Aims: Inconsistent treatment of the vegetation layers is one of the main problems in the floristic classification of forests. In this study I investigate whether a classification based solely on woody species leads to units similar to the Braun-Blanquet system or to something completely different. Study area: Austria (Central Europe) and adjacent regions. Methods: 23,681 forest relevés from the Austrian Vegetation Database were classified using TWINSPAN. Spruce and pine plantations and stands with a cover of non-native woody species > 5% were excluded from the dataset. Only native tree and shrub species were used in the classification while herbs, dwarf shrubs, cryptogams and all records of woody species in the herb layer were omitted. Results: The TWINSPAN classification revealed elevation (i.e., climate) as the main floristic gradient in the data set. Within lowland communities, soil moisture was the dominant factor. The higher units of the Braun-Blanquet system were mostly well reproduced. Conclusions: The higher levels of the phytosociological forest classification (class, order, partly also alliance) can basically be defined by taking only the shrub and tree layer into account. However, all past and current classifications suffer from arbitrary exceptions to this rule. This leads to many inconsistencies and blurs the main biogeographical patterns within European forests. Here I argue that using the tree and shrub species for defining the higher levels and the understorey species for defining the lower ones is best suited to meet the properties that users would expect from a good forest classification.

Taxonomic reference: Fischer et al. (2008).

Syntaxonomic reference: Mucina et al. (2016) if not stated otherwise.

Abbreviations: EVC = EuroVegChecklist (Mucina et al. 2016).

Braun-Blanquet approach, forest, shrub layer, tree layer, vegetation classification

The classes of the Braun-Blanquet system correspond to major floristic, biogeographical and ecological units (Pignatti et al. 1995; Loidi 2020). For European zonal forests, these are the Quercetea ilicis (mediterranean evergreen), Quercetea pubescentis (submediterranean deciduous), Quercetea robori-petraeae and Carpino-Fagetea (temperate deciduous), Vaccinio-Piceetea (boreal and temperate montane–subalpine coniferous) and Betulo-Alnetea viridis (subarctic-subalpine deciduous) (Figure 1). Azonal forests can be arranged into two groups: Erico-Pinetea, Pyrolo-Pinetea and Junipero-Pinetea include coniferous forests on very dry sites, while Salicetea purpureae, Alno-Populetea and Alnetea glutinosae are wetland forests. Finally, the Crataego-Prunetea, Franguletea and Robinietea comprise seral woodland and tall-scrub (Mucina et al. 2016). These classes can be incorporated into a global formation system, enabling broad-scale comparisons among continents (Willner and Faber-Langendoen 2021).

Figure 1. 

Main zonal formations of Europe (following Bohn et al. 2000) and corresponding EVC classes. Orange: mediterranean sclerophyllous forests and scrub (Quercetea ilicis); yellow-green: submediterranean deciduous broad-leaved forests (Quercetea pubescentis); dark green: cool-temperate deciduous broad-leaved forests (Quercetea robori-petraeae, Carpino-Fagetea); lilac: boreal, hemiboreal and temperate-montane coniferous and mixed broad-leaved-coniferous forests (Vaccinio-Piceetea); pale pink: subarctic and temperate-subalpine open woodland and scrub (Betulo-Alnetea viridis).

For most of the 20^th century, the correspondence between classes and formations was much weaker because some classes (Vaccinio-Piceetea, Betulo-Adenostyletea, Epilobietea) included both forest and non-forest vegetation (e.g., Oberdorfer 1957). Splitting these physiognomically heterogeneous classes has been identified as one of the megatrends in phytosociology during the last 50 years (Guarino et al. 2018; Willner and Faber-Langendoen 2021). However, many uncertainties and inconsistencies still linger within the classes, blurring the biogeographical and ecological correspondences outlined above. For instance, the EuroVegChecklist (EVC, Mucina et al. 2016) classifies Western Caucasian Pinus brutia forests (Campanulo sibiricae-Pinion brutiae) in the Quercetea pubescentis while the Pinus brutia forests of the Eastern Mediterranean are included in the Quercetea ilicis (but see Bonari et al. 2021 for a different solution). Boreal-subarctic birch woods on nutrient-poor podzolic soils (Empetro hermaphroditi-Betulion pumilae) are included in the Vaccinio-Piceetea, those on nutrient-rich soils (Geranio sylvatici-Betulion pumilae) in the Betulo-Alnetea viridis. Temperate pine forests on acidic soils (Dicrano-Pinion sylvestris) are classified within the Vaccinio-Piceetea by the EVC, whereas Willner and Grabherr (2007) assign them to the Erico-Pinetea. Acidophytic beech forests are placed in the Carpino-Fagetea by some authors, but in the Quercetea robori-petraeae by others (Willner 2002; see also remark fag03 in Mucina et al. 2016, p. 35). High montane acidophytic beech forests of Central Europe (Calamagrostio villosae-Fagetum) were even assigned to the Vaccinio-Piceetea in Oberdorfer (1992). Temperate Abies alba forests are either considered as part of the Carpino-Fagetea (e.g., Chytrý 2013), or as part of the Vaccinio-Piceetea (e.g., Willner and Grabherr 2007; Mucina et al. 2016), whereas some authors split them between the two classes (Oberdorfer 1992).

All these examples have one question in common: Should one give higher weight to the tree layer or the herb layer composition when classifying forests? Floristic similarity is the main criterion in the Braun-Blanquet approach (Westhoff and Van der Maarel 1978). However, floristic similarity of the tree layer might suggest a different grouping than floristic similarity of the herb layer . As the European tree flora is rather poor in species compared to other continents (Leuschner and Ellenberg 2017), overall similarity is usually driven by the herb layer. Thus, following a purely floristic approach, communities dominated by the same tree species may end up in different classes, orders or alliances (Grabherr et al. 2003). Because such units are often very heterogenous in terms of physiognomy and at odds with broad-scale formations and biogeographical units, most authors consciously or unconsciously give higher weight to the tree species composition in at least some cases. However, in the absence of a general rule, these decisions are mostly subjective and arbitrary, resulting in a low stability of the forest classification in Europe.

Similar problems arise from the shrub layer, especially for communities without a tree layer. In the past, shrub communities were either joined with forests or with herb vegetation: The Prunetalia spinosae were part of the Querco-Fagetea, the Sambucetalia racemosae part of the Epilobietea angustifolii, the Betulo-Alnetea viridis included in the Betulo-Adenostyletea etc. (Oberdorfer 1992; Mucina et al. 1993). Whether shrub communities should be separated from forests on a high syntaxonomic level is still a controversial issue (e.g., Loidi 2020). Indeed, while the classes Crataego-Prunetea and Betulo-Alnetea viridis have been widely accepted in recent decades, they still include communities dominated by either shrubs or trees (see Mucina et al. 2016). The distinction between trees and tall shrubs is not always straightforward as many woody species have a rather high phenotypic plasticity. For instance, Fagus sylvatica often has a prostrate growth form at its upper elevational limit (Willner 2002), Quercus pubescens may be less than 4 m tall on dry sites with shallow soils (Jakucs 1961), while Corylus avellana can achieve 10 m in height despite having a shrubby branching pattern (Schütt and Lang 2014). Therefore, a priori separation of forests and tall scrub is inevitably at odds with the floristic principle (see also Willner and Faber-Langendoen 2021).

Interestingly, the traditional classification of tall shrub vegetation is almost exclusively based on the species composition of the shrub layer (Tüxen 1952; Weber 1997; Willner and Grabherr 2007). Herbs and grasses are often confined to the fringes and gaps of the scrub while the dense interior is almost completely devoid of a herb layer (Weber 1999). Therefore, non-woody species in relevés mainly reflect the neighbouring fringe and grassland vegetation, which belongs to a different successional stage, and their presence is strongly dependent on the delimitation of the sampling plot. Since the classification of scrub communities should reflect the ecological and biogeographical properties of the dominant shrubs rather than those of adjacent herbs and grasses, it makes sense to ignore the latter in the delimitation of higher syntaxa.

In an effort to increase the consistency of the Central European forest classification, Willner and Grabherr (2007) adopted an approach for the definition of higher forest syntaxa that was similar to that of tall shrub vegetation; that is, they suggested that the upper units of the system should be primarily based on the species composition of the tree layer, whereas the lower ones should be primarily based on the understorey composition (Willner 2017). Depending on the ecological amplitude of the dominant trees, the switch between over- and understorey was done at different hierarchical levels (e.g., between order and alliance for spruce forests, and between alliance and suballiance for most deciduous forests). However, the principle was not rigorously applied using numerical methods.

In the present study, I investigate whether a classification of Central European forests based solely on the woody species of the shrub and tree layer leads to units similar to the traditional Braun-Blanquet system or to something completely different.

The plot records (relevés) used in this study are from Austria (Central Europe) and adjacent regions in the SE Alps and NW Dinaric mountains (Figure 2). Austria covers most of the Eastern Alps and their foreland, the western part of the Pannonian Basin and the SE part of the Bohemian Massif. The elevation of the plots ranges from 120 m a.s.l. in Eastern Austria to the highest forests in the Alps at ca. 2300 m a.s.l. Annual precipitation ranges from 500 mm to 2000 mm (in the SE Alps locally up to 3000 mm). The mean annual temperature ranges from 1–2°C at the treeline to 10°C in the Pannonian lowland.

Figure 2. 

Plot locations in Austria (green dots) and adjacent areas (green numbers, indicating the number of plots from northern Italy, Slovenia, and Croatia, respectively).

Due to the large climatic gradient Austria has a large variety of forest types, and forests cover 46% of the country (ca. 3.88 million hectares) . Lowland forests are mostly deciduous, and oaks (Quercus spp.), hornbeam (Carpinus betulus), beech (Fagus sylvatica) and ash (Fraxinus excelsior) are dominant trees. The outer ranges of the Alps are occupied by mixed forests composed predominantly of beech and fir (Abies alba). The inner parts of the Alps, which have a strongly continental climate, and the whole subalpine belt are covered by coniferous forests with spruce (Picea abies), larch (Larix decidua), and Arolla pine (Pinus cembra) as dominants. The upper subalpine belt is often dominated by Pinus mugo krummholz (Mayer 1974).

The TWINSPAN classification resulted in 63 clusters (one division failed because the minimum group size was not reached). With a few exceptions, lowland forests and scrubs were separated from those at higher elevations at the first level of division. At the second division level, lowland communities were further divided along a moisture gradient, and montane communities were separated from subalpine ones (Table 1).

Specifically, the TWINSPAN clusters corresponded to the following vegetation types (numbers in brackets refer to the column number in Table 1 and Suppl. material 1, syntaxa follow the EVC system; the clusters are numbered from 1 to 64 to show the full TWINSPAN hierarchy; note that there is no cluster 48 because the corresponding level 6 division failed):

1–8 (1): nutrient-rich willow carrs with Salix cinerea (Salicion cinereae p.p.)

9–12 (2): submontane and montane alluvial willow scrub (Salicion eleagno-daphnoidis)

13 (3): alluvial forests with Salix alba (Salicion albae p.p.)

14 (4): alluvial forests with Salix fragilis (Salicion albae p.p.)

15–16 (5): lowland alluvial scrub with Salix triandra (Salicion triandrae)

17–20 (6): swamp forests with Alnus glutinosa (Alnion glutinosae)

21 (7): alluvial forests with Populus alba (Alnion incanae p.p.)

22 (8): alluvial forests with Alnus incana (Alnion incanae p.p.)

23 (9): alluvial forests with Alnus glutinosa (Alnion incanae p.p.)

24 (10): sycamore forests (Tilio-Acerion)

25 (11): moist oak-hornbeam forests with Quercus robur (Carpinion betuli p.p.)

26 (12): lime forests and mesic oak-hornbeam forests with Fraxinus excelsior (Melico-Tilion platyphylli, Carpinion betuli p.p.)

27 (13): mesic and dry oak-hornbeam forests with Quercus petraea (Carpinion betuli p.p.)

28 (14): acidophytic oak forests with Quercus petraea (Agrostio-Quercion petraeae)

29 (15): thermophilous oak forests on deeper soils (Quercion petraeae, Quercion pubescenti-petraeae p.p.)

30 (16): thermophilous oak forests on shallow soils with Quercus pubescens (Quercion pubescenti-petraeae p.p.)

31 (17): thermophilous seral scrub (Berberidion vulgaris, Urtico-Crataegion)

32 (18): lowland alluvial hardwood forests (Fraxino-Quercion roboris)

33–36 (19): beech forests (Fagetalia sylvaticae, Luzulo-Fagetalia sylvaticae)

37–38 (20): spruce forests (Piceetalia excelsae, Athyrio filicis-feminae-Piceetalia)

39–40 (21): montane elder scrub in forest clearings (Sambuco-Salicion capreae)

41 (22): Pinus sylvestris forests (Erico carneae-Pinion, Dicrano-Pinion sylvestris, Vaccinio uliginosi-Pinion sylvestris)

42 (23): Pinus nigra forests (Erico-Fraxinion orni)

43–44 (24): dry calcareous Ostrya carpinifolia forests on shallow soils (Fraxino orni-Ostryion)

45–47 (25): nutrient-poor willow carrs with Salix aurita (Salicion cinereae p.p.)

49–52 (26): subalpine krummholz with Pinus mugo (Pinion mugo, Erico-Pinion mugo)

53–54 (27): subalpine Larix decidua woodland (Piceion excelsae p.p.)

55–56 (28): subalpine Pinus cembra woodland (Piceion excelsae p.p.)

57–64 (29): subalpine green alder scrub (Alnion viridis)

The relevés used in this study are available upon request from the Austrian Vegetation Database (GIVD-ID EU-AT-001) managed by the author of this paper and from the European Vegetation Archive (https://euroveg.org/eva-database/).

I thank Don Faber-Langendoen, Jim Martin and two anonymous reviewers for their valuable comments on previous versions of the manuscript.

• Wolfgang Willner (Corresponding author, wolfgang.willner@univie.ac.at), ORCID: https://orcid.org/0000-0003-1591-8386