sp., (x– y) PV, (z) aperture SEM detail. (aa–bb) Apiaceae gen. indet. sp. 1, EV. (cc–dd) Apiaceae gen. indet. sp. 2, EV. (ee–ff) Angiosperm pollen fam. et gen. indet. sp., (ee) EV, (ff) mesocolpium SEM detail. Abbreviations: equatorial view (EV), polar view (PV). Scale bars 10 µm (LM, b, n, t, y, z, bb, dd), 1 µm (d, f, h, j, p, r, ff). Figure 6. Coexistence-Approach diagram showing coexistence intervals for MAT and CMMT. MAT and CMMT climate ranges of relict taxa a priori excluded from the analysis are shown on the left side of the diagram. Blue bars, coldest month mean temperature; red bars, 10–90 percentile climatic range; dark red extensions, full climatic range. Figure 7. Köppen signal diagram for the macrofossil and pollen floras of Vegora. To test and illustrate the stability of the climatic signal, gymnosperms (common alpine elements) and azonal elements (e.g. riparian or swamp vegetation) were excluded in some runs. Supplementary Material 22 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 Supplementary Material Tables S1 - Climatic parameters of NLR Table S1 (1). Fossil species and climatic parameters of the corresponding NLR (depending on the fossil-species and their botanical affinities, climate parameters of species, sections, subgenera, genera, or subfamilies are used as NLR). Table S1 (2). Climatic parameters of NLR. Supplementary Material Tables S2 - Köppen-Geiger climate type signatures. Table S2 (1). Scored Köppen-Geiger signatures of all NLR species of the macrofossil and pollen flora of Vegora. Tables S2 (2). Köppen-Geiger signature values and diagram of the macrofossil and pollen flora of Vegora. Supplementary Material S3 - Systematic palaeobotany and descriptions of palynomorphs from the plant fossil bearing strata of Vegora (sample S115992). Supplementary Material Tables S4 - Palynomorph abundance of sample S115992. Supplementary Material S5 - Coding of leaf physiognomic characters for morphotypes from the Vegora lignite mine macroflora. Output PDF files from online CLAMP analysis (http://clamp.ibcas.ac.cn). Supplementary Material S6 - Köppen-Geiger categories 23 bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. Figure 1. Figure 1. Fossil localities and lithological map of the Florina–Ptolemais–Servia Basin. Map redrawn after Steenbrink et al. (1999, 2006), Ognjanova-Rumenova (2005), Ivanov (2001) and Koufos (2006). Fossil localities: (1) Bitola Basin, Republic of North Macedonia, PF. (2) Vegora Basin, MF and PF (3) Dytiko, VF. (4) Prosilio, MF. (5) Lava, MF. (6) Likoudi, MF. (7) Serres Basin. (2–7) Greece. (8) Sandanski Graben, Bulgaria, PF. Abbreviations: Plant macrofossils (MF), palynoflora (PF), vertebrate fossils (VF). bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. Figure 2. Figure 2. Lithology and polarity zones of the Vegora section (redrawn after Steenbrink et al., 2006). Position of fossil bearing strata following Velitzelos and Schneider (1979) and Kvaček et al. (2002). bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. Figure 3. Figure 3. Light microscopy (LM) and scanning electron microscopy (SEM) micrographs of algae, fern and fern allies, and gymnosperm palynomorphs. (a) Botryococcus sp. cf. B. kurzii. (b) Spirogyra sp. 1/ Ovoidites elongatus. (c) Spirogyra sp. 2/Cycloovoidites cyclus. (d–e) Osmunda sp., (d) EV, (e) PV. (f) Cryptogramma vel Cheilanthes sp, PV. (g–h) Pteris sp., (g) PV, (h) DV. (i) Davalliaceae vel Polypodiaceae sp./ Verrucatosporites alienus (R.Potonié) P.W.Thomson et Pflug, 1953, EV. (j) Leavigatosporites haardti, EV. (k–l) Inaperturopollenites hiatus. (m) Abies sp., EV. (n–o) Cathaya sp., (n) PV, (o) SEM detail, nanoechinate sculpturing of cappa (PRV). (p) Cedrus sp., EV. (q) Pinus subgenus Pinus sp., EV. (r) Pinus subgenus Strobus sp., EV. (s–t) Tsuga sp. 1, (s) PV, (t) monosaccus and corpus detail, PRV. (u–v) Tsuga sp. 2, (u) PV, (v) monosaccus and corpus detail, PRV. Abbreviations: equatorial view (EV), polar view (PV), distal view (DV), proximal view (PRV). Scale bars 10 µm (LM, h, t, v), 1 µm (o). bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. Figure 4. Figure 4. LM and SEM micrographs of Poales, Vitales, Rosales, Fagales, Malpighiales, and Geraniales. (a) Typha sp, tetrad, PV. (b–c) Poaceae gen. indet., EV, (c) exine detail, PRV. (d–e) Monocotyledone indet., (d) PV, (e) PRV. (f–g) Parthenocissus sp., EV. (h) Ulmus vel Zelkova sp., PV. (I) Fagus sp., EV. (j–k) Quercus sect. Cerris sp., EV, (k) SEM detail, mesocolpium exine sculpturing. (l–m) Quercus sect. Ilex sp., EV, (m) SEM detail, mesocolpium exine sculpturing. (n–o) Quercus sect. Quercus sp., PV, (o) SEM detail, apocolpium exine sculpturing. (p–q) Castanoideae gen. indet. sp., EV, (q) SEM detail, mesocolpium exine sculpturing. (r) Carya sp., PV. (s) Platycarya sp., PV. (t) Engehardioideae gen. indet., PV. (u) Alnus sp., PV. (v) Betula sp., PV. (w) Carpinus sp., PV. (x) Corylus sp., PV. (y) Salix sp., EV. (z–aa) Geranium sp., (z) PV, (aa) clavae detail. Abbreviations: equatorial view (EV), polar view (PV), proximal view (PRV). Scale bars 10 µm (LM, e, g), 1 µm (c, k, m, o, q, aa). bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. Figure 5. Figure 5. LM and SEM micrographs of Sapindales, Malvales, Caryophyllales, Cornales, Asterales, Dipsacales, and Apiales. (a–b) Cotinus sp, EV. (c–d) Pistacia sp., (c) PV, (d) exine SEM detail. (e–f) Acer sp. 1, (e) PV, (f) mesocolpium SEM detail. (g–h) Acer sp. 2, (g) PV, (h) mesocolpium SEM detail. (i–j) Craigia sp., (i) PV, (j) apocolpium SEM detail. (k) Amaranthaceae gen. indet. sp. 1. (l) Amaranthaceae gen. indet. sp. 2. (m–n) Caryophyllaceae gen. indet. sp. (o–p) Nyssa sp., (o) PV, (p) exine sculpturing and aperture SEM detail. (q–r) Fraxinus sp., (q) EV, (r) mesocolpium SEM detail. (s–t) Olea sp., EV. (u) Cichorioideae gen. indet. sp., PV. (v) Asteroideae gen indet. sp. 1, PV. (w) Asteroideae gen indet. sp. 2, PV. (x–z) Valeria sp., (x–y) PV, (z) aperture SEM detail. (aa–bb) Apiaceae gen. indet. sp. 1, EV. (cc–dd) Apiaceae gen. indet. sp. 2, EV. (ee–ff) Angiosperm pollen fam. et gen. indet. sp., (ee) EV, (ff) mesocolpium SEM detail. Abbreviations: equatorial view (EV), polar view (PV). Scale bars 10 µm (LM, b, n, t, y, z, bb, dd), 1 µm (d, f, h, j, p, r, ff). bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. Figure 6. Figure 6. Coexistence-Approach diagram showing coexistence intervals for MAT and CMMT. MAT and CMMT climate ranges of relict taxa a priori excluded from the analysis are shown on the left side of the diagram. Blue bars, coldest month mean temperature; red bars, 10–90 percentile climatic range; dark red extensions, full climatic range. bioRxiv preprint doi: https://doi.org/10.1101/848747. this version posted November 25, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. Figure 7. Figure 7. Köppen signal diagram for the macrofossil and pollen floras of Vegora. To test and illustrate the stability of the climatic signal, gymnosperms (common alpine elements) and azonal elements (e.g. riparian or swamp vegetation) were excluded in some runs. bioRxiv https://doi.org/10.1101/848747 November 25, 2019. The copyright holder for this preprint (which was Table 1. Plant taxapreprint recordeddoi: from unit 1 (lignite seam) and unit 2. this (blueversion marls) posted of the Vegora section. not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. Vegora micro, meso and macro flora Taxon Element Reference Life form Ecology Vegetation units [BLD subgroups]a Algae Botryococcaceae Botryococcus sp. P 11 Algae Aquatic VU1 Zygnemataceae Spirogyra spp. P 11 Algae Aquatic VU1 P L 11 2, 8 Fern Fern Swamp, riparian, well-drained lowland forest Swamp, riparian, well-drained lowland forest VU3, VU4, VU5 VU3, VU4, VU5 Polypodiaceae Cryptogramma vel Cheilanthes P 11 Fern Swamp, riparian, well-drained lowland forest VU3, VU4, VU5 Pteridaceae Pteris sp. P 11 Fern Swamp, riparian, well-drained lowland forest VU3, VU4, VU5 Fam. incerta sedis Davalliaceae vel Polypodiaceae P 11 Fern Swamp, riparian, well-drained lowland forest VU3, VU4, VU5 P 11 Fern Swamp, riparian, well-drained lowland forest VU3, VU4, VU5 Gymnosperms Ginkgoaceae Ginkgo adiantoides L L 8, 9 Treegym Riparian forests VU4 Cupressaceae Cupressus rhenana L, R 6, 8 Treegym Conifer forest lowland, upland, peat-forming VU7 Sequoia abietina L 8, 10 Tree gym Conifer forest lowland, upland, peat-forming VU7 Cryptomeria anglica L 10 Treegym Conifer forest lowland, upland, peat-forming VU7 Glyptostrobus europaeus L, R 6, 8 Treegym Swamp forest VU3 Taxodium dubium L, R 6, 8 Tree gym Swamp forest VU3 P 11 Tree gym Indifferent Fern and fern allies Osmundaceae Osmunda sp. Osmunda parschlugiana Laevigatosporites haardti Papillate Cupressaceae Pinaceae Abies sp. P 11 Treegym Well-drained lowland and upland forests VU5, VU6, VU7 Cathaya sp. P 11 Treegym Conifer forest lowland, upland, peat-forming VU7 Cedrus sp. P 11 Treegym Well-drained lowland and upland forests VU7 Cedrus vivariensis R 4, 8 Treegym Well-drained lowland and upland forest VU7 Keteleeria hoehnei Pinus hampeana (diploxylon) R 6 Tree gym Conifer forest lowland, upland, peat-forming VU7 R 6, 8 Tree gym Well-drained lowland forest VU5 Pinus salinarum (diploxylon) R 4, 8 Tree gym Well-drained lowland forest VU5 L, R 8 Tree gym Indifferent Pinus spp. Pinus sp. diploxylon type P 11 Tree gym Indifferent Pinus sp. haploxylon type P, R 8, 11 Tree gym Indifferent Pinus vegorae (haplox.) R 4, 6 Tree gym Well-drained lowland forest VU5 Tsuga spp. P 11 Tree gym Conifer forest lowland, upland, peat-forming VU7 Angiosperms Cabombaceae Brasenia sp. R 4 Herb Aquatic VU1 Lauraceae Daphnogene pannonica L 8 Tree Well-drained lowland forest VU5 [BLD wet] Laurophyllum pseudoprinceps L 8 Tree Well-drained lowland forest VU5 [BLD wet] Sassafras ferrettianum Laurophyllum sp. L L 7, 8 8 Tree Tree Riparian, well-drained lowland forest Indifferent VU4, VU5 [BLD wet] [BLD wet] Potamogetonaceae Potamogeton sp. R 4 Herb Aquatic VU1 Arecaceae Chamaerops humilis fossilis L 1, 8 Palm Well-drained lowland forest or scrub VU0, VU5 [BLD drought] S 4 Herb Bogs, wet meadows VU2 P 11 Herb Aquatic, bogs, swamp forest, riparian forest VU1, VU2, VU3, VU4 P 11 Herb Indifferent S 3, 4 Herb Meadows Zingiberaceae Spirematospermum wetzleri Typhaceae Typha sp. Poaceae Poaceae gen. indet. Cyperaceae Bolboschoenus vegorae VU2 Cladium bioRxiv preprint doi: https://doi.org/10.1101/848747 .Herb this version posted November 25, 2019. The copyright holder S 4 Bogs, wet meadows VU2for this preprint (which was Ceratophyllaceae Ceratophyllum sp. not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. S 4 Herb Aquatic VU1 Platanaceae L 8 Tree Riparian, well-drained lowland forest VU4, VU5 [BLD drought] Vitaceae Parthenocissus sp. Platanus leucophylla P 11 Liana Swamp forest, riparian, well-drained lowland forest VU3, VU4, VU6 Fabaceae Leguminosites sp. L 8 Tree Indifferent [BLD] Ulmus plurinervia P L 11 8 Tree Tree Riparian, well-drained lowland forest Riparian, well-drained lowland forest VU0, VU4, VU5 [BLD] VU4, VU5 [BLD cold] Zelkova zelkovifolia L 8 Tree Mediterranean scrub, riparian, well-drained lowland forest VU0, VU4, VU5 [BLDdrought] Ulmaceae Ulmus vel Zelkova Fagaceae Castaneoideae gen indet. Castanea sp. Fagus sp. P 11 Tree Well-drained lowland forest VU5 [BLD] R P 6, 8 11 Tree Tree Well-drained lowland forest Well-drained lowland and upland forest VU5 [BLD] VU5, VU5 [BLD cold] L, R 6, 7, 8 Tree Well-drained lowland and upland forest VU5, VU6 [BLD cold] P 11 Tree Well-drained lowland forest VU5 [BLD] Quercus cerrisaecarpa Quercus gigas R L 6, 8 7, 8 Tree Tree Well-drained lowland forest Well-drained lowland forest VU5 [BLD] VU5 [BLD] Quercus kubinyii Quercus sect. Ilex L 8 Tree Well-drained lowland forest VU5 [BLD drought] P 11 Tree Mediterranean scrub, well-drained lowland forest VU0, VU5 [BLD drought] Quercus drymeja L 7, 8 Tree Well-drained lowland and upland forests VU5, VU6 [BLD drought] Quercus mediterranea L 7, 8 Tree Mediterranean scrub, well-drained lowland forest VU0, VU5 [BLD drought] Quercus sosnowskyi L 7, 8 Tree Well-drained lowland forest VU5 [BLD drought] Quercus sect. Quercus Quercus pseudocastanea P L 11 8 Tree Tree Riparian forest, well-drained lowland forest Well-drained lowland forest VU4, VU5 [BLD] VU5 [BLDcold] Quercus sp. R 8 Tree Indifferent [BLD] Juglandaceae Carya sp. Platycarya sp. P P 11 11 Tree Tree Riparian, well-drained lowland forest Riparian, well-drained lowland forest VU4, VU5 [BLD] VU4, VU5 [BLD] L P 7, 8 11 Tree Tree Riparian, well-drained lowland forest Riparian, well-drained lowland forest VU4, VU5 [BLD] VU4, VU5 [BLD] Fagus gussonii Quercus sect. Cerris Pterocarya paradisiaca Engelhardioideae gen. indet. Betulaceae Alnus sp. P 11 Tree Swamp, riparian forest, well-drained lowland forest VU3, VU4, VU5 [BLD cold] Alnus adscendens L 8 Tree Well-drained lowland forest VU5 [BLD cold] Alnus cecropiifolia Alnus cf. kefersteinii L 8 Tree Swamp, riparian forest VU3, VU4 [BLD cold] R 6, 8 Tree Indifferent [BLD cold] Alnus ducalis L 7, 8 Tree Well-drained lowland forest VU5 [BLD cold] Alnus gaudinii L 8 Tree Well-drained lowland forest VU5 [BLD cold] Alnus julianiformis Betula sp. L 8 Tree Riparian, well-drained lowland forest VU4, VU5 [BLD cold] P 11 Tree Riparian, well-drained lowland and upland forest VU4, VU5, VU6 [BLD cold] Betula pseudoluminifera Carpinus sp. L 8 Tree Well-drained lowland forest VU5 [BLD cold] P 11 Tree Well-drained lowland forest VU5 [BLD] Carpinus betulus fossilis Carpinus grandis Carpinus kisseri, group of C. tschonoskii Corylus sp. R L R P 8 8 6, 8 11 Tree Tree Tree Shrub Well-drained Well-drained Well-drained Well-drained VU5 VU5 VU5 VU5 L 8 Tree Riparian forest VU4 [BLD] L L P 8 7, 8 11 Tree Tree Tree Riparian forest Riparian forest Swamp, riparian forest VU4 [BLD] VU4 [BLD] VU3, VU4 [BLD] P 11 Herb Steppe, meadows, well-drained lowland forest VU0, VU2, VU5 R 4 Shrub Swamp VU3 P 8, 11 Shrub Well-drained lowland forest VU5 [BLD drought] P 11 Tree, shrub (Mediterranean) scrub, well-drained lowland forest VU0, VU5 Acer aegopodifolium Acer integrilobum Acer limburgense (sect. Macrophylla ) L L R 7, 8 8 6, 8 Tree Tree Tree Well-drained lowland and upland forests Well-drained lowland and upland forests Well-drained lowland and upland forests VU5, VU6 [BLD] VU5, VU6 [BLD] VU5, VU6 [BLD] Acer pseudomonspessulanum Acer pyrenaicum (sect. Rubra) L 8 Tree (Mediterranean) scrub, well-drained lowland forest VU0, VU5 [BLD drought] L 7, 8 Tree Well-drained lowland forest VU5 [BLD] Salicaceae Populus balsamoides Populus populina Populus spp. Salix sp. Geraniaceae Geranium sp. lowland lowland lowland lowland forest forest forest forest [BLD] [BLD] [BLD] [BLD] Lythraceae Decodon globosus Anacardiaceae Cotinus sp. (=Dicotylophyllum sp. 5) Pistacia sp. Sapindaceae bioRxiv preprint doi: https://doi.org/10.1101/848747 .Tree this version posted November 2019. this[BLD preprint (which was Acer subcampestre L 8 Well-drained lowland and25, upland forestThe copyright holder VU5,for VU6 ] drought not certified by peer review) is the author/funder. It is made available under a CC-BY-NC 4.0 International license. Acer tricuspidatum (sect. Rubra ) L 8 Tree Swamp, well-drained lowland and upland forest VU3, VU5, VU6 [BLD cold] Acer spp. Malvaceae Craigia sp. P, R 8, 11 Tree Indifferent [BLD] P 11 Tree Well-drained lowland forest VU5 [BLD] Craigia bronnii Droseraceae R 6, 8 Tree Well-drained lowland forest VU5 [BLD] Aldrovandia praevesiculosa Caryophyllaceae Caryophyllaceae gen. indet. sp. Amaranthaceae S 4 Herb Aquatic VU1 P 11 Herb Steppe, meadows, well-drained lowland and upland forest VU0, VU2, VU5, VU6 Amaranthaceae/Chenopodiaceae gen. indet. sp. Nyssaceae Nyssa sp. P 11 Herb Steppe, meadows VU0, VU2 P 11 Tree Swamp, well-drained lowland and upland forest VU3, VU5, VU6 [BLD wet] R, P P 8, 11 11 Tree Tree Riparian forest Mediterranean scrub, well-drained lowland forest VU4 [BLD] VU0, VU5 P 11 Herb Steppe, meadows, well-drained lowland and upland forest VU0, VU2, VU5, VU6, VU7 P 11 Herb Steppe, meadows, well-drained lowland and upland forest VU0, VU2, VU5, VU6, VU7 P 11 Herb Steppe, meadow, riparian VU0, VU2, VU4 L 7, 8 Liana Riparian, well-drained lowland forest VU4, VU5 P 11 Herb Steppe, meadows, well-drained lowland and upland forest VU0, VU2, VU5, VU6, VU7 Monocotyledone indet. Dicotylophyllum sp. 1–4, 6 (oder spp.) L L 8 8 Herb Swamp, riparian, lake margin VU2, VU3, VU4 Monocotyledone indet. Pollen indet. P P 11 11 Oleacaee Fraxinus sp. Olea sp. Asteraceae Asteraceae gen. indet. spp. Cichorioideae gen. indet. sp. Caprifoliaceae Succisa sp. Araliaceae Hedera multinervis Apiaceae Apiaceae gen. indet. spp. Incerta sedis a BLD (Broadleaf deciduous forest biome of Woodward et al., 2004) was divided in BLD drought, deciduous trees and shrubs that are drought resistant, and BLD cold, deciduous trees and shrubs that are cold tolerant; in addition we use BLD wet for trees and shrubs that typically occur in humid warm temperate regions. Vegetation Unit (VU) 0: Steppe-like meadows with shrubs and/or small trees scattered or in groups; Mediterranean scrub. VU 1: Aquatic. VU 2: Bogs, wet meadows. VU 3: Swamp forest. VU 4: Riparian forest. VU 5: Well-drained lowland forest -a ‘‘hot’’(Lauraceae, Chamaerops, Engelhardioideae, Olea ); -b ‘‘temperate’’ (Castanea , Carpinus , Tilia) including levee forests. VU 6: Well-drained upland forest (-a Quercus drymeja-mediterranea ; -b Fagus-Cathaya ). VU 7: Well-drained (lowland and) upland conifer forest including hammocks and raised bogs within peat-forming vegetation. L = leaves; P = palynomorph; R = reproductive structures 1) Velitzelos & Schneider, 1979; 2) Velitzelos & Petrescu, 1981; 3) Velitzelos et al., 1983; 4) Velitzelos & Gregor, 1985; 5) Mai & Velitzelos, 1992; 6) Mai & Velitzelos, 1997; 7) Velitzelos & Kvaček, 1999; 8) Kvaček et al., 2002; 9) Denk & Velitzelos, 2002; 10) Velitzelos & Denk, 2002; 11) this study Table 2. Estimated climate parameters for the pre-evaporitic Messinian of Vegora from two CLAMP calibration datasets and from CA. Climate parameter CLAMP CLAMP CA modified CA modified Physg3arcAZ PhysgAsia1 10-90%iles MAT (°C) 10–13.5 8.7–11.5 8.6–21.2 9.9–18.4 CMMT (°C) 1–5 -2.7–2.3 ≥ 1.2 WMMT (°C) 19.2–22.8 19–22.6 GROWSEAS (months) 6–8 5.5–7 MMGSP (mm) 110–160 100–160 Three_WET (mm) 500–780 400–750 Three_DRY (mm) 180–260 80–220 3_WET/3_DRY <4 < 5.5 MAT = mean annual temperature, CMMT = coldest month mean temperature, WMMT = warmest month mean temperature, GROWSEAS = duration of growing season, MMGSP = mean month growing season precipitation, Three_WET = precipitation of three consecutive wettest months, Three_DRY = precipitation of three consecutive driest months. Table 3. Vegetation types recognised for the pre-evaporitic Messinian of the Florina–Ptolemais–Servia Basin. a Vegetation type Swamp forest Main (and accessory) taxon/taxa Taxodium , Glyptostrobus Biome NLD Vegetation unit(s) VU3 Swamp forest Riparian forest BLD BLD VU3 VU4 BLD VU5b Well-drained forest Alnus , (Sassafras ) Pterocarya , Zelkova , Ulmus , (Sassafras ) Quercus kubinyi , Q. pseudocastanea , (Carpinus , Tilia etc.) Fagus , (Quercus pseudocastanea ) BLD VU5b Well-drained forest Fagus , Abies , Cedrus , Cathaya MIXED VU6b Well-drained laurophyllous forest Quercus drymeja , (Q. sosnowsky ) Quercus mediterranea , Chamaerops , Olea Poaceae BLE VU6a Well-drained forest Well-drained sclerophyllous forests/shrublands [?]c Grassland-steppe forest b Modern (Neogene) analogue Taxodium swamp forests SE USA; (Taxodium/Glyptostrobus swamp forests widespread in N Hemisphere Neogene) Alnus swamp forest Riparian and alluvial forest of Georgia and Iran Lowland oak-hornbeam forests; ("Quercetum mixtum") Lowland beech forests of N Turkey, Georgia, N Iran; ("Fagetum gussonii") Montane Fagus-Abies forest, montane Fagus-Cedrus-Pinus forest; Abant Gölü; Erbaa-Çatalan Quercus dilatata association (with Taxus , Pinus , Acer etc.) Mediterranean sclerophyllous forest/shrublands Forest-steppe of SE Europa to Afghanistan BLE/ VU0 SHRUBLAND GRASSLAND/ VU0 SHRUBLAND 1) Mai, 1995; 2) Dolezych & Schneider, 2007; 3) Denk et al., 2001; 4) Akhani et al., 2010; 5) Maharramova, 2015; 6) Kozlowski et al., 2018; 7) Kvaček et al., 2002; 8) Mayer & Aksoy, 1986; 9) Akkemik, 2003; 10) van Zeist & Bottema, 1991; 11) Freitag, 1971; 12) Erdős et al., 2018 a Biome classification follows the phsiognomic approach of Woodward et al., 2004. bVegetation units as in Table 1. c [?] expresses the uncertainty around a possible extra-regional signal in the Vegora pollen record. According to Erdös et al. (2018) steppe forest with Stipa and other grasses and different species of Quercus (forest-steppes of the type 'Region A - SE Europe') is characterized by MAP of 420-600 mm; this would be much drier than the inferred MAP for the FPS. References 1, 2 3, 4 1, 3, 4, 5, 6 4, 7 3, 4, 7 8, 9, 10 11 10 8, 10, 11, 12