The middle Miocene palynofloras of the Salihpaşalar lignite mine (Yatağan Basin, southwest Anatolia): environmental characterisation and comparison with palynofloras from adjacent basins

As the third part of an ongoing investigation of middle Miocene palynofloras in the Yatağan Basin (YB), southwestern Anatolia, the palynofloras of the Salihpaşalar lignite mine in the main YB were studied. Seven types of algal spores, aplanospores/zygospores or cysts, six types of lycophyte and fern spores, 12 types of gymnosperm pollen and 90 types of angiosperm pollen were identified. Of a total of ca. 140 plant taxa described from the YB, over 10% are confined to the Salihpaşalar assemblage. Differences between coeval palynofloras of the Sekköy Member might reflect changing or prograding depositional environments. A number of rare accessorial taxa reflect these local differences: Pilularia, Valeriana, Drosera and Persicaria aff. amphibia only occur at Salihpaşalar and are typical of shallow water or temporary ponds associated with a lake shore. Apart from this, all the palynofloras, originating from the lignite seams and overlying limnic limestones (uppermost Turgut and Sekköy Member), of the YB are strongly indicative of extensive woody vegetation with a dominance of diverse Fagaceae and Pinaceae. In addition, a list comparing the well-documented YB palynomorphs to morphologically similar palynomorphs of published late early to middle Miocene plant assemblages of western Anatolian was compiled. Such a comparison reveals that in many instances different taxon names have been used to denote the same taxa. Hence, resolving these synonymies is a prerequisite of any meaningful comparison of palynofloras in the region.


Introduction
High tectonic activity in western Anatolia during the late Cenozoic associated with the collision of the Arabian and Anatolian plates resulted in orogeny, volcanic activity and basin formation (Rögl 1999;Popov et al. 2004; van Hinsbergen and Schmid, 2012). In southwestern Anatolia, the Denizli, Söke and Yatağan Basin (YB) among others formed with the onset of the Miocene subsidence (Alçiçek 2010). Age correlation of these Miocene fault-bounded basins is complicated and previously suggested ages for individual basins based on different proxies have been highly controversial. Specifically, it has been demonstrated that palynological data alone are not sufficient to reliably infer the age of a particular basin . Examples from western Anatolia for conflicting ages inferred from palynological data and radiometric dating include the Soma Basin, for which Akgün et al. (2007) suggested a middle Miocene age whereas radiometric dates from Ersoy et al. (2014) suggest an early Burdigalian age, the Gördes Basin, for which Akgün et al. (2007) suggested an early to middle Serravallian age, while Purvis et al. (2005) based on radiometric dates inferred a late Aquitanian to early Burdigalian age based on radiometric dates, and the Bigadiç Basin, for which Akyol and Akgün (1990) inferred a middle to earliest late Miocene age, while Erkül et al. (2005) suggested an early Burdigalian age, based, again, on radiometric dating. Complementary data, such as vertebrate fossils or radiometrically dated layers, are hence of utmost importance for reliably inferring the age of a fossil locality.
Another issue that hampers the comparison of Neogene sedimentary basins in western Anatolia relates to the two concepts applied for the naming of dispersed fossil pollen and spores. A substantial number of studies have been using the artificial fossil nomenclature (strictly morphology based) for palynomorphs, while in more recent publications, conventional botanical names are used more often. Therefore, it maybe difficult to compare studies for a particular area, potentially leading to highly redundant taxon lists.
The present contribution is the third in a series of papers dealing with the palynology of middle Miocene (MN6-MN7/ 8) deposits of the upper Turgut and Sekköy Member of the Eskihisar Formation in southwestern Anatolia . Previous palynological studies have concentrated on the Eskihisar and the Tınaz lignite mines and vertebrate localities (e.g. Benda 1971;Gemici et al. 1990;Erdei et al. 2002;Akgün et al. 2007;Yavuz-Işık et al. 2011;Koç, 2017;Bouchal et al. 2016Bouchal et al. , 2017. The present investigation focuses on the Salihpaşalar lignite mine located to the south of the main YB that was exploited for a short period between 2012 and 2015. The sections studied in previous papers and in the present paper are strongly correlated by the palynological content of the lignite seams overlying limestones and their age is wellconstrained by vertebrate fossils recently identified from the main coal seam of the Eskihisar lignite mine (electronic supplementary material in Bouchal et al., 2017) and the fossil vertebrate locality Yeni Eskihisar (Sickenberg 1975;Saraç 2003;Wessels 2009;The NOW Community 2018).
In this paper, I will (i) describe palynological assemblages from a stratigraphic column at Salihpaşalar using a combined scanning electron and light microscopy approach, (ii) compare the palynological records of three lignite mines in the YB, and (iii) discuss the palaeoenvironmental signal of the palynological data from the YB palynofloras. In addition, a list comparing the well-documented YB palynomorphs to morphologically similar pollen of other late early to middle Miocene palynoassemblages of western Anatolian strata was compiled [Supplementary Material (S) 2]. This is an attempt to accommodate previous concepts of palynological nomenclature (strictly morphology based) and modern concepts of applying botanical names to palynomorphs for this region.

Geological setting
The YB, Muğla, southwestern Turkey, is a southeast-trending graben, up to 50 km long and 15 km wide (Fig. 1a, b). The basin is divided by an axial bedrock horst (metamorphic cover rocks of the Menderes Massif) in the smaller (ca 12 km long, 1 km wide, south east trending) Tınaz subbasin (Fig. 1b) and the main YB. The northern area of the main basin between Eskihisar and Turgut (west and north of Aladağ Mountain) is segregated from the southern part by a bedrock sill (Fig. 1b;Atalay 1980;Becker-Platen 1970). The Neogene basin fill is up to 600 m thick and grouped into the Eskihisar Formation (early to middle Miocene), the Yatağan Formation (late Miocene to early Pliocene), and the Milet Formation (middle to late Pliocene). The Eskihisar Formation is subdivided in Turgut Member (reddened alluvial-fan deposits followed by fluviatile deposits and lignites) and Sekköy Member (limnic marls and limestones; Fig. 1c). For the present study, the type profiles of the Turgut and Sekköy members as designated by Becker-Platen (1970, p. 22 ff. and pp. 26-27) were used as reference points and the boundary placed between the two members where the change from siliciclastic (noncarbonatic) to carbonate sedimentary rocks occurs. Five lignite fields (Eskihisar field, Tınaz field, Bağyaka field, Bayir field, Turgut field) have been prospected in the YB. All economically exploited lignite seams of the YB are confined to the transition zone of the Turgut and Sekköy members of the Eskihisar Formation (Becker-Platen 1970;Inaner et al. 2008).

Age of the Salihpaşalar lignite mine section
The stratigraphic chart of the mediterranean and parathetyan realm has been going through several changes since the 1970s; for stratigraphic ages and ranges of Becker-Platen (1970), the updated corresponding age is provided based on the compiled stratigraphic chart of Neubauer et al. (2015). In the overview map of the YB, Becker-Platen (1970, pl. 3) assigned an area (ca 2 km long, 0.5 km wide) east of the village of Salihpaşalar to the Sekköy Member (=Sekköy Formation in Becker-Platen). Additionally, the following age constraining fossils are listed (sample localities P419, P431, P434 in Becker-Platen 1970): pulmonate gastropods-Clausilla sp.; Ostracoda-Cytheromorpha zinndorfi, Loxoconcha granosa, Hemicythere folliculosa, Cyprideis cf. littoralis → Helvet to Pannon (=middle to late Miocene age); pharyngeal teeth of cyprinids-"Cypridopsis" biplanata → Torton (=early late Miocene). Atalay (1980) reported two vertebrate fossil localities NEE of the village Salihpaşalar (Salihpaşalar-Kemikalan, Salihpaşalar-Karaağaç; both assigned to MN12, late Tortonian to early Messinian; The NOW Community 2018) and assigned the rocks of this area to the late Miocene Yatağan Formation. So far, the stratigraphic relevant animal taxa of the Salihpaşalar lignite mine have not been investigated (Serdar Mayda, personal communication 2017). The here documented palynoflora corresponds to previously reported pollen floras and leaf assemblages of the YB lignite mines (Tables 1 and 2; Bouchal et al. 2016Bouchal et al. , 2017Güner et al. 2017); for the macro and micro floras, an age of MN6 to MN7/8 is proposed due to their lithostratigraphic position between the vertebrate locality Eskihisar Gallery (MN6) and the younger Yeni Eskihisar locality [MN(7)/8] Wessels 2009). Tuff layers from the upper Sekköy Member (Yeni Eskihisar II mammal locality) produced radiometric dates of 13.2 Ma ± 0.35 and 11.2 ± 0.2 (Becker-Platen et al. 1977).

Santalaceae
Arceuthobium sp. This study S153637 Typhaceae field is up to 18 m thick and split by several faults, leading to highly varying thickness of overburden (5-15 m at  Salihpaşalar lignite mine, up to 493 m in other parts of the  lignite field, see table 3 in Inaner et al. 2008). Three further opencast lignite mines are currently active in the YB, (i) the Eskihisar lignite mine located west of the Aladağ Mountain exploiting the Eskihisar and Turgut lignite fields; (ii) the Tınaz lignite mine, located in the southern part of the Tınaz subbasin; and (iii) the Bağyaka lignite mine, located in the northen part of the Tınaz subbasin (Fig. 1b). In recent years, the plant macrofossils and palynogical assemblages of the Eskihisar and Tınaz lignite mines have been the focus of a rather high number of studies (Akgün et al. 2007;Benda 1971;Erdei et al. 2002;Gemici et al. 1990;Güner et al. 2017;Koç, 2017;Yavuz-Işık et al. 2011).

Sampled stratigraphic section
A stratigraphic column comprising 25 m (hereafter Salihpaşalar section; Fig. 1c) was sampled at 1-m intervals (mainly siliciclastic and marly sedimentary rocks alternated by lignite seams). Twenty-seven samples were collected of which 16 were suitable for palynological analysis (detailed information on samples is provided in S 1). The base of the Salihpaşalar section starts below the excavated lignite seam (uppermost part of the Turgut Member) with weakly consolidated, blueish-grey to dark grey, micaceous clayey siltstones. They are followed by a section of thin lignite seams (2-20 cm) intercalated with silty claystones (1-10 cm); indeterminable plant debris is common in these layers. This section is succeeded by the main lignite layers, which have a combined thickness of three to four meters. Its lower part is intercalated with thin clayey siltstones (> 1 cm), which are replaced in the upper part by light-grey to blueish-grey marls and clayey limestones (> 1 cm). The succeeding section comprises 19 m of grey clayey limestones and blueish-grey marls (0.01-2 m) alternating with lignites (1-50 cm). Leaf and root fossils as well as gastropods and mollusc shell debris occur in this part of the section but are rare or restricted to single horizons. The following yellowish-grey to yellowish-red clayey limestones were not further investigated or sampled. In 2015, the mining activity ceased and the pit was flooded with groundwater; therefore, the profile is no longer accessible.

Sample processing and data
Sedimentary rock samples were processed following the protocol described in Grímsson et al. (2008) and the same pollen grains were investigated using LM and SEM (single grain method, Zetter 1989). LM photographs of dispersed fossil pollen were taken with an Olympus BX51          Specimens were sputter coated with gold and photographed using a Hitachi S-4300 cold field emission scanning elect r o n m i c r o s c o p e . F o r t h e p o l l e n d i a g r a m , 4 0 0 palynomorphs were counted per sample. The terminology for pollen morphology follows mostly Punt et al. (2007) and Hesse et al. (2009). The pollen diagram ( Fig. 2) was generated in C2 vers. 1.7.6, maps and sections were drawn in Adobe Illustrator 15.0.0 and photographs were cropped in Adobe Photoshop 12.0. Sediment samples, processed samples and SEM stubs are stored at the Swedish Museum of Natural History, Stockholm, under accession numbers S153622 to S153648. Table 1 lists all taxa identified for and their occurences through the Salihpaşalar section. A number of taxa have previously been described, discussed and figured from adjacent localities and coeval strata of the upper Turgut and lower Sekköy members . These taxa, although from different lignite mines of the YB, are not described here. Light microscopy and SEM documentation of all palynotaxa is provided in S 3.
Remarks: Non-papillate pollen is commonly produced by Cupressaceae. No further allocation is possible owing to poor preservation and general lack of diagnostic morphological characters of the available specimens (compare Van Campo-Duplan 1951, 1953Kedves 1985). Spindle-shaped non-papillate, ruptured pollen ( Fig. 5a and S 3- Fig. 3a 1988). Pollen of this type commonly ruptures from proximal to the papillate distal pole during the hydration process preceding pollen germination (Southworth 1988) leading to its characteristic shape (Fig. 5c). Owing to this deformed state, no further generic determination is recommended.
Remarks: Euphorbia sp. 2 corresponds in size and sunken colpi to type 2-Euphorbia nutans of El-Ghazaly and Chaudhary (1993), which is produced by several not closely related Euphorbia species.
Betula sp. 2 (Fig. 7m, n and S 3-Fig. 7d- Genus Carpinus L., 1753 Carpinus sp. Genus Corylus L., 1753 Corylus sp. (Fig. 7p and S 3-Fig. 7j-l) Remarks: Dispersed and poorly preserved pollen of Corylus, Ostrya and Morella vel Myrica is difficult to unambiguously assign to one of these genera using only LM, owing to a high degree of morphological similarities (cf. Edwards 1981). Therefore, these three taxa were combined in the pollen diagram in the category Corylus/Ostrya/Morella vel Myrica. Genus Quercus L., 1753.

Remarks:
The morphological characteristics of this pollen type are shared by a number of families within the Sapindales (Erdtman 1952), and hence determination down to the family level is not possible.
Remarks: Amaranthaceae gen. indet. 3 differs by its smaller size and operculi with a higher number of nanoechini from Amaranthaceae gen. indet. 1 and 2. Strong morphological similarities between the Ranunculaceae Thalictrum L. (Clarke et al. 1991) and the fossil species Thalictrumpollis thalictroides Stuchlik (Stuchlik et al. 2009) can be observed using LM. Using SEM Amaranthaceae pollen reveals nanoechinate and perforate exine sculpturing, whereas Thalictrum lacks perforations and its nanoechini are more evenly spaced (compare Clarke et al. 1991).
Remarks: Caryophyllaceae gen. indet. 1 and 2 were combined in the pollen diagram. Caryophyllaceae gen. indet. 2 differs from Caryophyllaceae gen. indet. 1 by smaller pores, smaller perforations and fewer apertures. Pollen of this family has been reported previously from YB (Table 2).
Remarks: Extant pollen of Drosera rotundifolia L. (Halbritter et al. 2012) and the D. rotundifolia type of Punt et al. (2003) correspond the Salihpaşalar specimen by the presence of echini and microechini.
Order Asterales Link, 1829 Family Asteraceae Berchthold et Presl, 1820 Subfamily Asteroideae Lindl Asteroideae type 1 ( Fig. 11g and S 3-Fig. 14d-f) Remarks: For detailed descriptions and remarks, see Bouchal et al. (2017, p. 25, as Asteroideae type 1). Asteroideae types 1 to 5 clearly belong to this subfamily (compare Punt and Hoen 2009); due to poor preservation (folded, broken, deformed grains), the encountered echinate pollen cannot be assigned to particular genera. The here used types are purely artificial and of no taxonomic value.
Artemisia and Asteroideae types 1 to 5 were combined in the pollen diagram.
Remarks: Asteroideae type 3 differs by thinner exine and small, more widely spaced echini from the remaining four Asteroideae pollen types.
Remarks: The available specimen corresponds by its exine architecture and microechinae situated on weak verrucae to the Valeriana officinalis type of Clarke and Jones (1977) and the Valeriana pollen type I of Clarke (1978). Previous reports of Valerianoideae pollen from middle Miocene localities of western Turkey: Valerianaceae (Yavuz-Işık 2007;Yavuz-Işık et al. 2011).
Remarks: Apiaceae types 1 to 6 were combined in the pollen diagram. For the morphological description of Apiaceae type pollen, the nomenclature of Punt (1984) is used. Apiaceae type 1 has strong morphological similarities (colpus length, exine thickened in mesocolpium) to the Conium type of Beug (2004), with closest similarities to Falcaria vulgaris Bernh. (Beug, 2004, plate 23, figs 19-20), and the F. vulgaris type of Punt (1984) but differs by its smaller size.
Remarks: Apiaceae type 4 is poorly preserved, its dimensions (size, colpus length) are similar with Apiaceae type 3 but differences in exine sculpturing set this type apart from cooccurring pollen types.
Remarks: Apiaceae type 6 has similar dimensions (size, colpus length) with Apiaceae type 5 but differs in its exine sculpturing.

Incerta sedis
Remarks: Pollen of uncertain determination was not included in the pollen diagram.

Discussion
Palynoflora and pollen zones of the Salihpaşalar lignite mine section The most frequently observed palynomorphs of the Salihpaşalar section belong to Fagaceae (14.5-47.5%) and bisaccate gymnosperm pollen (8.75-34.5%); through the entire section arboreal pollen (AP) (73-94%) is distinctly more frequent than non-arboreal pollen (NAP). The composition of the palynofloras of the Salihpaşalar section is rather homogenous, but two pollen zones can be distinguished. Pollen Zone (PZ) 1 is characterised by a higher abundance of spores (7.75-14%) and azonal elements (Decodon 4.5-12.75%) and has previously been reported from the Eskihisar and Tinaz mine sections by Bouchal et al. (2016Bouchal et al. ( , 2017. At the Salihpaşalar mine, PZ1 is restricted to the lower part of the sampled section (S153622-S15327) comprising the thickest lignite seam, its intercalated sediments and the underlying micaceous siltstones. PZ 1 mainly concurs in its composition with PZ 2, differing only in the presence of Lycopodium, Lycopodiella, Selaginella, Pilularia, Tsuga and Drosera.
PZ 2 is confined to the upper part of the section overlying the main lignite seam consisting of clayey limestones and marls intercalated by thin lignite beds (samples S15329-S15346). In this zone, occasional abundance spikes of algal cysts (> 20%, in S15329 and S15339) and Apiaceae (9.5% in S15338) are recorded and spores are less frequent (0.5-2.75%). Pollen of Buxus, papillate Cupressaceae, Ilex, Persicaria, Sapotaceae, Valerianoideae and Viburnum have only been encountered in PZ 2.
Comparing the palynoassemblages of the YB Palynofloras from different strata and localities in the YB, most of them belonging to the Turgut and Sekköy members, have been the focus of a high number of studies (this study ;Nakoman 1967;Benda 1971;Gemici et al. 1990;Erdei et al. 2002;Jiménez-Moreno 2005;Akgün et al. 2007;Yavuz-Işık et al. 2011;Koç, 2017;Bouchal et al. 2016Bouchal et al. , 2017. Most of these studies agree that Fagaceae and Pinaceae are the most taxonomically diverse and abundant families in the palynofloras of the Turgut and Sekköy members (Table 2). This study and Bouchal et al. (2016Bouchal et al. ( , 2017 distinguished four pollen zones for sections of the YB lignite mines (Eskihisar, Tınaz, Salihpaşalar): (i) PZ 1 is mainly dominated by fern spores (S 4, abundances ranging from < 10-60%) and azonal elements (e.g. Alnus) and is restricted to the lignite seams and intercalated and underlying layers. Hence, in this PZ, a strong facies signal is seen rather than a stratigraphical one. Gemici et al. (1990 , Table 1) report fern spore abundances of > 80% and Yavuz-Işık et al. (2011, samples E in fig. 2; spores were not investigated in this study) report Alnus pollen values of < 40% for their samples from Eskihisar; this would possibly place these samples in PZ 1. (ii) In PZ 2, zonal AP taxa are most diverse and abundant (e.g. Pinaceae, Fagaceae, Juglandaceae, Ulmaceae) while the herbaceous component is diverse but of low abundance. In the Eskihisar, Tınaz and Salihpaşalar mines, this PZ is present in the first 10-20 m of marls and limestones overlying the lignite seams. A similar signal is reported from samples from Çatakağyaka (ca 12 km southeast of Tınaz) by Jiménez-Moreno (2005, samples 1 and 2 in fig. 4.77) and Yavuz-Işık et al. (2011, samples Ç in fig. 3). Bouchal et al. (2017) reported a (iii) transitional zone between PZ 2 and PZ 3 for their Tınaz section. Here, the ratio between AP and non-AP fluctuates strongly and diversity and abundance of herbaceous taxa increases while AP taxa decrease (S 4). Possibly, the upper part of the Eskihisar section of Bouchal et al. (2016, samples S153582-S153593 in fig. 2) could belong to this PZ, because of an abundance peak of Amaranthaceae and Apiaceae. Similar increased abundance of herbaceous taxa has been reported from Çatakbağyaka (< 25% NAP, Jiménez-Moreno 2005, sample 3 in fig. 4.77) and Yeni Eskihisar 1 (< 30%, Yavuz-Işık et al. 2011, sample Y in fig. 3). It is important to note that these similarities reflect similar facies but not necessarily same age. (iv) In PZ 3, present in a single sample from the Tınaz section of Bouchal et al. (2017), AP abundance and diversity have further declined and a value of < 20% of AP is reported (S 4).
Noteworthy is also the presence/absence of some of these accessorial elements. While documented by Gemici et al. (1990), the presence of the rare and distinct Sapotaceae pollen type could not be verified in samples of the Eskihiar mine investigated by Benda (1971, "Eskihisar Pollenbild") and Bouchal et al. (2016), but single ocurrences are documented for Tınaz and Salihpaşalar (this study; Bouchal et al. 2017). Another example is Buxus pollen, which is absent from the Tınaz section but documented for Eskihisar and Salihpaşalar, while leaf remains of Buxus pliocenica Saporta et Marion are reported from Tınaz and Eskihisar Güner et al. 2017).

Depositional environment
The alternating 19-m sequence of pelitic sediment (limestones and marls) intercalated with lignites at the Salihpaşalar lignite mine shows no signs of deltaic sedimentation (absence of channelfills and coarse-grained sediments). The formation of lignite at the margin of the YB was most likely controlled by water table fluctuations within the palaeo-lake, either by transient subsidence of the basin or change in fluvial water supply. During a deepening phase (more subsidence/higher water input) of the palaeo-lake, the pellitic limestones and marls were deposited; coal swamps probably developed when the water table receded to a level that still provided a water-saturated lake margin. In the first few centimetres (1-5 cm) of limestone or marl beds following a lignite bed, poorly preserved monocotyledon root horizons (affinities to Phragmites or Typha) were observed. These probably represent the development of a reed belt that developed in response to the flooding of the shallow, swampy palaeo-lake margin areas. The overlying marls and pellitic limestones do not contain such root horizons, indicating a further deepening of the palaeo-lake. Summarising, the deposits of the Salihpaşalar section accumulated in a marginal part of the Yatağan palaeo-lake. No clastic payload incoming from a fluviatile system affected the sedimentation. Whether cosmic cyclicities or subsidence did lead to the alternating lignite and limestone pattern of the Salihpaşalar section has not been the focus of this study. Future geological investigations on more complete sections (e.g. drill cores) of the main YB could help resolve this aspect.
Regional vegetation inferred from the palynofloras of the YB Table 2 lists cyst, spore and pollen taxa reported from the YB palynofloras. The ecological properties of their potential modern analogues are indicated and assigned to vegetation units (VU). Only very few taxa (e.g. Engelhardioideae p.p., Sapotaceae) are today restricted to lowland forests in warm temperate to tropical climates (VU 5a). Also Carpinus, Ostrya and Tilia are nowadays commonly found in lowland forests of temperate regions. The few lianas reported here may also have been part of lowland forests (both well-drained and swamp forests). The majority of tree species reported from the YB palynofloras grow in well-drained lowland and upland forests (VU 5b-7) and comprise both angiosperms and conifers. Among conifers, it is noteworthy that coal-forming taxodiaceous taxa are not well-represented suggesting that other plants contributed to the coal in the Yatağan Basin. In contrast, Pinaceae were fairly diverse. Although modern distributions of taxa such as Cathaya, Cedrus or Picea are typically confined to high elevations, the mass occurrence of Cathaya and Cedrus leafy shoots, cones and cone scales in late Miocene deposits of northern Greece (Velitzelos et al. 2014) suggests that these elements could also have grown at lower elevation. Similar patterns of shifting of ecological niches are seen in ancestors of mountain goats (Solounias and Moelleken 1992). Among angiosperms, trees and shrubs include both deciduous and evergreen taxa and may have contributed to different types of forest vegetation. Fagaceae (Fagus, Quercus sect. Cerris, Q. sect. Ilex) most likely were the dominating forest elements in the surroundings of the YB.
Among herbaceous plants, a great number might have been associated with open lake shore vegetation including areas that temporarily fell dry. Typha and Typha-allies may have formed a reed belt in some places. Only few herbaceous taxa (e.g. the ferns Pteris and Osmunda) are characteristic elements of forests.
In summary, the regional vegetation as inferred from the extensive palynological record of the YB floras consisted of different forest types, including broadleaved thermophilous lowland forests, temperate lowland and upland forests as well as mixed broadleaved and conifer forests at mid and high elevations. In addition, rich riparian forests were present around the lake and rivers. Coal-forming swamp forests may have included taxodiaceous conifers and angiosperms such as Myrica, Alnus and Decodon. Open vegetation was restricted to disturbed and shore areas.
In this and two previous studies , a combined LM and SEM approach was used to achieve best possible botanical determination for palynomorphs for three coeval sections in the YB. At the same time, it was noticed that in the published literature for the same fossil-taxon, a variety of botanical affinities had been indicated by different authors (S 2). Using the well-documented (LM and SEM) studies on YB lignite mine sections (this study; Bouchal et al. 2016Bouchal et al. , 2017 as a taxonomical and morphological baseline, I compared 24 published accounts on (preferably figured) dispersed palynomorphs from early to middle Miocene deposits of western Anatolia. Although the palynological record of this time span and region likely is more diverse and complex than that of three sections of a single basin, a fairly high degree of homogeneity in composition (not in abundance) at regional scale has previously been noted by other authors (Benda 1971;Takahashi and Jux 1991;Akgün et al. 2007). In Table S 2, both botanical and abotanical taxa were compiled (i) to visualise the various palynological concepts used for this region since the 1960ies; (ii) to discriminate well-documented (common) taxa from those that so far lack documentation; and (iii) to narrow down the list of over 400 names that have been used in reviewed literature to the ca 170 taxa that are botanically (using LM and SEM) recognisable for the YB. The resulting redundancies of names are highly misleading when discussing past biodiversity but may also affect taxonomic interpretation and as such the inference of palaeoenvironments. Finally, (iv) they were compiled to provide a basis for further scientific discussion. In the following, I provide a few examples to illustrate why a standardised palynomorph nomenclature with updated botanical affinities will improve regional environmental comparisons and palaeoecological inference.
The name Laevigatosporites haardti is commonly used for psilate monolete spores. This stratigraphically widespread and rather featureless spore type is found in several extant fern families (e.g. Aspleniaceae, Davalliaceae, Dryopteridaceae, Gleicheniaceae, Lomariopsidaceae, Oleandraceae, Thelyp teridaceae, Vittariaceae, Polypodiaceae; see Tyron and Lugardon 1991) but has mainly been associated with Polypodiaceae in the reviewed literature (S 2). If using this name, it should be clearly stated that the exact botanical affinities of this taxon cannot be established using spore morphology.
The former Taxodiaceae, "Taxodiaceae pollen", commonly used as synonymous of papillate cupressaceaeous pollen, are currently included within the Cupressaceae and its former members have been assigned to the subfamilies Taiwanioideae, Athrotaxoideae, Sequoioideae and Taxodioideae (Earle 2010;Farjon 2005). Several studies demonstrated the limited taxonomic value of papillate and non-papillate Cupressaceae pollen (Kedves 1985;Bortenschlager 1990;Grimm et al. 2016). For identification to subfamily level, LM and SEM investigation of exceptionally well-preserved and complete fossil material (3D preservation) is needed (Grímsson and Zetter 2011); such type of preservation has not been reported for any of the reviewed pollen floras of western Anatolia. Since modern members of the Cupressaceae are ecologically very diverse, overly specific (to genus level) identifications run the risk of great error when inferring palaeoenvironments.
The botanical affinity of the fossil species Pityopollenites libellus (R.Potonié) Nakoman has been assumed to be with the mainly southern hemispheric family Podocarpaceae or its genus Podocarpus (S 2). Under SEM, all YB specimens falling within the morphological range of this fossil-taxon show unequivocal characteristics (microechini) of the Pinaceae Cathaya . Thus, numerous reports of Podocarpus for the Neogene of western Anatolia would need to be verified or rejected by SEM investigations before any biogeographic or palaeoenvironmental conclusions can be drawn. Tricolporopollenites megaexactus, a fossil species affiliated to Cyrillaceae, is commonly reported and figured in the reviewed literature (S 2). Combined LM and SEM investigation demonstrates that this fossil species shares morphological characteristics (size, shape, aperture) with pollen of the Lythraceae genus Decodon. A first, revised record of Decodon for Turkey shows a stratigraphic range, at least, from middle Miocene to early Pleistocene (this study; Bouchal et al. 2016Bouchal et al. , 2017Alçiçek et al. 2017).
Likewise, dispersed pollen in Neogene western Anatolian strata has frequently been determined as Castanea, Castanopsis and/or Lithocarpus or assigned to the fossil species Tricolporopollenites cingulum, T. cingulum oviformis or T. oviformis for which affinities with Castanopsis and Lithocarpus were assumed. However, pollen of extant Castanoideae is strikingly uniform and not sufficiently distinct in LM and SEM to distinguish different genera of this polyphyletic group (see, e.g. Praglowski 1984). Furthermore, poorly preserved pollen of an extinct Castanoideae, Trigonobalanopsis, is indistinguishable in LM from extant castanoids and can only be securely determined using SEM. Trigonobalanopsis is not uncommon in YB palynofloras and hence may be hidden behind some of the names mentioned above from other localities of western Anatolia. Hence, the great diversity of castanoid Fagaceae implied by the reviewed palynological literature is highly misleading. For the reviewed LM investigations, only the botanical affiliation Castaneoideae gen. indet. ought to be used for such palynomorphs.
These are but a few examples out of many which illustrate the urgent need of a catalogue of palynomorphs from Neogene strata of western Anatolia, standardising nomenclature and taxonomic affinities and documenting cysts, spores and pollen using LM and SEM. The skills, knowledge, experience and archives of all palynologists currently working in this geographical region would need to be joined to achieve this task.