136Vegetation Ecology and DiversityVED3033-1447Italian Society of Vegetation Science (SISV)10.3897/ved.177153177153Research ArticleAngiospermaePlant Ecology and SynecologySyntaxonomy and NomenclatureNumerical classification and syntaxonomical revision of the Belgrade Forest vegetation (Istanbul, Türkiye)ÇobanSüleymanhttps://orcid.org/0000-0003-1570-97951BayraktarSelimsbayraktar@iuc.edu.trhttps://orcid.org/0000-0002-8023-41352ŞahinNilüferhttps://orcid.org/0000-0003-0342-72723Department of Forest Engineering, Faculty of Forestry, Istanbul University-Cerrahpaşa, Istanbul, TürkiyeDepartment of Forest Engineering, Faculty of Forestry, Istanbul University-CerrahpaşaIstanbulTürkiyehttps://ror.org/01dzn5f42Department of Landscape Architecture, Faculty of Forestry, Istanbul University-Cerrahpaşa, Istanbul, TürkiyeDepartment of Landscape Architecture, Faculty of Forestry, Istanbul University-CerrahpaşaIstanbulTürkiyehttps://ror.org/01dzn5f42Marmara Forestry Research Institute, Fatih Forest Campus, Istanbul, TürkiyeMarmara Forestry Research InstituteIstanbulTürkiye
Corresponding author: Selim Bayraktar (sbayraktar@iuc.edu.tr)
Academic editor: Bianca Ott Andrade
20260104202663e177153E35C3B6B-A06C-548E-80FC-A94BD7C1E1120511202519032026Süleyman Çoban, Selim Bayraktar, Nilüfer ŞahinThis is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
The Belgrade Forest, located in the European part of Istanbul, represents an ecological and biogeographical transition zone where nemoral and Mediterranean elements coexist. Although comprehensive phytosociological studies have been conducted in the study area, differences in data structure, sampling intensity, and nomenclatural problems have prevented the establishment of a unified vegetation classification. In the present study, a total of 685 relevés derived from previous surveys were compiled and digitized in TURBOVEG, standardized, and analysed using numerical classification and ordination methods. The vegetation data were classified by TWINSPAN, and diagnostic species were identified based on the phi coefficient and constancy ratio. Detrended correspondence analysis (DCA) was applied to determine major compositional gradients, and Ellenberg-type indicator values were used to interpret ecological differentiation. The analysis resulted in four forest associations and one pseudomaquis community, corresponding to Alnus glutinosa-Carpinus betulus riparian forest, Acer campestre-Carpinus betulus forests on humid soils, meso-thermophilous mixed beech-oak-hornbeam forests, and pseudomaquis vegetation. Among these associations, Equiseto telmateiae-Alnetum glutinosae ass. nov. and Hyperico calycini-Ericetum arboreae ass. nov. are here described as new. The principal ecological gradient reflected a transition from humid to relatively dry site conditions, consistent with slope aspect (site exposure to solar radiation) and soil depth variations. The resulting syntaxonomic scheme provides a harmonized framework for the classification and monitoring of forest vegetation in the Belgrade Forest and offers a baseline for assessing future ecological change and restoration priorities.
Forest vegetation of Türkiye has been increasingly studied by foreign and local researchers since the 1960s (Çoban et al. 2016). The Belgrade Forest, an ecologically important remnant of Istanbul’s natural forest cover, has attracted numerous local and foreign scientists due to its rich flora and proximity to Istanbul. The first records related to the flora of this region date back to the 16th century (Belon 1553; Wheler 1862; Forskål 1775; Dallaway 1797; De Tchihatchef and De Verneuil 1853). The forest has long attracted scientific interest due to its naturalness and historical significance. During the Ottoman period, the Belgrade Forest was protected under special regulations aimed at safeguarding the city’s water resources, whose infrastructure had been developed continuously from the Byzantine through the Ottoman periods (Çolak and Kırca 2013; Kırca and Çolak 2013). There are seven reservoirs within the forest, connected to the city by a joint aqueduct (Çolak et al. 2013). Beyond its cultural heritage, the Belgrade Forest represents a key ecological refuge within the rapidly urbanized landscape of Istanbul (Bayraktar and Yener 2025).
The Belgrade Forest is located in the western Euxine part of the Euro-Siberian phytogeographical region of Türkiye. The flora of Belgrade Forest contains Balkan, Colchic, Mediterranean and Central European elements. The co-occurrence of nemoral and Mediterranean elements supports earlier findings that the forest marks the northern limit of Mediterranean influence in Northwestern Anatolia (Regel 1959). Indeed, the Quercus petraea subsp. iberica dominated forests approximate continental conditions, whereas the colline Fagus orientalis stands, which occur at the limit of the species’ natural distribution, develop under the influence of the Black Sea climate.
Previous floristic studies conducted in the Belgrade Forest have reported the presence of 19 tree species, 46 shrub species, and 316 herbaceous species. Among endemic species, Taraxacum aznavourii is a local endemic that is only known from its type locality and others are endemic to Northwestern Anatolia (Çolak 2013; Özalp 2013; Özhatay and Yüzbaşıoğlu 2013). This high floristic richness, including the presence of endemic and rare species, underscores the ecological and conservation importance of the Belgrade Forest, highlighting the need for continued monitoring and protection of its biodiversity (Bayraktar and Yener 2025). However, despite this richness, rapid urban expansion and recreational use have increased anthropogenic pressure on these ecosystems (Bayraktar et al. 2024).
The floristic composition of 12 different forest types was first studied by Yaltırık (1963) according to the Braun-Blanquet phytosociological method. Based on this vegetation data, Yaltırık et al. (1983) classified the vegetation of the Belgrade Forest into four plant communities: a) Carpinus betulus-Acer campestre, b) Fagus orientalis-Ilex aquifolia, c) Quercus petraea subsp. iberica-Lathyrus niger, d) Erica arborea-Erica verticillata communities. A second study was carried out by Yönelli (1986) who combined forest subcommunities to a higher unit named Quercus petraea subsp. iberica-Carpinus betulus which is composed of Quercus frainetto, Fagus orientalis, and typical (Castanea sativa) subcommunities. In addition, these studies differ in sampling intensity, nomenclatural issues, and the interpretation of diagnostic species, which limit comparability across datasets.
Recent years have witnessed major progress in the digitization and integration of phytosociological data, both in Türkiye and across Europe (Mucina et al. 2016). Large-scale vegetation databases now provide a comprehensive and standardized background for numerical vegetation classification (Chytrý et al. 2016; Mucina et al. 2016). In Türkiye, newly developed forest vegetation databases (Çoban 2017; Çoban and Willner 2019; Bonari et al. 2019) have greatly accelerated the classification of forest communities at higher syntaxonomical ranks such as alliance, order, and class (Kavgacı et al. 2012; Uğurlu et al. 2012; Bonari et al. 2021; Kavgacı et al. 2023). However, the harmonization of classification schemes and the resolution of inconsistencies in syntaxonomical units, particularly at the association level, remain challenging, underscoring the need for more detailed and standardized regional analyses.
Although comprehensive vegetation surveys have been conducted in this forest for decades, a unified and consistent classification has not yet been achieved. Given this long research history and the absence of an integrated syntaxonomic framework, the study aims to harmonize the historical phytosociological data. Accordingly, the main objective is to establish a digitized vegetation database for the Belgrade Forest and to develop a coherent and reproducible syntaxonomic framework.
Material and methodsStudy area
The Belgrade Forest is situated between 28°53'25"–29°00'55"E and 41°09'44"–41°14'40"N in the European part of Istanbul. Strandzha (Istranca) Mountains descend progressively southeastwards, ending on a plain around the Bosphorus where the Belgrade Forest is located (Figure 1). The region has an average elevation of 135 m. The highest point is Kartaltepe (230 m) in the north and the lowest is Kurudere (40 m) to the south. The total land area of the Belgrade Forest, including all vegetation types and the reservoirs, is 5,524 hectares, of which 475 ha are openings (OGM 2012a, 2012b, 2012c). In terms of Mayr’s silvicultural climate zone classification, the study area falls within the Castanetum-Fagetum transition zone (Yaltırık and Efe 1996). Within the study area, there are permanent streams that carry water throughout the year and seasonally flowing streams. In addition, lowland areas with a higher water table are also found. Historical water dams, which have been built since the Byzantine period and continued during the Ottoman period, contribute to the high ecological diversity of the study area.
Location of the study area. (a) geographic position of Türkiye, (b) location of Istanbul within Türkiye, and (c) extent of the Belgrade Forest, which constitutes the study area.
https://binary.pensoft.net/fig/1579121Climate
While the climate of Istanbul is submediterranean, characterized by several arid months in summer, a rapid change in climate is evident towards the Black Sea coast and the Belgrade Forest. According to the Thornthwaite climate classification, the study area is characterized by a humid, mesothermal, oceanic climate with a moderate summer water deficit (Kantarcı 1980; Özhan et al. 2005; Aydın et al. 2008; Arslangündoğdu and Yılmaz 2011). The vegetation period lasts approximately 7.5 months (230 days). Mean annual precipitation is around 1,111 mm, and the mean annual temperature is 12.7 °C (Table 1). In addition, the climate of the Belgrade Forest corresponds to the Euxine type (Aydın et al. 2008).
Climatic data from Bahçeköy meteorological station (1980–2009).
MONTHS
ANNUAL
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
Mean temperature (°C)
4.6
4.3
6.3
10.7
15.2
19.6
22.0
22.1
18.3
14.3
9.5
6.5
12.7
Highest temperature (°C)
18.3
22.2
25.4
32.9
33.9
37.3
40.7
37.9
34.1
33.1
25.2
19.9
40.7
Lowest temperature (°C)
-11.0
-11.4
-8.2
-2.2
0.3
6.1
8.2
8.0
6.4
0.2
-2.8
-7.6
-11.4
Precipitation (mm)
150
111
101
56
44
42
39
64
71
117
133
183
1111
Relative humidity (%)
81.4
78.3
76.9
75.0
77.2
76.6
77.9
78.5
78.4
80.4
80.5
80.6
78.4
Days covered by snow
3.9
5.6
2.0
-
-
-
-
-
-
-
0.1
2.1
13.7
Mean cloudy days
2.5
2.1
2.4
2.3
1.0
0.1
0.1
0.1
0.4
1.3
1.8
1.2
14.3
Wind direction
NE
NE
NE
NE
NE
NE
NE
NE
NE
NE
NE
NE
NE
Data analysis
Vegetation data sampled according to the Braun-Blanquet method were taken from Yönelli (1986), Yaltırık (1966), Çoban et al. (2019), and Şahin (2022). A total of 685 sample plots (349 species) were gathered and entered into the TURBOVEG database (Hennekens and Schaminée 2001). For consistency, species occurring in different tree layers were combined to ensure comparability among studies. In addition, tree species recorded in the shrub and herb layers were omitted. Subspecies and varieties were combined at the species level. Plants only identified at the genus level were removed.
The data were exported to the JUICE 7.1.3 software (Tichý 2002) and classified using the Two-Way Indicator Species Analysis (TWINSPAN; Hill 1979) with pseudospecies cut levels of 0, 2, 5, 10, and 20. Several alternative cut-level options were tested during preliminary analyses. The final cut-level combination was selected because it yielded the most ecologically coherent and floristically interpretable classification units for the vegetation of the Belgrade Forest. Because the historical and recent plot datasets differed in sampling design and occasionally included ecotonal relevés, a limited expert-based refinement was applied to ensure ecological coherence. Only a small number of relevés were reassigned, and these adjustments improved the ecological distinctness and internal consistency of the final vegetation units.
Diagnostic species were defined using the phi coefficient and Fisher’s exact test, the constancy ratio for understorey species, and total cover value for tree species. A species was considered diagnostic only when all criteria were simultaneously met, namely: (1) a phi coefficient ≥ 0.2 and a significant concentration of occurrences (p < 0.05, Fisher’s exact test), and (2) the constancy of the mean tree cover of the species across all plots within the target unit, at least twice as high as in any other unit of the same rank (Çoban and Willner 2019). Diagnostic species for higher syntaxonomical ranks were taken from Akman (1995) and Mucina et al. (2016). The nomenclature of the associations was revised in accordance with the International Code of Phytosociological Nomenclature (Theurillat et al. 2021).
Unconstrained ordination was applied to identify the major gradients in species composition and thus describe the general pattern in species distribution along the gradients. Vegetation data were recorded using the classical Braun-Blanquet cover–abundance scale (r, +, 1, 2, 3, 4, 5), and for Detrended Correspondence Analysis (DCA) the values were converted to percentage cover and log-transformed prior to ordination. The dataset was subjected to DCA using CANOCO 5 (ter Braak and Šmilauer 2002).
Species richness for each relevé was calculated in JUICE and used to compare floristic differences among vegetation units. Ellenberg-type indicator values were obtained from Tichý et al. (2023) and assigned to all species in the vegetation dataset. For each relevé, mean indicator values were calculated and subsequently averaged at the vegetation-unit level to compare ecological characteristics and to interpret the underlying environmental gradients. Physiographic variables (elevation and slope) were evaluated separately using boxplots to characterize and compare the topographic niche conditions of the identified vegetation units.
ResultsClassification of forest plant communities
The vegetation of the Belgrade Forest consists of forest and pseudomaquis vegetation. In addition, plantations of Pinus nigra, Pinus sylvestris, and Pinus pinaster were established on degraded shrublands and open areas surrounding the natural forest for experimental and economic purposes. In the classification process, forest vegetation was analyzed separately from pseudomaquis vegetation. The classification of forest vegetation data (654 relevés) distinguished forests on wet soils (Alnus glutinosa and Carpinus betulus-dominated) and mesophilous beech forests from meso-thermophilous hornbeam and beech forests at the first level of division. At the second level of division, riparian forests and thermophilous oak forests were separated. Mesophilous beech forest was identified at the third level of the division. Pseudomaquis vegetation, which is represented by 19 relevés in the dataset, contains evergreen species mixed with some deciduous species.
As a result of the classification 4 forest associations and 1 pseudomaquis association were distinguished (Table 2), of which 2 associations are here determined as new.
Abbreviated percentage constancy synoptic table of vegetation data. T, Tree layer; H, Herb layer. Grey: diagnostic species. Non-diagnostic species with less than 25% constancy were omitted (See Suppl. material 2 for the full table).
This riparian forest vegetation type occupies narrow corridors along streambeds. The width of these stands increases where the streambed widens on flatter terrain. The unit is characterized by the highest moisture and soil nutrient indicator values (Figure 3).
Acer campestre-Carpinus betulus forests on humid soils (Unit 2)
Carpinus betulus forms mainly mixed stands in valleys composed of Acer campestre, Alnus glutinosa, and Quercus robur. This unit occurs under humid soil conditions but has lower moisture and nutrient indicator values compared to Unit 1. Species richness is higher than in Unit 1, with a mean species number of 28.9 (Figures 2, 3).
Related syntaxa: Geranio robertiani-Carpinetum betuliKavgacı et al. 2011, Trachystemo orientalis-Carpinetum betuliKavgacı et al. 2016
Mesophilous oak, hornbeam and beech forests
Mesophilous Fagus orientalis forest (Unit 3)
Pure Fagus orientalis forests occur only on northern exposures (N, NE, NW) with deep soils. Fagus orientalis dominates the tree layer with high total cover, accompanied by Carpinus betulus. The main difference from other communities is that oaks are not present in their structure. The unit has higher moisture and soil nutrient indicator values but lower light indicator values compared to the oak-dominated unit (Unit 4–5) (Figure 3). Species richness (mean species number 16.9) is lowest in this unit due to dense tree cover (Figure 2). For this reason, this unit is represented by few diagnostic species (Table 2).
Related syntaxa: Cyclamini coi-Fagetum orientalis Tzonev at al. 2006, Ilici colchicae-Fagetum orientalisQuézel et al. 1992, Rubo hirti-Fagetum orientalisÖzen 2010
This unit is dominated by Fagus orientalis, Quercus petraea, Carpinus betulus, Quercus frainetto, and Quercus cerris in the tree layer and generally occurs on sunny exposures on relatively shallow soils. These site conditions are reflected in the mean indicator values as follows: lower moisture and soil nutrient indicator values, but higher light indicator values (Figure 3). This community forms varied stand tree composition from the edge of valley bottoms to hill ridges over a wide ecological gradient. Natural stand composition has been changed in many places (Yaltırık 1968), which makes consistent classification difficult. For this reason, the unit has a higher species richness and also many common species with degraded pseudomaquis stands (Figures 2, 4).
Related syntaxa: Salvio forsskaolei-Quercetum cerridis Akman et al. ex Quézel et al. 1992, Fago orientalis-Quercetum ibericae Altay et al. 2015
Based on the classification, this unit could be divided into two subunits, but without distinct diagnostic species. Unit 4 can be differentiated by the co-dominance of the Fagus orientalis in the tree layer and Festuca drymeja in the herb layer. Unit 5 can be differentiated by the dominance of Quercus frainetto in tree layer and Erica arborea, Galium verum, and Brachypodium pinnatum in the understorey layer that indicates drier site conditions. This difference is also reflected in the moisture and soil nutrient indicator values, as well as in the richness (Figures 2, 3) (Table 2).
Pseudomaquis community (Unit 6)
The pseudomaquis vegetation community is found in large forest openings and degraded forest areas on sunny exposures of the study area. This community has many common species with the oak forests of unit 5 with poor canopy cover. The occurrence of tree species such as Quercus frainetto and Quercus petraea in shrubby form (Yaltırık 1963) indicate forest disturbances (Table 2).
Plantation areas are dominated by Pinus nigra, Pinus sylvestris, and Pinus pinaster (Unit 7). The stands are accompanied mainly by Carpinus betulus, Pinus brutia, Quercus petraea, and several shrub species such as Erica arborea, Erica manipuliflora, Pyracantha coccinea, and Quercus coccifera, along with a range of typical herb-layer species (Table 2). Species richness is relatively low (mean species number 28.9).
Ecological differentiation of forest communities
DCA presents a high compositional diversity (gradient length is 4.79 SD units) on the first axis which represents the strongest gradient in the data. The first axis explains 8.67% of species variation which is less on other axes. The first axis is associated with an increasing moisture gradient from left to right. When the two vegetation datasets were compared, it was found that the data set of Yaltırık (1963) is more heterogeneous (5.2 SD) compared to the vegetation data set of Yönelli (1986) (3.838 SD).
In the ordination diagram, humid Alnus glutinosa and Carpinus betulus dominated forest (Units 1 and 2) and mesophilous beech forest are clearly separated from meso-thermophilous Fagus orientalis and Quercus petraea dominated forests along the first axis. In addition, similarities in species composition between pseudomaquis (Unit 6) and Quercus petraea dominated forest are evident in the ordination space (Figure 4).
Streams and topographical differences affect the distribution of forest communities in the study area. Although elevational differences are not very strong, slope and aspect are the factors contributing to shaping microclimatical conditions. For instance, Fagus orientalis dominated forest (Unit 3) occurs on steeper slopes (around 30–40 degrees) and at relatively higher elevations (Figure 5).
Niche space of vegetation units based on elevation (m) and slope degrees.
https://binary.pensoft.net/fig/1579125
The synoptic table summarizes the floristic composition and constancy values of the six vegetation units and 1 plantation forest identified in the Belgrade Forest (Table 2). Each unit exhibits distinct combinations of diagnostic, constant, and accompanying species reflecting ecological gradients and physiognomic differentiation within the forest. Tree-layer composition is dominated by Fagus orientalis, Carpinus betulus, and Quercus petraea in the mesic units, whereas Pinus nigra and Pinus pinaster prevail in the plantation stands. The herb layer reveals a clear transition from mesophilous nemoral elements (Mercurialis perennis, Melica uniflora, Brachypodium sylvaticum) toward thermophilous and acidophilous taxa (Erica arborea, Cistus creticus, Brachypodium pinnatum) in the southern and degraded sites.
Discussion
The flora of the Belgrade Forest, which contains Balkan, Colchic, Mediterranean and Central European floristic elements, does not present a uniform structure in terms of phytogeographical composition. More than half of the floristic composition belongs to European floristic elements (Yaltırık 1963). The dominant tree species are Quercus spp., which cover 75% of the total area, followed by Fagus orientalis, Carpinus betulus, and Castanea sativa. Although the tree layer has similarities with Central European forests, the understory layer, especially forest openings and meadows, has Mediterranean influence (Acatay 1943; Bornmüller 1897).
According to Aydın et al. (2008), dominant woodland types of the Belgrade Forest are Thracian mixed oakwoods and (Thracian)-Euxine beech woods. In addition to pure and mixed stands of the study area, a pseudomaquis vegetation also occurs in forest surroundings. Among the Mediterranean floristic elements, species such as Erica arborea, Erica manipuliflora, Calluna vulgaris, Arbutus unedo, Cistus salviifolius, Cistus creticus, and Spartium junceum are found in both forest openings and open areas. Their presence is influenced not only by climatic conditions but also by anthropogenic factors. Soil erosion following forest destruction caused shallow soil conditions, which allowed the development of pseudomaquis vegetation (Irmak 1940).
Previous studies differ considerably in sampling design and in the range of forest types they represent. Yaltırık (1963) sampled twelve stand types that varied in tree composition, age, structure, and canopy density, whereas Yönelli (1986) focused predominantly on Quercus petraea and Carpinus betulus stands and classified the forest into Quercus frainetto, Fagus orientalis, and typical Castanea sativa subcommunities under the Quercus petraea subsp. iberica-Carpinus betulus community. As a result, communities with narrow distributions, such as riparian Alnus glutinosa forests (Unit 1), Acer campestre-Carpinus betulus stands on humid soils (Unit 2), and Chamaecytisus pygmaeus-Erica manipuliflora pseudomaquis (Unit 6), were not adequately represented in Yönelli’s dataset. In contrast, the relevés of Yaltırık (1963) encompass most vegetation types of the Belgrade Forest. More recently, humid biotopes were described by Şahin (2022), providing complementary information. Standardization of all datasets was therefore essential to achieve a consistent and comparable classification, including the combining of vegetation layers, the removal of tree species recorded in shrub and herb layers, the exclusion of genus-level taxa and the elimination of biased plots. These procedures ensured a coherent and ecologically interpretable dataset.
The Acer campestre-Carpinus betulus community on humid soils (Unit 2) predominantly contains diagnostic species of the class Carpino-Fagetea sylvaticae. Species such as Euonymus europaeus, Corylus avellana, Sanicula europaea, Cardamine bulbifera, and Mercurialis perennis are commonly found in Fagus orientalis and Abies nordmanniana forests in Northwestern Anatolia (Çoban and Willner 2019), in Fagus sylvatica forests in Europe (Tzonev et al. 2006; Willner et al. 2017), and in Great Britain (Rodwell 1998). These species are generally absent from the Fagus orientalis forests (Unit 3) but preferentially occur in Unit 2. This floristic pattern supports assigning the community to the Carpinetalia betuli order and the class Carpino-Fagetea sylvaticae.
The Alnus glutinosa-Carpinus betulus riparian forest (Unit 1) is confined to streambeds feeding the reservoir system. Comparable riparian communities have been described from İğneada (Geranio robertiani-Carpinetum betuliKavgacı et al. 2011) and Bursa-Yeniköy (Apocyno veneti-Fraxinetum angustifoliae (Özen 2010) Kavgacı et al. 2016), however, both show low floristic similarity to the Belgrade Forest (8% and 18%, according to the Sørensen similarity index, respectively). Consequently, Unit 1 is described here as a new association, Equiseto telmateiae-Alnetum glutinosae ass. nov. In contrast, Unit 2 shows greater affinity to the association Trachystemono orientalis-Carpinetum betuli described from drier sites in the İğneada floodplain forest and is therefore assigned accordingly.
According to Willner et al. (2017), Fagus orientalis forests are mainly divided into thermo-basiphytic and meso-basiphytic beech forests in Anatolia. Çoban and Willner (2019) indicated that thermo-basiphytic beech forests lack distinct characteristics species as they share the diagnostic species with thermophilous oak and mixed broadleaved forests. For this reason, both thermophilous beech forests and mesic beech and fir-beech forests were classified under the alliance Fagion orientalis. Thermophilous beech forests were included under the suballiance Fagenion orientalisÇoban and Willner 2019.
Aydın et al. (2008) conducted a field excursion in northwestern of Türkiye including the Belgrade Forest and took three vegetation plots in order to determine forest communities. They named the Fagus orientalis-dominated forest (Unit 3) as Ilici colchicae-Fagetum orientalis Akman, Barbéro et Quézel 1978. This association occurs on a higher elevational belt (750–1250 m) in Kastamonu-Cide region with a different floristic composition such as the existence of Rhododendron ponticum and Rhododendron luteum which are absent in the Belgrade Forest. For this reason, Çoban and Willner (2019) accepted Cyclamini coi-Fagetum orientalisTzonev et al. 2006 for this unit which was described from a low elevational belt of Strandzha Mountains similar to Belgrade Forest. The type relevé of this unit has a high similarity (66%) with the Fagus orientalis-dominated forest of the study area.
In this study, the classification of Quercus petraea-dominated forests into a physiognomically coherent unit proved to be challenging. The tree-layer composition exhibited considerable heterogeneity, reflecting a range of stand structures. Such variation may be attributable to long-term degradation processes that have altered forest structure and tree-species diversity. Quercus petraea-dominated forests were classified as Salvio forskahlei-Quercetum cerris quercetosum frainetto Akman et al. ex Quezel et al. 1992 (Aydın et al. 2008) and Tanaceto cinerei-Quercetum ibericaeKavgacı et al. 2010 (Kavgacı et al. 2010). Kavgacı et al. (2010) suggested that Quercus ibericae-dominated forests in Northwestern Thrace show similar features to those from the Belgrade Forest and must be classified under Tanaceto cinerei-Quercetum petraea subsp. ibericaeKavgacı et al. 2010. However, the floristic similarity between this community and Quercus petraea subsp. ibericae dominated forest was found to be 46%. In addition, Tanacetum corymbosum subsp. cinereum is not found in the Belgrade Forest. For this reason, Salvio forskahlei-Quercetum cerris is more suitable for Quercus petraea-dominated forests (Unit 4–5). According to Akman (1995), Quercus frainetto-Quercus cerris forests occur on non-calcareous metamorphic bedrocks on a wide area in the southern parts of the Marmara Sea. These forests were defined as Salvio forskahlei-Quercetum cerridis which were subdivided into quercetosum frainetto and helleboretosum orientalis subassociations. The former subassociation occurs on relatively more humid part of the distribution area of the association (Quézel et al. 1992). However, Fagus orientalis co-dominated unit of this association was accepted as another subassociation, namely as Salvio forskahlei-Quercetum cerridis fagetosum orientalis, which has a transitional character between mesophilous beech forests (Unit 3) and thermophilous beech and oak forests (Unit 5).
Maquis vegetation was not adequately studied in western Euxine region. Therefore, it was not possible to assign pseudomaquis vegetation to a validly published and physiognomically similar association. Özyiğit et al. (2014) described two maquis associations in İstanbul Princes’Island as Arbuto unedonis-Phillyretum latifoliae Özyiğit et al. 2014 and Phillyreo latifoliae-Pinetum brutiae Schwarz 1936. However, both units differ from that of the study area in terms of dominant species and also similarity of species composition (5% similarity). For this reason, pseudomaquis vegetation which mostly occurs on degraded forest openings was defined as a new association. Çoban et al. (2019) also reported distribution of this association on powerline corridors which are periodically cleared for forest fire prevention around Belgrade Forest.
Based on the results of the numerical analyses, the following syntaxonomic scheme could be proposed:
Alno glutinosae-Populetea albae P. Fukarek et Fabijanić 1968
Alno-Fraxinetalia excelsioris Passarge 1968
Alno-Quercion roboris Horvat 1950
Equiseto telmateiae-Alnetum glutinosae ass. nov. (Unit 1)
Holotypus (Suppl. material 1, Relevé 5)
Location: Ayvad Bendi, relevé area (m2): 400; 41.2151015°N, 28.94198553°E, elevation: 96 m a.s.l., slope: 13°
This study reconciles vegetation surveys to provide a coherent and reproducible vegetation typology for the Belgrade Forest. The resulting syntaxonomic scheme, supported by diagnostic species and DCA gradients, clarifies the delimitation of mesophilous beech, meso-thermophilous mixed oak–beech–hornbeam and riparian forest communities, and distinguishes pseudomaquis under degraded conditions. These outcomes establish a standardized baseline for long-term monitoring and management, particularly for tracking secondary succession and prioritizing restoration.
Acknowledgements
We sincerely thank Milan Chytrý for his thorough review and valuable insights, which significantly strengthened the manuscript. We also thank the anonymous reviewers for their constructive comments, which helped improve the clarity and overall quality of the paper.
Additional information
Conflict of interest
The authors have declared that no competing interests exist.
Ethical statement
No ethical statement was reported.
Use of AI
No use of AI was reported.
Funding
The authors gratefully acknowledge the Italian Society of Vegetation Science (SISV) for granting a full publication fee waiver under the Young Researcher Support initiative.
Author contributions
Süleyman Çoban, Conceptualization, Data compiling, Methodology, Research design, Data analysis, Statistical interpretation, Drafting of the manuscript, Manuscript revision and response to reviewers. Selim Bayraktar, Conceptualization, Data compiling, Methodology, Research design, Drafting of the manuscript, Statistical interpretation, Review and editing, Manuscript revision and response to reviewers, Corresponding author. Nilüfer Şahin, Data providing, Contribution to manuscript writing and revisions.
Author ORCIDs
Süleyman Çoban https://orcid.org/0000-0003-1570-9795
Selim Bayraktar https://orcid.org/0000-0002-8023-4135
All of the data that support the findings of this study are available in the main text or Supplementary Information.
ReferencesAcatay A (1943) İstanbul çevresi ve bilhassa Belgrad Ormanı’ndaki zararlı orman böcekleri, mücadeleleri ve işletme üzerinde tesirleri [Harmful forest insects in and around Belgrade Forest, their control and effects on forest management]. Yüksek Ziraat Enstitüsü Çalışmaları 142: 1–88. Ziraat Vekâleti Yayınları, Ankara, Türkiye.Arslangündoğdu Z, Yılmaz E (2011) The effects of tree stand layers on resident bird species in Belgrade Forest, Istanbul, Turkey. Allgemeine Forst und Jagdzeitung 182: 25–29.Aydın C, Bergmeier E, Walentowski H (2008) Vegetation und Standort in der Waldlandschaft der Nordwest-Türkei. Önel Druck & Verlag, Köln.Akman Y (1995) Türkiye orman vejetasyonu [The forest vegetation of Turkey]. Ankara Üniversitesi Fen Fakültesi Yayınları, Ankara, Türkiye, 290 pp.Bayraktar S, Cegielska K, Sökmen ED, Noszczyk T, Yener ŞD, Kukulska-Kozieł A (2024) Directions of land degradation in the Greater Istanbul Metropolitan Area: A view from four decades. Land Degradation & Development 35(5): 1656–1672. https://doi.org/10.1002/ldr.5012Bayraktar S, Yener ŞD (2025) Floristic diversity and conservation in peri-urban landscapes: The case of Sarıyer, Istanbul. Artvin Çoruh Üniversitesi Orman Fakültesi Dergisi 26(1): 69–79. https://doi.org/10.17474/artvinofd.1613920Belon P (1553) Les observations de plusieurs singularitez et choses memorables trovvees en Grece, Asie, Iudée, Egypte, Arabie et autres pays estranges. Corrozet et Canellat, Paris. https://doi.org/10.5962/bhl.title.127374Bonari G, Knollová I, Vlčková P, Xystrakis F, Çoban S, … Chytrý M (2019) CircumMed Pine Forest Database: an electronic archive for Mediterranean and Submediterranean pine forest vegetation data. Phytocoenologia 49(3): 311–318. https://doi.org/10.1127/phyto/2019/0311Bonari G, Fernández‐González F, Çoban S, Monteiro‐Henriques T, Bergmeier E, … Chytrý M (2021) Classification of the Mediterranean lowland to submontane pine forest vegetation. Applied Vegetation Science 24(1): e12544. https://doi.org/10.1111/avsc.12544Bornmüller J (1897) Ein Maiausflug in dem “Wald von Belgrad” bei Constantinopel. Mitteilungen des Thüringischen Botanischen Vereins Neue Folge 15–16: 1–12.Chytrý M, Hennekens SM, Jiménez‐Alfaro B, Knollová I, Dengler J, … Yamalov S (2016) EuropeanVegetation Archive (EVA): an integrated database of European vegetation plots. Applied VegetationScience 19: 173–180. https://doi.org/10.1111/avsc.12191Çoban S (2017) Development of a national vegetation database for Turkey: a preliminary study. In: 37th Meeting of the European Association for the Study of Vegetation (EADSVE), Prizren (Kosovo), 13–16 July 2017.Çoban S, Willner W (2019) Numerical classification of the forest vegetation in the Western Euxine Region of Turkey. Phytocoenologia 49(1): 1–19. https://doi.org/10.1127/phyto/2018/0274Çoban S, Bayraktar S, Akgül M (2016) Forest vegetation maps and its development in Turkey: a case from Istanbul Belgrade forest. Forest Review 47(1): 7–16.Çolak AH (2013) Belgrad Ormanı’nın Ağaçları [Trees of Belgrade Forests]. In: Çolak AH (Ed.) Belgrad Ormanı – Bir Doğa ve Kültür Mirası. İstanbul: T.C. Orman ve Su İşleri Bakanlığı, 1. Bölge Müdürlüğü, Tor Ofset San. Ve Tic. Ltd. Şti., 154–186.Çolak AH, Kırca S (2013) Belgrad Ormanı flora araştırmaları tarihçesi [History of Floristic Research in Belgrade Forest]. In: Çolak AH (Ed.) Belgrad Ormanı – Bir Doğa ve Kültür Mirası. İstanbul: T.C. Orman ve Su İşleri Bakanlığı, 1. Bölge Müdürlüğü, Tor Ofset San. Ve Tic. Ltd. Şti., 25–45.Çolak AH, Çoban S, Özalp G (2013) Belgrad Ormanı’nın floristik analizi ve yetişme ortamı özellikleri [Floristic Analysis and Site Characteristics of the Belgrade Forest]. In: Çolak AH (Ed.) Belgrad Ormanı – Bir Doğa ve Kültür Mirası. İstanbul: T.C. Orman ve Su İşleri Bakanlığı, 1. Bölge Müdürlüğü, Tor Ofset San. Ve Tic. Ltd. Şti., 278–352.Dallaway J (1797) Constantinople Ancient and Modern: With Excursions to the Shores and Islands. London: T. Cadell Jun. And W. Davies.De Tchihatchef PA, De Verneuil PP (1853) Asie Mineure: description physique, statistique et archéologique de cette contrée. Gide et J. Baudry; L. Guérin, Paris.Forskål P (1775) Flora aegyptiaco-arabica: sive descriptiones plantarum, quas per Egyptum inferiorem et Arabiam felicem detexit, illustravit Petrus Forskål. Hauniae: Möller. https://doi.org/10.5962/bhl.title.6625Hennekens SM, Schaminée JHJ (2001) TURBOVEG, a comprehensive database management system for vegetation data. Journal of Vegetation Science 12: 589–591. https://doi.org/10.2307/3237010Hill MO (1979) TWINSPAN: a FORTRAN program for arranging multivariate data in an ordered two-way table by classification of the individuals and attributes. ISection of Ecology and Systematics, Cornell University, thaca, NY.Irmak A (1940) Belgrad Ormanı toprak münasebetleri [Soil relationships in the Belgrade Forest]. Yüksek Ziraat Enstitüsü Çalışmaları 70: 1–52.Kantarcı MD (1980) İstanbul ve civarının orman ekosistemlerinin yetişme ortamı özellikleri [Site characteristics of forest ecosystems in and around Istanbul]. İstanbul Üniversitesi Orman Fakültesi Dergisi A 30(1): 1–140.Kavgacı A, Čarni A, Tecimeni B, Özalp G (2010) Diversity and ecological differentiation of oak forests in NW Thrace (Turkey). Archives of Biological Sciences 62(3): 705–718. https://doi.org/10.2298/ABS1003705KKavgacı A, Čarni A, Tecimen HB, Özalp G (2011) Diversity of floodplain forests in the Igneada region (NW Thrace-Turkey). Hacquetia 10(1): 73–93. https://doi.org/10.2478/v10028-011-0003-9Kavgacı A, Arslan M, Bingöl Ü, Erdoğan N, Čarni A (2012) Classification and phytogeographical differentiation of oriental beech forests in Turkey and Bulgaria. Biologia 67(3): 461–473. https://doi.org/10.2478/s11756-012-0029-6Kavgacı A, Yalçın E, Korkmaz H (2016) Numerical classification and ordination of the floodplain forests in the Euxine region of Turkey. Turkish Journal of Botany 40(2): 164–175. https://doi.org/10.3906/bot-1504-31Kavgacı A, Karaköse M, Keleş ES, Balpınar N, Arslan M, … Čarni A (2023) Classification of forest and shrubland vegetation in central and eastern Euxine Turkey and SW Georgia. Applied Vegetation Science 26(4): e12753. https://doi.org/10.1111/avsc.12753Kırca S, Çolak AH (2013) Belgrad Ormanı tarihçesi. In: Çolak AH (Ed.) Belgrad Ormanı – Bir Doğa ve Kültür Mirası. İstanbul: T.C. Orman ve Su İşleri Bakanlığı, 1. Bölge Müdürlüğü, Tor Ofset San. Ve Tic. Ltd. Şti., 11–24.Mucina L, Bültmann H, Dierßen K, Theurillat JP, Raus T, … Tichý L (2016) Vegetation of Europe: hierarchical floristic classification system of vascular plant, bryophyte, lichen and algal communities. Applied Vegetation Science 19: 3–264. https://doi.org/10.1111/avsc.12257OGM (2012a) Bahçeköy Orman İşletme Müdürlüğü Bendler Orman İşletme Şefliği. Orman Amenajman Planı (2012–2021) [Forest Management Plan (2012–2021)]. Orman Genel Müdürlüğü (General Directorate of Forestry), İstanbul, Türkiye.OGM (2012b) Bahçeköy Orman İşletme Müdürlüğü Kurtkemeri Orman İşletme Şefliği. Orman Amenajman Planı (2012–2021) [Forest Management Plan (2012–2021)]. Orman Genel Müdürlüğü (General Directorate of Forestry), İstanbul, Türkiye.OGM (2012c) Bahçeköy Orman İşletme Müdürlüğü Atatürk Arboretumu Orman İşletme Şefliği. Orman Amenajman Planı (2012–2021) [Forest Management Plan (2012–2021)]. Orman Genel Müdürlüğü (General Directorate of Forestry), İstanbul, Türkiye.Özalp G (2013) Belgrad Ormanı’nın çalıları [Shrubs of Belgrade Forests]. In: Çolak AH (Ed.) Belgrad Ormanı – Bir Doğa ve Kültür Mirası. İstanbul: T.C. Orman ve Su İşleri Bakanlığı, 1. Bölge Müdürlüğü, Tor Ofset San. Ve Tic. Ltd. Şti., 187–216.Özen F (2010) Yeniköy (Bursa) higrofil, orman ve maki vejetasyonunun sinekolojik ve sintaksonomik analizi [Syntaxonomical and synecological analysis of hygrophile, forest and maquis vegetation of Yeniköy (Bursa)]. Ekoloji 19: 50–64. [in Turkish] https://doi.org/10.5053/ekoloji.2010.766Özhan S, Balcı AN, Özyuvacı N, Hızal A, Gökbulak F, Serengil Y (2005) Determination of cover and management factor of the Universal Soil-Loss Equation for forest ecosystems in Marmara Region, Turkey. Forest Ecology and Management 214: 118–123. https://doi.org/10.1016/j.foreco.2005.03.050Özhatay N, Yüzbaşıoğlu S (2013) Belgrad Ormanı’nın otsu bitkileri (Herbs of Belgrade Forests). In: Çolak AH (Ed.) Belgrad Ormanı – Bir Doğa ve Kültür Mirası. İstanbul: T.C. Orman ve Su İşleri Bakanlığı, 1. Bölge Müdürlüğü, Tor Ofset San. Ve Tic. Ltd. Şti., 217–277.Özyigit S, Altay V, Özyigit I, Yarci C (2014) Vegetation ecology of the Princes’ Islands, Istanbul Turkey. Journal of Environmental Biology 36(1): 113.Quézel P, Barbero M, Akman Y (1992) Typification de syntaxa décrits en région méditerranéenne orientale. Ecologia Mediterranea 18: 81–87. https://doi.org/10.3406/ecmed.1992.1708Regel C (1959) Vegetationszonen und Vegetationsstufen in der Türkei. Feddes Repertorium Beiheft 138: 230–282.Rodwell JS [Ed.] (1998) British plant communities: volume 1, woodlands and scrub (Vol. 1). Cambridge University Press.Şahin G (2022) Nemli biyotop bitki toplumları ve haritalaması; Belgrad Ormanı örneği [The plant communities of humid biotopes and their mapping: example of Belgrade Forest]. PhD Thesis, Istanbul University-Cerrahpaşa, Institute of Graduate Studies, Istanbul, Türkiye.ter Braak CJF, Šmilauer P (2002) CANOCO for Windows Version 4.5. Wageningen: Biometris, The Netherlands.Theurillat JP, Willner W, Fernández-González F, Bültmann H, Čarni A, … Weber H (2021) International code of phytosociological nomenclature. Applied Vegetation Science 24: e12491. https://doi.org/10.1111/avsc.12491Tichý L (2002) JUICE, software for vegetation classification. Journal of Vegetation Science 13: 451–453. https://doi.org/10.1111/j.1654-1103.2002.tb02069.xTichý L, Axmanová I, Dengler J, Guarino R, Jansen F, … Schaminée JHJ (2023) Ellenberg-type indicator values for European vascular plant species. Journal of Vegetation Science 34: e13168. https://doi.org/10.1111/jvs.13168Tzonev R, Dimitrov M, Chytry M, Roussakova V, Dimova D, … Ganeva A (2006) Beech forest communities in Bulgaria. Phytocoenologia 36(2): 247–279. https://doi.org/10.1127/0340-269x/2006/0036-0247Uğurlu E, Roleček J, Bergmeier E (2012) Oak woodland vegetation of Turkey – a first overview based on multivariate statistics. Applied Vegetation Science 15(4): 590–608. https://doi.org/10.1111/j.1654-109x.2012.01192.xWheler G (1862) A Journey into Greece. Cademan, Kettlewel, Awnsham Churchill, London.Willner W, Jiménez‐Alfaro B, Agrillo E, Biurrun I, Campos JA, … Chytrý M (2017) Classification of European beech forests: a Gordian Knot?. Applied Vegetation Science 20(3): 494–512. https://doi.org/10.1111/avsc.12299Yaltırık F (1963) Belgrad Ormanı vejetasyonunun floristik analizi ve ana meşcere tiplerinin kompozisyonu üzerine araştırmalar. İstanbul Üniversitesi Orman Fakültesi Dergisi A 13: 1–75.Yaltırık F (1968) Belgrad Ormanı ve civarının orman kuruluşu üzerine araştırmalar. İstanbul Üniversitesi Orman Fakültesi Dergisi A 18: 45–67.Yaltırık F, Efe A (1996) Otsu bitkiler sistematiği. İstanbul Üniversitesi Orman Fakültesi Yayınları, İstanbul.Yaltırık F, Akman Y, Ketenoğlu O (1983) A phytosociological research in the Belgrad Forest. Communications Faculty of Sciences University of Ankara Series C 27: 55–72. https://doi.org/10.1501/commuc_0000000041Yönelli V (1986) Belgrad Ormanındaki orman toplumlarının yapısı ve silvikültürel değerlendirilmesi [Structure and silvicultural evaluation of forest communities in Belgrade Forest]. PhD Thesis, Istanbul University, Istanbul, Türkiye.
Topical Collection: “Vegetation classification: from classic to numeric approaches”.
https://binary.pensoft.net/file/1579126This dataset is made available under the Open Database License (http://opendatacommons.org/licenses/odbl/1.0). The Open Database License (ODbL) is a license agreement intended to allow users to freely share, modify, and use this Dataset while maintaining this same freedom for others, provided that the original source and author(s) are credited. Süleyman Çoban, Selim Bayraktar, Nilüfer Şahin10.3897/ved.177153.suppl224C67B81-4AC6-531A-8F30-063C4A839FC1
Full synoptic table
https://doi.org/10.3897/ved.177153.suppl2
xlsx
https://binary.pensoft.net/file/1579127This dataset is made available under the Open Database License (http://opendatacommons.org/licenses/odbl/1.0). The Open Database License (ODbL) is a license agreement intended to allow users to freely share, modify, and use this Dataset while maintaining this same freedom for others, provided that the original source and author(s) are credited. Süleyman Çoban, Selim Bayraktar, Nilüfer Şahin