Phytosociological Analysis of Natural and Artiﬁcial Pine Forests of the Class Vaccinio-Piceetea Br.-Bl. in Br.-Bl. et al. 1939 in the Sudetes and Their Foreland (Bohemian Massif, Central Europe)

: Research Highlights: Differentiation of Scots pine forests of the class Vaccinio-Piceetea in Poland has been the subject of numerous studies, including revisions. Despite that, the area of southwestern Poland was hitherto practically unexplored in this respect. Background and Objectives: The aim of this work was therefore (i) to present the diversity of the pine forests in the Sudetes and their foreland; (ii) to compare the ecology of studied communities. Materials and Methods: We analyzed 175 phytosociological relev é s collected between 1991 and 2020 in natural and anthropogenic pine stands. To identify vegetation types, we used the modiﬁed TWINSPAN algorithm; principal coordinate analysis, distance-based redundancy analysis and permutational tests were applied to identify the variation explained and the main environmental gradients shaping the studied plant communities. Results: Five associations were distinguished: thermophilous Asplenio cuneifolii-Pinetum sylvestris Pišta ex Husov á in Husov á et al. 2002, which develops on shallow soils over ultrabasic substrates, Hieracio pallidi-Pinetum sylvestris Stöcker 1965, which prefers outcrops of acidic rocks; Betulo carpaticae-Pinetum sylvestris Mikyška 1970, which is relict in origin and occurs on the upper Cretaceous sandstones, the peatland pine–birch forests of the Vaccinio uliginosi-Betuletum pubescentis Libbert 1933 and the Vaccinio myrtilli-Pinetum sylvestris Juraszek 1928. Moreover, community Brachypodium sylvaticum-Pinus sylvestris with the occurrence of many thermophilous and basiphilous species was also found on limestone substratum. The analysis of the species composition of pine plantations established on deciduous and mixed forests habitats revealed that these anthropogenic communities were marked by a random combination of species in which a certain group of common forest generalists participated. The distinguished communities differed clearly among each other also in habitat characteristics. Particularly important for their differentiation were soil reaction and nutrients, supported by differences in moisture, temperature and light availability. Apart from the edaphic factors, altitude and the bedrock type proved to be equally important. Conclusions: Our study provides new remarks to the typology and synecology of pine forest communities in SW Poland.


Introduction
The history of research on pine forests in Poland dates back to the 1930s of the previous century [1]. However, more intensive investigations began at the turn of the 1950s and 1960s [2][3][4], and they continue today. Over the last seventy years, there have been several revisions on the national scale relating both to the phytosociological diversity of pine forests [5][6][7][8] and the main changes they undergo [9]. However, it should be emphasized

Field Sampling
Vegetation composition was sampled by using phytosociological relevés according to the Braun-Blanquet approach [25]. Field sampling was conducted from 1991 to 2020. We focused on not only potentially natural pine forests (rock outcrops, peat bogs), but also anthropogenic stands. Collected data originated from the stands where Scots pine's cover reached more than 25% in the tree layer. The relevés area ranged from 25 square m up to 400 square m (on flat stands and within floristically homogenous plots). The inclusion of relevés of less than 100 square m to document the full variability of Pinus sylvestris L. communities was necessary because most of the sites within hard-to-reach places on rock ledges and pinnacle rocks have an area of 25-50 square meters. Excluding them from the analysis would mean omitting a key part of ecological and phytosociological diversity of the studied communities.

Environmental Variables
In order to identify ecological conditions of the pine forests within the study area, different environmental variables were analyzed. Altitude (measured in m a.s.l. and divided by 1000 to standardize value), heat load [26] and bedrock type were used as explanatory variables. The bedrock type at each site was obtained from a Detailed Geological Map of the Sudetes mountains. (Polish Geological Institute, National Research Institute, http://www.pgi.gov.pl/pl/oddzial-dolnoslaski/separator1/geologiadolnego-slaska/mapageologiczna-sudety.html). Based on the criterion of mineral composition and major geological processes [27], nine categories of rocks were proposed as explanatory variables: Quaternary deposits (sands, clays and gravels), Quaternary peat bogs, serpentines, limestones, metamorphic rocks (gneisses, schists, mud-and claystones, greywackes, Paleozoic

Field Sampling
Vegetation composition was sampled by using phytosociological relevés according to the Braun-Blanquet approach [25]. Field sampling was conducted from 1991 to 2020. We focused on not only potentially natural pine forests (rock outcrops, peat bogs), but also anthropogenic stands. Collected data originated from the stands where Scots pine's cover reached more than 25% in the tree layer. The relevés area ranged from 25 square m up to 400 square m (on flat stands and within floristically homogenous plots). The inclusion of relevés of less than 100 square m to document the full variability of Pinus sylvestris L. communities was necessary because most of the sites within hard-to-reach places on rock ledges and pinnacle rocks have an area of 25-50 square meters. Excluding them from the analysis would mean omitting a key part of ecological and phytosociological diversity of the studied communities.

Environmental Variables
In order to identify ecological conditions of the pine forests within the study area, different environmental variables were analyzed. Altitude (measured in m a.s.l. and divided by 1000 to standardize value), heat load [26] and bedrock type were used as explanatory variables. The bedrock type at each site was obtained from a Detailed Geological Map of the Sudetes mountains. (Polish Geological Institute, National Research Institute, http://sudety.pgi.gov.pl/). Based on the criterion of mineral composition and major geological processes [27], nine categories of rocks were proposed as explanatory variables: Quaternary deposits (sands, clays and gravels), Quaternary peat bogs, serpentines, limestones, metamorphic rocks (gneisses, schists, mud-and claystones, greywackes, Paleozoic conglomerates), granitoids, trachytes and two kinds of sandstones-very different in terms of their properties. The first one labelled as Upper Jointed Sandstone of Late Turonian/Coniacian age [28] is subjected to very slow weathering processes; therefore, only initial podzols derive from it. The second one labelled as Lower Jointed Sandstone (Coenomanian/early Turonian) has got structures that are more prone to weathering; thus, deeper podzols and even cambisols derive from it. Heat load index (HL), is a direct measure of incident radiation calculated from slope inclination, aspect and latitude [26]. As there were no direct measurements of light and soil condition Ellenberg indictor values (EIVs) [29], corrected by datasets of Berg et al. [30] with reference to values of continentality were used. EIVs weighted by percentage species' cover were calculated for each relevé using the JUICE software [31].

Phytosociological Analysis
Occurrences of the same woody species in different vertical layers were merged using the procedure implemented in JUICE-under the assumption that the overlap of layers is random [31,32]. Plant nomenclature follows Euro+Med PlantBase [33] for vascular plants and Ochyra et al. [34] and Fałtynowicz [35] for bryophytes and lichens, respectively. The nomenclature of phytosociological alliances and classes is in accordance with Mucina et al. [23].
The vegetation types were identified using a modified TWINSPAN algorithm [36] with Total Inertia measure of heterogeneity using JUICE software [37]. The obtained number of clusters coincided with the results of crispness analysis [38], suggesting the division of relevés into nine groups.
Diagnostic species were determined using the Φ coefficient as a measure of fidelity for clusters of equalized size [39][40][41]. Species with Φ ≥ 20 (0.10 × 100), constancy ≥ 20%, constancy ratio [42] higher than 1.5 and significant concentration in a particular cluster, tested by the Fisher's exact test (p < 0.05), were considered to be diagnostic. A species was considered diagnostic for more than one cluster with the Φ > 20 (0.2 × 100) in at least two clusters, regardless of the constancy ratio. Species with constancy ratio < 1.5 and Φ > 20 in only one cluster were not considered diagnostic. Constant species were defined as species with frequency of at least 60% in a cluster. Distribution maps of the recognized clusters were prepared using DMAP software [43].

Ecological Analysis
Principal coordinate analysis (PCoA) [44] was performed both to explore differentiation of recognized clusters and check the percentage of variation explained. Matrix of distance (175*175) was calculated using Sorensen distance and square-root species cover transformation. Distribution of the sample groups on PCoA diagram was visualized and interpreted.
To identify the statistical significance of correlations (using Spearman's coefficient) between the PCoA sample scores obtained from CANOCO and mean randomized EIVs for relevés a modified permutation test with 499 unrestricted permutations was conducted. The test was performed with MoPeT_v1.2.r script [45] in R software [46]. Permutational analysis of variance (one-way ANOVA on the mean randomized EIVs) and modified permutation test (with 499 unrestricted permutations) were also calculated using MoPeT_v1.2.r [45], to determine which EIVs differentiate the selected communities. Using permutation ANOVA is an alternative to other tests under non-normal conditions, because it does not operate under the assumption of normality and uses actual scores [47].
Distance-based redundancy analysis (db-RDA) embedded in CANOCO 5.0 [44], with Sorensen distance and square-root species transformation, was implemented to check the main ecological drivers affecting the diversity of distinct groups, and variation explained [48]. A standard Monte Carlo permutation test with 499 unrestricted permutations under the full model was conducted to identify the significance of the simple term and conditional effects of environmental variables (such as altitude, heat load and bedrock type) on the species composition of the analyzed samples [44]. The conditional effect expresses the variation explained by a single explanatory variable, whereas the others are used as covariables. The simple effect expresses the variation explained by the single explanatory variable without covariables.

Results
In the analyzed material we distinguished nine groups of Scots pine forests (both natural and semi-natural, and planted, Table 1, Figures 2 and 3), clearly different from each  other in terms of the species composition and main ecological indicators (Tables 2 and 3, Figures 4 and 5).

Natural Or Semi-Natural Communities
Among the analyzed material, six groups (clusters) of the relevés represent natural or semi-natural phytocoenoses, possible for phytosociological identification, even if some of their patches are anthropogenic in character. The identified phytocoenoses can be arranged according to the following syntaxonomic classification.   These are mostly open-canopy, managed and even-aged forests of Pinus sylvestris on shallow soils derived from serpentine or peridotite bedrock. However, there are some sites known with spontaneous Scots pine regeneration. On such localities the Scots pine trees are dwarf and looks malnourished, in contrast to young trees with straight trunks on intensively managed stands. Rich in species herb layer (mean 32 species per relevè) contains a differentiated composition of grasses, dwarf shrubs and thermophilic forbs, with many locally rare or endangered species. Community stands out from the other of Scots pine forests by the frequent occurrence of Silene vulgaris (Moench) Garcke, Galium verum L., Asplenium cuneifolium Viv. and Potentilla alba L., which locally prefers serpentine soils. This is rare community, known from four isolated localities in the Grochowa Massif (Pogórze Paczkowskie Plateau), Kiełczyńskie and Oleszeńskie Hills (Ślęża ophiolite massif) and near Janowice (Rudawy Janowickie mountains). This community is characterized by high EIVs for nutrients and temperature, and low EIV for moisture ( Figure 4). The community develops at altitude 320-460 m a.s.l.
On deeper soils derived from serpentine bedrock Pinus-dominated communities are developed usually as artificial forest, included in cluster 9, rarely 7 (see Figure 5), with dominance of mesophytic forest generalists and the absence of thermophilic species typical of open canopy forests.    14.9 This is the most common association in the study area. It occurs on rocky outcrops on steep slopes in the areas of acidic bedrocks (granites, gneisses); on deep soils resulting from weathering serpentinites; on other neutral rocks; and on weathered and blown sands and the podzolic soils derived from them. In most cases, its phytocoenoses are of anthropogenic origin. However, in some localities in the Sudetes they may be close to natural in character, as evidenced by the natural habit of trees, which is not the result of forest management pressure. Herb layer is extremely poor (13 species per relevé) and consists of the common species of acidophilic grasses and ericoid dwarf shrubs; acidophilic bryophytes often dominate or co-dominate here. The community is marked by low EIVs for moisture, soil reaction and nutrients (Figure 4), and develops at altitudes from 220 m to 735 m a.s.l.      This very rare community has been known so far from only three localities-one in the Trzcińskie Mokradła peat-bog (Kotlina Jeleniogórska) and two in the Stołowe mountains, where it develops on drained moors on sandstone plateau. The phytocoenoses are loose; the cover of the tree layer reaches 40%-50%; however, in contrast to the remaining natural or semi-natural forests, the shrub layer is well developed, and the undergrowth is abundant, with coverage close to 90%-100%. Depending on their location, apart from Scots pine, they are co-dominated by Betula pubescens var. pubescens Ehrh. or Picea abies. In the shrub layer, the presence of Salix aurita L, S. cinera L. and Frangula alnus is significant. The understory may be dominated by graminoids (such as Molinia caerulea (L.) Moench, Eriophorum sp., Carex sp., Juncus sp.), dwarf shrubs (mainly Vaccinium myrtillus L.) and bryophytes (numerous species of the genus Sphagnum sp. and Polytrichum sp.). In these phytocoenoses on average occur 17 plant species in the relevé. The community is marked by the highest EIV for moisture from all the recorded syntaxa ( Figure 4). The community develops at altitudes from 400 m to 720 m a.s.l.

Artificial Scots Pine Forests
Among the analyzed material, we also distinguished three groups of fully artificial forests dominated by Pinus sylvestris, differing in floristic composition, fertility of the habitat and soil reaction. All these communities, due to domination of Scots pine in the tree layer may be conditionally included in the Pinetalia sylvestris order as non-hierarchical phytocoenons. A characteristic feature that distinguishes them from natural or seminatural communities (except for Vaccinio myrtilli-Pinetum, whose patches are also partially anthropogenic) is low light availability (Figure 4).
The most common recorded community with the dominance of Pinus sylvestris, but with a constant and high proportion of nemoral species. The undergrowth includes common forest species of the Carpino-Fagetea sylvaticae Jakucs ex Passarge 1968 class (e.g., Acer platanoides L., Corylus avellana L., Brachypodium sylvaticum (Huds.) P. Beauv., Dryopteris filix-mas (L.) Schott, Drymochloa sylvatica (Pollich) Holub), along with nitrophilous (Urtica dioica L., Rubus fruticosus agg., R. idaeus L.) and alien species (Impatiens parviflora L.) is particularly common in this cluster). Among plantation forests, this community has the highest EIVs for soil reaction and nutrients, which indicates that pine was planted here on fertile habitats of deciduous forests of the Carpino-Fagetea sylvaticae class, most often in sub-mountainous locations (Carpinion betuli Issler 1931 alliance), less often in nutrient-rich beech forest habitats of the Fagetalia sylvaticae Pawł. in Pawł., Sokoł. et Wall. 1928. However, the identified combination of species indicates important transformation of the ecosystem, in which native forest species play a week role in relation to non-forest species. The mean number of species per relevé is quite high (23.6).    Submountain community (as evidenced by the constant share of Luzula luzuloides (Lam.) Dandy & Wilmott or Fagus sylvatica), similarly to the previous one in terms of high share of mesophilous general forest species but with higher requirements as to the soil moisture (Athyrium filix-femina (L.) Roth, Maianthemum bifolium (L.) F. W. Schmidt, Luzula pilosa (L.) Willd., Molinia caerulea). The share of alien species is much lower here than in the two previous groups (only Quercus rubra seems to be common). Among plantation forests, this community has the lowest EIVs for soil reaction and nutrients, but a constant share of forest species (e.g., Corylus avellana, Acer pseudoplatanus L.) suggests that pine was planted here on mesotrophic deciduous and mixed forests' habitats (submountain forms of the Tilio-Carpinetum and beech forests of the Luzulo-Fagion sylvaticae Lohmeyer et Tüxen in Tüxen 1954 alliance). Among artificial forest, this community retained its character most closely to the natural. The mean number of species per relevé is 21.4.

Ecological Differentiation
PCoA ordination diagram ( Figure 3) shows a species compositional pattern within all distinguished associations and communities. The PCoA results derived from CANOCO revealed that the first and the second PCoA axes explained 13.64% and 5.92% of compositional variability of studied communities, respectively. The first PCoA axis was significantly correlated with the EIVs for soil reaction and nutrients (both p < 0.01) and for temperature (p < 0.05). The second PCoA axis was significantly correlated (p < 0.01) with the EIV for light ( Table 2). The obtain results suggest that the studied communities are arranged along the first axis from those developing on nutrient-rich substrates with higher pH and favorable thermal conditions towards poor, acidic and oligothermic sites. The second axis determines a gradient along which studied forests are distributed from shaded ones to those with loose structure and higher light availability.
ANOVA of the six EIVs for the nine relevé groups suggested that most of the analyzed factors (except for continentality index) played a significant role in shaping the diversity of the studied vegetation types (Figure 4), as reflected by p < 0.01 for soil reaction, moisture and nutrients and p < 0.05 for light and temperature.
The db-RDA revealed that the explanatory variables used in the analysis accounted for 27.26% (adjusted explained variation was 22.76%) of variation in species composition. Figure 5 presents the db-RDA diagram of studied samples and main environmental gradients derived from: altitude, heat load index and the main types of bedrock. However, their contribution to the explained variability varies depending on whether we consider simple term effects or conditional effects (Table 3). The most important variable (both in simple and conditional effects) is altitude, but the effect of some bedrock types (especially two different kinds of sandstones, postglacial formations, limestones and serpentinites) is also significant. The db-RDA revealed that the explanatory variables used in the analysis accounted for 27.26% (adjusted explained variation was 22.76%) of variation in species composition. Figure 5 presents the db-RDA diagram of studied samples and main environmental gradients derived from: altitude, heat load index and the main types of bedrock. However, their contribution to the explained variability varies depending on whether we consider simple term effects or conditional effects (Table 3).   The most important variable (both in simple and conditional effects) is altitude, but the effect of some bedrock types (especially two different kinds of sandstones, postglacial formations, limestones and serpentinites) is also significant.

Discussion
The present study suggests that pine forests of the Sudetes mountains and their foreland are more differentiated than previously reported in phytosociological literature. Instead of three associations of pine forests known so far, five (and one community) were

Discussion
The present study suggests that pine forests of the Sudetes mountains and their foreland are more differentiated than previously reported in phytosociological literature. Instead of three associations of pine forests known so far, five (and one community) were distinguished, and their distinctiveness was confirmed both by the differences in species combinations and ecological conditions under which they develop.
Cluster 1 embraces phytocoenoses of pine forest communities developing on limestone substrate and classified as the community Brachypodium pinnatum-Pinus sylvestris, resembling more the Festuco-Pinion sylvestris alliance phytocoenoses than typical Dicrano-Pinion forest. These basiphilous, species-rich pine forests have been found so far neither in the Sudetes, nor in Poland at all (except for Erico-Pinetea Horvat 1959 relic Carpathian forest). They are reported in the north-western part of the Czech Republic [49] and Slovakia [50]. Communities described from Slovakia as Brachypodio pinnati-Pinus sylvestris Michalko 1980 are very different from the Sudeten phytocoenoses, in terms of the occurrence of Carpathian and sub-Mediterranean species, such as Calamagrostis varia (Schrad.) Host, Pulsatilla helleri subsp. slavica (G. Reuss) Zāmelis, Tanacetum corymbosum subsp. subcorymbosum (Schur) Pawł., Aster amellus L., Cotoneaster melanocarpus (Bunge) Loudon, Aegonychon purpurocaeruleum (L.) Holub and Euphorbia epithymoides L. The species composition of our community is almost identical with Czech basiphilous Scots pine forests [49], which are, however, probably erroneously identified with thermophilic, but acidophilous Festuco ovinae-Pinetum Kobendza 1930 association, described from rather acidic dunes in Central Poland [8]. The other type of subcontinental pine forests with the occurrence of thermophilous species (ass. Peucedano-Pinetum W.Mat. (1962) 1973 subass. pulsatilletosum and ass. Serratulo-Pinetum J.Mat 1981) is recorded from the central and north-western part of Poland and belongs to the Dicrano-Pinion alliance. These forests prefer poor in nutrients podzolic soils derived from Quaternary postglacial formations [7]. According to the Central European classification, the alliance Festuco-Pinion sylvestris is included in the separate class Pyrolo-Pinetea sylvestris Korneck 1974 embracing Euro-Siberian (sub)continental psammophilous (sub)thermophilous steppic pine forests [23]. Therefore, for the time being, we kept the classification of comm. Brachypodium pinnatum-Pinus sylvestris within the Dicrano-Pinion alliance. Certainly, this community requires further research based on wider phytosociological material from southern Poland.
Cluster 2 includes xerothermic pine forests developing on shallow soils formed on ultra-metamorphic bedrock, included in the Czech Republic in the Asplenio cuneifolii-Pinetum sylvestris Pišta ex Husová in Husová et al. 2002 [51]. The differences between phytocoenoses described from the Czech Republic and Poland are insignificant-in the Czech Republic, Erica carnea L., which is not present in Poland, is one of the diagnostic species, and Larix decidua Mill., which is a symptom of phytocoenoses deformations due to forest management, was also more frequent in the Czech stands. Other species, such as Asplenium cuneifolium, Silene vulgaris and Festuca ovina L., are diagnostic for the association in both countries. However, analogous serpentine communities in Austria (Festuco eggleri-Pinetum Eggler 1954 corr Wallnofer 1993) already differ significantly in terms of floristic composition [52,53]. The Asplenio cuneifolii-Pinetum sylvestris association has not been recorded in Poland so far.
Cluster 3 includes rocky, thermophilic pine forests of the Hieracio pallidi-Pinetum sylvestris Stöcker 1965 association, found so far on three sites on the peaks of granite hills in the Kotlina Jeleniogórska and its surroundings (Chojnik Mt., Witosza Mt. and Krzyżna Góra Mt.). The phytocoenoses known from Chojnik were originally described by Matuszkiewicz A. and W. [12,13] as Leucobryo-Pinetum, then byŚwierkosz [54] as Betulo carpaticae-Pinetum; the other sites have no published phytosociological documentation. Of the diagnostic species [55], the most important ones are bryophytes (Polytrichum piliferum Hedw., Ceratodon purpureus (Hedw.) Brid), although one of the plots on Chojnik Mt. also recorded another important species for the diagnosis of the association, Hieracium schmidtii Tausch [54] (Table 1, erroneously as "Hieracium glaucinum"). The phytocoenoses of rocky pine forests in the Sudetes are severely impoverished, probably due to their presence on the northern border of the range and a small number of typical species' sites. They are also exposed to the invasion of alien species, especially Pseudotsuga menziesii (Mirb.) Franco. These communities require further research, especially on the associated lichen flora and the origin of the pine trees. This association has not been recorded in Poland so far, but is found in the close vicinity of its borders in the Czech Republic [55] and Germany [56].
Cluster 4 embraces a typical forms of mesic pine forests which are of both anthropogenic and probably natural origin, belonging to the association Vaccinio myrtylli-Pinetum with the typical species composition of vascular plants and bryophytes [7,11]. The part of the phytocoenoses reported in the past from the Karkonosze mountains [12,13], theŚlęża Massif and Oleszeńskie Hills [14] and the Grochowa Massif [15] under the name Leucobryo-Pinetum (earlier synonym) in fact represent Vaccinio myrtylli-Pinetum. The association is frequent throughout Central Europe [7,11,50,52,53,57].
The relict, rocky pine forests on the sandstones of lower mountain zone (cluster 5), occurring on both sides of the Polish-Czech border. They were first described as Betulo carpaticae-Pinetum by Mikyška [10] and then reported by Passarge [58] from the Zittau mountains (Zittauer Gebirge) on the Czech-German border. This is how they were treated in some synthetic studies from the Bohemian Massif [59], and even suggested that they form an endemic association of this structural unit [60]. This concept has not yet been adopted and the name is now treated as a synonym for the Vaccinio myrtilli-Pinetum [11], while the association's distinctiveness is supported by many specific floral and ecological characteristics. Pine trees occurring here are a postglacial relicts as shown by Bobowicz [61] and Krzakowa, Lisowska [62] and differ significantly from the lowland populations in Poland. The closest to them, in biochemical term, populations of pine are found in Finland [62]. The species composition of phytocoenoses from SW Poland includes both Betula pubescens specimens with distinct features of subalpine B. pubescens var. glabrata [63,64] and hybrids with Betula pubescens var. pubescens [65]. Another species recorded in these phytocoenoses is subalpine Pinus mugo Turra and its hybrids with P. sylvestris-P. x rhaetica (=P. uliginosa Neumann) [66,67]. The herb layer is composed of rare in the study area species of boreocontinental range type (Empetrum nigrum L., Vaccinium uliginosum). The similarities of these phytocoenoses to subarctic ones are so clear that Kącki et al. [68] proposed their inclusion in the Betuletum pubescentis var. glabratae Lohmeyer and Bohn 1962 from the class Betuletea pendulo-pubescentis Julve 2016 (now as Vaccinio myrtilli-Betuletalia pubescentis Mucina et Willner ined.) [23], embracing European boreo-subarctic and orotemperate birch woods and krummholz on nutrient-poor podzolic soils. Regardless of the final solution adopted at a higher syntaxonomic level, it seems that this community should be given the status of an independent unit in the rank of an association.
Equally intriguing as the previous one is cluster 6, which includes bog pine-birch forests with the participation of Betula pubescenes var. pubescens (communities with Pinus sylvestris belonging to the class Oxycoco-Sphagnetea Br.-Bl. et Tüxen ex Westhoff et al. 1946 were not the subject of present study). The species composition of phytocoenoses included here clearly indicates their affiliation with the Vaccinio uliginosi-Betuletum pubescentis Libbert 1933 (e.g., through the presence of Salix aurita and some bog species of Cyperaceae and the lack of Rhododendron tomentosum Harmaja); in the Polish phytosociological literature this name refers to communities in the north-western part of Poland occurring in the sub-Atlantic climate range [7,8]. It is not known where this limitation comes from, since the association is also recorded in the southern part of Germany [57]; in the Czech Republic [69], Austria [52] [38].
In present study we also distinguished three different communities of anthropogenic pine forests established by planting Pinus sylvestris in deciduous and mixed forest habitats. Most of these patches represent anthropogenic ecosystems with random combinations of species, in which a group of common forest generalists also participates. Only the communities on nutrient-poor substrates (comm. Pinus sylvestris-Molinia caerulea) refer to the floristic composition of the native communities within the Luzulo-Fagion sylvaticae alliance.
A distinct feature of these anthropogenic and artificial forest ecosystems is the presence of non-forest species which find here optimal conditions and benefit from changes caused by forest management and other anthropogenic processes of global character-eutrophication [71,72], invasions of geographically alien species [73,74] and thermophilization-all connected with climate change and enhanced by direct human impact on forests [75,76]. These processes favor the entry of species which have not been present in native forest ecosystems so far, and their impact is particularly noticeable in the group of secondary (artificial) communities. The same refers to neophyte species whose presence is particularly pronounced in the discussed communities of anthropogenic character (Table 4).
Among the neophytes, the most common are Impatiens parviflora, Quercus rubra and Robinia pseudoacacia L. for which artificial pine forests are favorable habitats, especially when compared with pine communities of natural character. Among the latter ones, neophytes were encountered only occasionally (clusters 3 and 4) or were not recorded at all (clusters 5 and 6). This suggests that anthropogenically altered forests, heavily overexposed and with a disturbed undergrowth structure, may be a kind of gateway through which alien species enter the local pool of forest communities. Such a mechanism is known and well documented from both forests [77][78][79] and other types of ecosystems worldwide [80]. Research conducted in both tropical [81] and temperate forests [82] indicates that tree felling creating large gaps, combined with wide roads for transporting timber, enhances the penetration of alien species into native communities. The share of neophytes in the undergrowth increases as the crown density decreases [82]. Although many attempts of invasion fail and alien species disappear as the forest regenerates and light availability declines [81], some of them become invasive. They start to modify the structure of ecosystems significantly, e.g., by limiting the growth of seedlings [83], directly competing with native species and changing the physical, chemical and biotic properties of environment [83,84]. Among the species considered highly invasive in Europe [83], and reported also during present study, at least five (Impatiens parviflora, Solidago gigantea, Quercus rubra, Robinia pseudoacacia and Prunus serotina) fully meet the definition of "transformers"-species which change the structures of ecosystems [85]. Their high share in forest plantations creates a serious risk of transition to native forest communities in the closest vicinity, as previously reported from many forests of tropical [81] and subtropical zones [84].
A similar relationship was observed for nitrophilous species (EIV ≤ 7)-Rubus fruticosus agg., Sambucus nigra L., S. racemosa L., Senecio ovatus (G. Gaertn. & al.) Hoppe, Stellaria media (L.) Cirillo, Galium aparine L., Urtica dioica L., Geum urbanum L., Elytrigia repens (L.) Nevski, Alliaria petiolata (M. Bieb.) Cavara & Grande, Geranium robertianum L., Aegopodium podagraria L., Chaerophylum temulum L., Chelidonium majus L. and Ranunculus repens L. were found only in secondary pine forests (comm. Impatiens parviflora-Pinus sylvestris, comm. Prunus serotina-Pinus sylvestris, comm. Molinia caerulea-Pinus sylvestris and ass. Asplenio cuneifolii-Pinus sylvestris). This may indicate the role that secondary forests play in the general decline of native forest biodiversity in Europe [72,86]. It should be emphasized that the majority of the analyzed patches are located within larger forest complexes, and the Polish rules of forest management do not require additional fertilization of plantations established in such areas. Therefore, the presence of alien and nitrophilous species can be associated with the direct impact of the management itself, i.e., the harvesting of wood and then the renewal of the area by natural or artificial pine planting.

Conclusions
In this study we distinguished five associations: thermophilous Asplenio cuneifolii-Pinetum sylvestris, which develops on shallow soils over ultrabasic substrates, Hieracio pallidi-Pinetum sylvestris, which prefers outcrops of acidic rocks; Betulo carpaticae-Pinetum sylvestris, which is relict in origin and occurs on the upper Cretaceous sandstones; the peatland pine-birch forests of the Vaccinio uliginosi-Betuletum pubescentis and the Vaccinio myrtilli-Pinetum sylvestris previously described as the Leucobryo-Pinetum. Moreover, community Brachypodium sylvaticum-Pinus sylvestris with the occurrence of many thermophilous and basiphilous species was also found on limestone substratum. Three of the abovementioned syntaxonomical units are new for Poland. The obtained results also indicate that the anthropogenic pine plantations established in deciduous and mixed forests habitats are marked by a random combination of species in which a certain group of common forest generalists participated. The distinguished syntaxonomical units differ clearly among each other in habitat characteristics as well. Particularly important for their differentiation are edaphic conditions reflected by soil reaction and nutrients. Studied communities show their distinctiveness also in terms of moisture, temperature, light availability, altitude and bedrock type.
Aliens, including invasive species and nitrophilous, non-forest species were recorded within secondary forest communities (comm. Impatiens parviflora-Pinus sylvestris, comm. Prunus serotina-Pinus sylvestris, comm. Molinia caerulea-Pinus sylvestris, ass. Asplenio cuneifolii-Pinus sylvestris) almost exclusively. This may indicate the general role of such secondary forests in the decreasing of native forest biodiversity not only on a local, but also on a European scale.
Author Contributions: Investigation, K.R., P.P. and K.Ś.; conceptualization, K.R. and K.Ś.; writingoriginal draft, K.R. and K.Ś.; writing-review and editing, K.R., P.P. and K.Ś. All authors have read and agreed to the published version of the manuscript.