How to classify forests? A case study from Central Europe

Abstract

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).

Full text

How to classify forests?
A case study from Central Europe
Wolfgang Willner1,2
1 Department of Botany and Biodiversity Research, University of Vienna, Vienna, Austria
2 Vienna Institute for Nature Conservation & Analyses, Vienna, Austria
Corresponding author: Wolfgang Willner (wolfgang.willner@univie.ac.at)
Academic editor: Jorge Capelo
Received 21 December 2023♦Accepted 24 January 2024♦Published 1 March 2024
Abstract
Aims: Inconsistent treatment of the vegetation layers is one of the main problems in the oristic classication of forests.
In this study I investigate whether a classication based solely on woody species leads to units similar to the Braun-Blan-
quet system or to something completely dierent. Study area: Austria (Central Europe) and adjacent regions. Methods:
23,681 forest relevés from the Austrian Vegetation Database were classied using TWINSPAN. Spruce and pine plan-
tations 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 classication while herbs, dwarf shrubs, cryptogams and all records of woody species
in the herb layer were omitted. Results: e TWINSPAN classication revealed elevation (i.e., climate) as the main
oristic gradient in the data set. Within lowland communities, soil moisture was the dominant factor. e higher units
of the Braun-Blanquet system were mostly well reproduced. Conclusions: e higher levels of the phytosociological
forest classication (class, order, partly also alliance) can basically be dened by taking only the shrub and tree layer into
account. However, all past and current classications suer from arbitrary exceptions to this rule. is leads to many
inconsistencies and blurs the main biogeographical patterns within European forests. Here I argue that using the tree
and shrub species for dening the higher levels and the understorey species for dening the lower ones is best suited to
meet the properties that users would expect from a good forest classication.
Taxonomic reference: Fischer et al. (2008).
Syntaxonomic reference: Mucina et al. (2016) if not stated otherwise.
Abbreviations: EVC = EuroVegChecklist (Mucina et al. 2016).
Keywords
Braun-Blanquet approach, forest, shrub layer, tree layer, vegetation classication
Introduction
e classes of the Braun-Blanquet system correspond
to major oristic, biogeographical and ecological units
(Pignatti et al. 1995; Loidi 2020). For European zonal
forests, these are the Quercetea ilicis (mediterranean ev-
ergreen), Quercetea pubescentis (submediterranean de-
ciduous), Quercetea robori-petraeae and Carpino-Fagetea
(temperate deciduous), Vaccinio-Piceetea (boreal and
temperate montane–subalpine coniferous) and Betulo-Al-
netea viridis (subarctic-subalpine deciduous) (Figure 1).
Azonal forests can be arranged into two groups: Erico-Pin-
etea, Pyrolo-Pinetea and Junipero-Pinetea include conifer-
ous forests on very dry sites, while Salicetea purpureae,
Alno-Populetea and Alnetea glutinosae are wetland forests.
Finally, the Crataego-Prunetea, Franguletea and Robinietea
Copyright Wolfgang Willner. This 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.
Vegetation Classification and Survey 5: 17–26
doi: 10.3897/VCS.117703
RESEARCH PAPER
International Association for Vegetation Science (IAVS)
CLASSIFICATION OF EUROPEAN FORESTS

Wolfgang Willner: How to classify forests?18
comprise seral woodland and tall-scrub (Mucina et al.
2016). ese classes can be incorporated into a global for-
mation system, enabling broad-scale comparisons among
continents (Willner and Faber-Langendoen 2021).
For most of the 20th century, the correspondence be-
tween classes and formations was much weaker because
some classes (Vaccinio-Piceetea, Betulo-Adenostyletea,
Epilobietea) included both forest and non-forest vegeta-
tion (e.g., Oberdorfer 1957). Splitting these physiognom-
ically heterogeneous classes has been identied 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 biogeograph-
ical and ecological correspondences outlined above. For
instance, the EuroVegChecklist (EVC, Mucina et al. 2016)
classies Western Caucasian Pinus brutia forests (Cam-
panulo sibiricae-Pinion brutiae) in the Quercetea pubescen-
tis while the Pinus brutia forests of the Eastern Mediter-
ranean are included in the Quercetea ilicis (but see Bonari
et al. 2021 for a dierent solution). Boreal-subarctic birch
woods on nutrient-poor podzolic soils (Empetro hermaph-
roditi-Betulion pumilae) are included in the Vaccinio-Picee-
tea, those on nutrient-rich soils (Geranio sylvatici-Betulion
pumilae) in the Betulo-Alnetea viridis. Temperate pine for-
ests on acidic soils (Dicrano-Pinion sylvestris) are classied
within the Vaccinio-Piceetea by the EVC, whereas Willner
and Grabherr (2007) assign them to the Erico-Pinetea. Ac-
idophytic 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 for-
ests 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 ap-
proach (Westho and Van der Maarel 1978). However, o-
ristic similarity of the tree layer might suggest a dierent
grouping than oristic similarity of the herb layer . As the
European tree ora is rather poor in species compared to
other continents (Leuschner and Ellenberg 2017), overall
similarity is usually driven by the herb layer. us, follow-
ing a purely oristic approach, communities dominated by
the same tree species may end up in dierent classes, or-
ders or alliances (Grabherr et al. 2003). Because such units
are oen very heterogenous in terms of physiognomy and
at odds with broad-scale formations and biogeographical
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 (Quer-
cetea 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).

Vegetation Classification and Survey 19
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, result-
ing in a low stability of the forest classication 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: e 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 sep-
arated 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). e distinction between trees and
tall shrubs is not always straightforward as many woody
species have a rather high phenotypic plasticity. For in-
stance, Fagus sylvatica oen has a prostrate growth form
at its upper elevational limit (Willner 2002), Quercus pu-
bescens 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). erefore, a priori separation of
forests and tall scrub is inevitably at odds with the oristic
principle (see also Willner and Faber-Langendoen 2021).
Interestingly, the traditional classication of tall shrub
vegetation is almost exclusively based on the species com-
position of the shrub layer (Tüxen 1952; Weber 1997;
Willner and Grabherr 2007). Herbs and grasses are oen
conned to the fringes and gaps of the scrub while the
dense interior is almost completely devoid of a herb lay-
er (Weber 1999). erefore, non-woody species in relevés
mainly reect the neighbouring fringe and grassland vege-
tation, which belongs to a dierent successional stage, and
their presence is strongly dependent on the delimitation
of the sampling plot. Since the classication of scrub com-
munities should reect the ecological and biogeographical
properties of the dominant shrubs rather than those of ad-
jacent herbs and grasses, it makes sense to ignore the latter
in the delimitation of higher syntaxa.
In an eort to increase the consistency of the Cen-
tral European forest classication, Willner and Grabherr
(2007) adopted an approach for the denition of higher
forest syntaxa that was similar to that of tall shrub vegeta-
tion; that is, they suggested that the upper units of the sys-
tem should be primarily based on the species composition
of the tree layer, whereas the lower ones should be primar-
ily 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 dierent hierarchical levels (e.g., between order and al-
liance for spruce forests, and between alliance and subal-
liance for most deciduous forests). However, the principle
was not rigorously applied using numerical methods.
In the present study, I investigate whether a classi-
cation 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 dierent.
Study area
e 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 cov-
ers most of the Eastern Alps and their foreland, the west-
ern part of the Pannonian Basin and the SE part of the
Bohemian Massif. e 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). e mean annual temperature ranges from
1–2°C at the treeline to 10°C in the Pannonian lowland.
Due to the large climatic gradient Austria has a large
variety of forest types, and forests cover 46% of the coun-
try (ca. 3.88 million hectares) . Lowland forests are mostly
deciduous, and oaks (Quercus spp.), hornbeam (Carpi-
nus betulus), beech (Fagus sylvatica) and ash (Fraxinus
excelsior) are dominant trees. e outer ranges of the Alps
are occupied by mixed forests composed predominantly
of beech and r (Abies alba). e 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. e upper subalpine belt is oen
dominated by Pinus mugo krummholz (Mayer 1974).
Methods
Dataset preparation
Initially, all relevés of forest and shrub communities were
selected from the Austrian Vegetation Database (GIVD-ID
EU-AT-001; Willner et al. 2012). Spruce and pine planta-
tions and stands with a cover of non-native woody species
> 5% were excluded from the dataset. Also excluded were
forest relevés where the cover of trees was not estimated
separately for the tree and herb layer, relevés with a cover
of woody species determined only at the genus level > 5%,
and relevés dominated by (>25%) low shrubs [i.e., shrub
species not exceeding 2 m, including all Rubus species].
e 2 m threshold was chosen following the denition
of the forest and woodland formation class (Willner and
Faber-Langendoen 2021). Finally, relevés with a total cover
of trees and tall shrubs < 15% were omitted. is resulted in
a dataset of 23,681 relevés, with 22,588 plots from Austria
and 1,093 plots from neighbouring countries (Figure 2).
Only native tree and tall shrub species in the shrub and
tree layer were used in the classication while all other taxa
(including records of woody species in the herb layer and

Wolfgang Willner: How to classify forests?20
taxa determined only at the genus level) were omitted. e
omission of non-native trees and shrubs follows the con-
sideration that the syntaxonomic system of European for-
est and shrub communities should be based on the native
species (though syntaxa for communities dominated by
non-native species might be added in a second step). Re-
cords of native tree and tall shrub species in dierent lay-
ers were merged using the algorithm published by Fischer
(2015). Altogether, 111 taxa were kept in the nal dataset.
All data handling was done with JUICE 7.1 (Tichý 2002).
Numerical classification
e matrix of 23,681 relevés and 111 taxa was classied using
the original TWINSPAN algorithm (Hill 1979). Parameter
settings were three pseudospecies cutlevels (0%, 5%, 25%),
six levels of division and a minimum group size for division
of two. For species sorting and interpretation, the diagnostic
value of woody species for phytosociological classes accord-
ing to Mucina et al. (2016) was used. If a species was given as
diagnostic for two or more classes occurring in Central Eu-
rope, the diagnostic value according to Willner and Grabherr
(2007) was followed. Within each class, species were sorted
by decreasing phi coecient (Chytrý et al. 2002; Tichý and
Chytrý 2006) using a threshold of 0.3. e phi coecient was
calculated assuming equal group size, and positive phi values
were only accepted if the dierence in species constancy be-
tween the target unit and the rest of the data set was signi-
cant according to Fisher’s exact test at p < 0.05.
Results
e TWINSPAN classication 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 rst level of division. At the second division level, low-
land communities were further divided along a moisture
gradient, and montane communities were separated from
subalpine ones (Table 1).
Specically, 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.)
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).

Vegetation Classification and Survey 21
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 (Quer-
cion 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-Quer-
cion roboris)
33–36 (19): beech forests (Fagetalia sylvaticae, Luzulo-Fag-
etalia sylvaticae)
37–38 (20): spruce forests (Piceetalia excelsae, Athyrio
licis-feminae-Piceetalia)
39–40 (21): montane elder scrub in forest clearings (Sam-
buco-Salicion capreae)
41 (22): Pinus sylvestris forests (Erico carneae-Pinion, Dicra-
no-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)
Discussion
Syntaxonomy
e TWINSPAN classication revealed elevation (i.e., cli-
mate) as the main oristic gradient in the data set. Within
lowland communities, soil moisture was the dominant
factor. Interestingly, the higher units of the Braun-Blan-
quet system were mostly well reproduced, with clusters
1–8 corresponding to the Franguletea, clusters 9–16 to the
Salicetea purpureae, clusters 17–20 to the Alnetea glutinos-
ae, clusters 21–23 to the Alno-Populetea, clusters 24–27 to
the Carpino-Fagetea and so on. Notable exceptions are the
classes Quercetea pubescentis, Quercetea robori-petraeae
and Crataego-Prunetea, which were all intermingled with
the Carpino-Fagetea. is could be interpreted as sup-
port for the more traditional concept of a broadly dened
class Querco-Fagetea (e.g., Oberdorfer 1992; Loidi 2020).
However, because the Quercetea pubescentis and Querce-
tea robori-petraeae have their main distribution outside
the study area, this question will not be further discussed
in the present paper. e strange position of the Fraxino
orni-Ostryion in the TWINSPAN table reects the fact
that Ostrya carpinifolia forests reach their northern dis-
tribution limit in the study area, where they are conned
to dry, calcareous sites similar to those of pine forests. In-
deed, Mucina et al. (1993) classied these communities
within the class Erico-Pinetea. Alluvial hardwood forests
(Fraxino-Quercion roboris) were widely separated from
the Alnus and Populus woods of the Alnion incanae, which
suggests keeping them in the class Carpino-Fagetea. Lime
forests (Melico-Tilion platyphylli) were grouped togeth-
er with oak-hornbeam forests (Carpinion betuli). Abies
alba forests were not reproduced as a separate cluster, but
mostly included in beech forests, supporting the concept
of Chytrý (2013). e position of subalpine Larix decidua
and Pinus cembra woodland seems at odds with the EVC
system, but it ts well with the classication in Willner
and Grabherr (2007), where both units were included in a
broadly dened Pinion mugo.
On the whole, the traditional Braun-Blanquet system
of forests seems to have given more weight to the tree
species combination than is generally acknowledged in
textbooks. As expected, the syntaxonomic rank of the
TWINSPAN clusters varies vastly, from a single asso-
ciation (e.g., cluster 13: Salicetum albae) to a group of
classes (cluster 41: Pinus sylvestris forests). is reects
the dierent ecological amplitude of the dominant spe-
cies. In most cases, however, the woody species combi-
nation seems most suitable for the denition of orders
and alliances. Some ecological gradients (e.g., calcare-
ous versus acidic soils) are only visible in the herb layer
(including dwarf shrubs) and are therefore not reected
in the table.
What do we expect from a good forest classifi-
cation?
Loidi (2020) suggested three criteria for a “good” phy-
tosociological class: (1) biogeographical-evolutionary
criterion: common origin and evolution, (2) oristic
criterion: common set of characteristic species, and (3)
application criterion: coherence in the presentation. For

Wolfgang Willner: How to classify forests?22
Table 1. Synoptic table of the TWINSPAN result. Values are percentage constancy. Bold values indicate an average cover > 14%, grey shading indicates high fidelity (i.e., phi > 0.3).
Species are sorted according to their diagnostic value for classes. The vertical lines represent the division level 2. Groups 21 and 25 (with 14 and 10 relevés, respectively) and species
not reaching 10% constancy in any column are omitted. The full table is provided in Suppl. material 1.
Group number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 22 23 24 26 27 28 29
TWINSPAN level 3 4 6 6 5 4 6 6 6 6 6 6 6 6 6 6 6 6 4 5 6 6 5 4 6 6 3
Number of relevés 121 307 318 110 109 855 250 833 1087 1013 393 1847 653 397 749 546 299 96 5232 5090 1181 255 327 1000 149 133 307
Franguletea
Salix cinerea 95 1 3 3 4 14 1 3 2 1 1 . . . . . 1 . . 1 1 . . 1 . . .
Salicetea purpureae
Salix eleagnos .57 4 9 3 . 1 5 3 5 . 1 . . 1 1 1 . 1 1 3 1 3 1 . . 1
Myricaria germanica . 17 . . . . . 1 . . . . . . . . . . . . 1 . . . . . .
Salix myrsinifolia . 21 5 . 1 1 1 6 1 1 . . . . . . 1 . . 1 1 . . 1 . . 1
Salix daphnoides . 15 3 . 2 . . 2 1 1 . . . . . . . . . 1 . . . 1 . . 1
Salix fragilis 19 6 20 91 9 9 2 3 12 1 1 1 . . . . . 1 . . . . . . . . .
Salix × rubens 1 1 2 24 6 2 1 . 1 . 1 . . . . . . . . . . . . . . . .
Salix triandra 1 10 14 30 92 1 . 1 1 . . . . . . . . . . 1 . . . . . . .
Salix viminalis . 3 8 14 39 1 . 1 1 1 . 1 . . . . . . . . 1 . . . . . .
Salix purpurea 182 31 57 33 1 2 6 6 1 1 1 . 1 1 . 1 6 . 1 1 . . 1 . . .
Salix alba 5 16 97 32 62 2 25 8 7 1 1 1 . 1 1 . 1 2 1 . 1 . . . . . .
Alnetea glutinosae
Alnus glutinosa 27 2 4 34 399 9 2 50 7 10 5 1 1 1 . 1 5 1 1 1 . . 1 . . .
Frangula alnus 9 4 3 1 . 30 6 8 8 3 22 4 1 23 8 4 4 18 3 4 17 2 1 3 . . .
Alno-Populetea albae
Populus nigra . 18 36 12 6 1 22 5 4 1 1 1 1 . 1 . 1 7 1 . . . . . . . .
Populus alba 1 1 4 . 1 2 62 3 7 1 2 1 1 1 2 1 1 33 1 1 . . . . . . .
Alnus incana 2 43 39 35 4 4 46 99 24 26 2 1 . 1 1 . . . 1 6 2 . 1 1 . 3 2
Prunus padus 4 6 27 17 . 36 50 40 72 11 8 7 1 1 4 1 3 7 1 1 1 . . 1 . . .
Ulmus minor . . 3 . . 1 14 1 3 1 12 3 1 1 19 4 16 70 1 . . . . . . . .
Fraxinus angustifolia 1 . 1 1 1 2 6 1 1 . 11 1 1 . 1 . . 32 .........
Viburnum opulus 7 4 17 9 1 17 18 15 25 7 7 6 1 2 5 1 5 8 1 1 1 1 . . . . .
Ulmus laevis . . 9 8 . 1 30 2 9 1 4 2 1 . 1 . 1 24 1 1 . . . . . . .
Crataego-Prunetea
Cornus sanguinea 2 5 36 18 3 7 90 31 48 11 22 22 4 13 51 44 44 78 1 1 2 2 1 . . . .
Ligustrum vulgare 1 1 4 1 . 4 15 6 16 1 12 13 12 18 76 53 60 45 1 1 7 4 4 . . . .
Viburnum lantana . 1 2 1 . 1 2 6 9 7 3 14 2 2 43 53 22 2 1 1 12 5 10 1 . . .
Prunus spinosa 1 . 1 . . 1 5 1 1 1 7 1 1 5 18 12 68 21 1 . 1 1 . . . . .
Rosa canina agg. 1 1 1 1 . 1 8 1 3 1 4 3 4 9 24 29 77 17 1 1 1 4 1 . . . .
Juniperus communis . 1 . . . 1 . 1 . 1 2 1 1 4 5 8 4 . 1 2 53 7 7 3 . 1 1
Amelanchier ovalis . . . . . . . . . 1 1 2 1 . 1 33 . . 1 1 41 70 30 6 . . .
Cotoneaster tomentosus . . . . . . . . . . . 1 . . . 6 . . 1 1 7 25 15 1 . . .
Crataegus monogyna 1 2 6 1 2 2 30 12 17 5 22 22 12 17 73 68 74 64 1 1 4 8 4 . . . .
Corylus avellana . 2 3 3 . 5 3 9 49 62 55 49 9 40 30 29 9 23 11 4 9 5 8 1 . . 1
Berberis vulgaris . 1 3 . . 3 3 10 6 6 5 7 1 7 17 47 11 . 2 2 34 47 9 1 . . .
Euonymus europaeus 2 1 10 9 1 13 27 19 33 6 12 12 3 2 25 12 37 18 1 1 1 . 1 . . . .
Rhamnus cathartica . . 1 1 . 9 3 1 2 1 5 5 2 3 14 35 30 19 1 1 3 3 3 . . . .
Prunus mahaleb . . . . . . . . . . . 1 . . 1 13 7 . . . 1 1 . . . . .
Quercetea pubescentis
Quercus cerris . . . . . . . . . 1 3 5 28 11 48 9 6 3 1 1 . . . . . . .
Quercus pubescens . . . . . . . . . . 1 2 1 . 26 76 4 1 1 . 1 6 3 . . . .
Cornus mas . . . . . . 1 . 1 1 3 16 9 1 39 64 3 3 1 1 . 9 . . . . .

Vegetation Classification and Survey 23
Group number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 22 23 24 26 27 28 29
TWINSPAN level 3 4 6 6 5 4 6 6 6 6 6 6 6 6 6 6 6 6 4 5 6 6 5 4 6 6 3
Number of relevés 121 307 318 110 109 855 250 833 1087 1013 393 1847 653 397 749 546 299 96 5232 5090 1181 255 327 1000 149 133 307
Euonymus verrucosus . . . . . . . 1 1 1 2 10 3 2 20 38 2 2 1 . 1 . 3 . . . .
Ostrya carpinifolia . . . . . . . . . . . 1 . . 2 3 . . 3 1 1 4 98 1 . . .
Fraxinus ornus . . . . . . 1 1 . . . 1 1 . 3 4 . . 2 1 3 13 83 . . . .
Sorbus aria agg. . . . . . 1 . . 1 1 2 14 7 6 9 63 1 . 11 3 30 69 55 6 1 . .
Sorbus torminalis . . . . . . . . 1 . 2 5 16 11 27 25 2 . 2 1 . 3 . . . . .
Pyrus pyraster . 1 . . . 1 . 1 1 1 5 2 4 9 15 15 1 9 1 1 1 1 1 . . . .
Quercetea roboris
Quercus robur 1 1 1 5 . 9 10 2 23 4 88 22 4 11 30 7 13 63 3 1 5 . 3 . . . .
Quercus petraea . . . . . 1 . . . 1 9 24 87 95 54 11 1 . 11 1 5 1 6 . . . .
Carpino-Fagetea
Ulmus glabra . 3 1 5 . 1 3 3 17 60 2 34 2 1 5 5 1 1 10 1 1 1 . . . . .
Acer pseudoplatanus . 6 2 1 . 6 2 18 49 89 13 59 4 2 9 9 3 4 45 10 9 7 6 12 9 . 12
Acer platanoides . 1 1 . . 1 . 1 4 6 1 29 3 1 5 9 1 2 3 1 1 1 1 . . . .
Tilia platyphyllos . . 1 . . 1 . 1 2 10 2 30 6 5 5 18 1 . 2 1 1 2 2 . . . .
Acer campestre . . 1 . . 1 8 1 9 3 24 34 34 4 53 34 17 58 2 1 1 . . . . . .
Fagus sylvatica . 1 . . . 2 . 1 5 39 15 61 46 16 5 17 . . 97 21 9 25 29 5 2 . 1
Fraxinus excelsior 1 10 9 16 . 35 43 39 91 88 17 74 13 3 45 57 9 21 16 1 2 7 1 1 . . .
Carpinus betulus . 1 1 1 . 5 1 1 16 9 64 69 92 54 36 18 3 16 6 1 1 . 2 . . . .
Lonicera xylosteum . 7 7 3 1 2 9 21 34 25 8 24 4 3 26 15 4 4 5 2 10 1 5 1 1 . 1
Tilia cordata . . 1 1 . 1 2 1 12 8 29 37 32 23 15 13 1 5 2 1 1 1 2 . . . .
Prunus avium . 1 1 1 . 1 1 2 8 5 23 24 17 17 28 9 20 5 3 1 1 . . . . . .
Crataegus laevigata . . . . . 1 . 1 2 1 9 12 11 4 19 14 2 14 1 1 1 . . . . . .
Staphylea pinnata . . . . . . . . 1 3 1 13 2 . 4 14 1 . 1 . . . . . . . .
Populus tremula . 1 1 . . 1 1 1 3 2 20 3 2 8 2 1 3 8 1 1 2 1 2 . 1 . 1
Vaccinio-Piceetea
Abies alba . . . . . 1 . 1 2 8 3 7 3 2 1 1 . . 50 26 1 1 1 5 7 . 1
Picea abies . 11 1 1 . 23 1 28 17 51 34 21 9 14 6 4 1 . 69 99 62 26 38 34 60 60 43
Larix decidua . 1 . . . 1 . 1 1 1 3 2 2 4 1 1 . . 15 48 17 2 11 23 99 61 29
Pinus cembra . . . . . . . . . . . . . . . . . . 1 7 1 . . 3 51 100 11
Pinus uncinata . . . . . . . . . . . . . . . . . . 1 1 11 . . 2 1 . .
Lonicera caerulea . . . . . . . 1 . 1 1 . . . . . . . 1 1 1 . 1 6 11 3 5
Robinietea
Sambucus nigra 3 6 40 45 2 15 46 41 62 45 12 18 3 2 13 1 30 13 4 1 1 . . . . . .
Erico-Pinetea
Pinus sylvestris 2 5 1 . . 3 . 2 1 1 20 5 8 52 12 8 1 1 12 15 93 23 42 1 . 1 .
Pinus nigra . . . . . . . . . . . 2 2 1 2 34 2 . 2 1 3 98 10 . . . .
Rhamnus fallax . . . . . . . . . 1 . . . . . . 1 . 1 1 . 1 15 1 . . .
Roso pendulinae-Pinetea mugo
Sorbus chamaemespilus . . . . . . . . . . . . . . . . . . 1 1 6 . . 22 25 2 2
Pinus mugo (s. str.) . 1 . . . 1 . . . . . 1 . . . . . . 1 4 11 1 5 98 72 29 7
Rosa pendulina . . . . . 1 . 1 1 2 . 1 . . . 1 . . 2 2 3 2 1 16 15 . 1
Betulo-Alnetea viridis
Alnus alnobetula . . . 1 . 1 . 1 . 1 . 1 . . . . . . 1 3 1 . 1 9 24 13 96
Salix appendiculata . 13 1 . . 1 . 3 1 5 1 1 . . 1 . 1 . 1 3 2 1 2 23 16 2 24
other species
Sorbus aucuparia 1 1 1 1 . 8 . 5 7 5 13 6 2 8 2 1 1 1 8 21 15 2 4 34 64 23 47
Betula pendula . 7 3 6 . 2 2 4 4 2 25 6 14 25 3 1 2 2 4 7 7 1 2 1 3 3 14

Wolfgang Willner: How to classify forests?24
the third criterion, he noted that “it is very dicult to ar-
gue, in a teaching context, that forests dominated by the
same species belong to dierent classes”. Obviously, the
second and third criterion can only be simultaneously
fullled if the class is oristically dened by the species of
the dominant layer. In this way, most European tree spe-
cies become character species on some hierarchical level
while in many traditional systems they are only treated as
companion species.
We might complement Loidi’s criteria by four gener-
al properties that users might reasonably expect from a
good forest classication: (a) e upper levels of the hi-
erarchy are more easily recognizable than the lower lev-
els. (b) e upper level units are more stable over time
in terms of vegetation history. (c) e factors shaping
global vegetation patters are reected on the upper lev-
els, while the factors responsible for regional and local
patterns are reected on the lower levels. (d) e upper
levels are consistent with global formation and biome
classications.
As shown above, the higher levels of the phytosocio-
logical forest system can basically be dened by taking
only the tall shrub and tree layer into account. However,
this has never been formulated as a rule, and all past and
current classications suer from arbitrary weighting of
the layers, leading to inconsistencies and blurring the
main biogeographical patterns within European forests.
e distribution of European tree species is mainly con-
trolled by broad climatic gradients as well as dierences
in soil moisture (Table 1) – the same ecological factors
that are reected in global biome and formation systems
(Walter 1976; Faber-Langendoen et al. 2016; Loidi et al.
2022). Dierences in calcium content, on the other hand,
have a more regional signicance, and are mostly visible
in the herb layer composition (Leuschner and Ellenberg
2017). It is therefore advisable to reect the latter on
lower hierarchical levels, e.g., by uniting basiphytic and
acidophytic beech forests in the same class or even or-
der (Moor 1978; Oberdorfer 1992; Willner and Grabherr
2007; Willner et al. 2017), or by transferring temperate
dry Pinus sylvestris forests on acidic bedrock (alliance
Dicrano-Pinion) from the Vaccinio-Piceetea to the Eri-
co-Pinetea – a solution that is also supported by numer-
ous common understorey species (Willner and Grabherr
2007). However, given the extremely broad amplitude of
Pinus sylvestris, it is also justied to classify the forests
dominated by this species in at least two dierent class-
es, Erico-Pinetea (temperate dry pine forests, including
the Erico-Pinion, Dicrano-Pinion, Ononido-Pinion and
other alliances) and Vaccinio uliginosi-Pinetea (boreal
and temperate wet pine forests, including the Vaccinio
uliginosi-Pinion and Cladonio stellaris-Pinion; see Erma-
kov and Morozova 2011). Submediterranean pine forests
dominated by Pinus nigra should probably be placed in a
separate class.
Understorey species may have markedly dierent
biogeographical histories than the tree species they are
currently associated with (Záveská et al. 2021; Willner
et al. 2023). us, we can assume that vegetation units
dened by tree species have been more stable over time
than syntaxa dened by species from dierent layers.
In fact, vegetation units solely dened by understorey
species can be completely independent of the tree layer
and even exist outside the forest. A classic example is
Braun-Blanquet’s Rhododendro-Vaccinion (Braun-Blan-
quet et al. 1939), which comprised both coniferous for-
ests and treeless dwarf shrub heaths. Carrying this idea
to extremes, Gillet (1988) proposed to independently
classify the herb, shrub and tree layers (see also Gillet
and Julve 2018). While agreeing with these authors on
the basic problem, I suggest a dierent and less radi-
cal solution: By using the tree and (tall-)shrub layer for
dening the upper levels and the herb and cryptogam
layer for dening the lower levels of the system, the
basic units (i.e., the associations) represent the whole
forest community. At the same time, over- and under-
storey composition are not mixed in an arbitrary and
oen confusing manner for the denition of the higher
units as in most traditional systems. However, it must be
emphasised that “upper” and “lower” level is meant in a
purely relative sense here. e lowest appropriate rank
to be dened by the tree and tall shrub layer depends
both on the ecological amplitude of the resulting units
and the oristic heterogeneity of the herb and cryptog-
am layer within these units; thus, it may vary from asso-
ciation or suballiance (though this will be uncommon)
to a group of classes (see examples above).
Previous proposals have suggested separating forests
and tall-scrub on the one hand and non-woody vegeta-
tion (including dwarf-shrub heaths) on the other hand as
two a-priori structural types in syntaxonomy (Bergmeier
et al. 1990; Dengler et al. 2005). If trees and tall-shrubs
are used to dene the classes of woody vegetation, this
separation becomes a natural component of the oristic
classication, without a sudden change of criteria. At the
same time, phytosociological classes dened by the com-
bination of woody species can be easily tted into a global
formation system (Willner and Faber-Langendoen 2021)
as well as in most biome systems (Mucina 2019; Keith et
al. 2022).
Data availability
e 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/).
Acknowledgements
I thank Don Faber-Langendoen, Jim Martin and two
anonymous reviewers for their valuable comments on
previous versions of the manuscript.

Vegetation Classification and Survey 25
References
Bergmeier E, Härdtle W, Mierwald U, Nowak B, Peppler C, Flintrop
T (1990) Vorschläge zur syntaxonomischen Arbeitsweise in
der Panzensoziologie. Kieler Notizen zur Panzenkunde in
Schleswig-Holstein und Hamburg 20: 92–103.
Bohn U, Gollub G, Hettwer C (2000) Karte der natürlichen Vegetation
Europas. Map of the natural vegetation of Europe. Bundesamt für
Naturschutz, Bonn-Bad Godesberg, DE.
Bonari G, Fernández-González F, Çoban S, Monteiro-Henriques T, Berg-
meier E, Didukh YP, Xystrakis F, Angiolini C, Chytrý K, … Chytrý
M (2021) Classication of the Mediterranean lowland to submon-
tane pine forest vegetation. Applied Vegetation Science 24: e12544.
https://doi.org/10.1111/avsc.12544
Braun-Blanquet J, Sissingh G, Vlieger J (1939) Klasse der Vaccinio-Picee-
tea [Prodromus der Panzengesellschaen 6]. Comité International
du Prodrome Phytosociologique, Montpellier, FR.
Chytrý M [Ed.] (2013) Vegetace České republiky 4. Lesní a křovinná
vegetace [Vegetation of the Czech Republic 4. Forest and scrub
vegetation]. Academia, Praha, CZ.
Chytrý M, Tichý L, Holt J, Botta-Dukát J (2002) Determination of di-
agnostic species with statistical delity measures. Journal of Vege-
tation Science 13: 79–90. https://doi.org/10.1111/j.1654-1103.2002.
tb02025.x
Dengler J, Berg C, Jansen F (2005) New ideas for modern phytosociolog-
ical monographs. Annali di Botanica 5: 193–210.
Ermakov N, Morozova O (2011) Syntaxonomical survey of boreal oligo-
trophic pine forests in northern Europe and Western Siberia. Applied
Vegetation Science 14: 524–536. https://doi.org/10.1111/j.1654-
109X.2011.01155.x
Faber-Langendoen D, Keeler-Wolf T, Meidinger D, Josse C, Weakley A,
Tart D, Navarro G, Hoagland B, Ponomarenko S, … Helmer E (2016)
Classication and description of world formation types. U.S. Depart-
ment of Agriculture, Forest Service, Rocky Mountain Research Sta-
tion [General Technical Report RMRS-GTR-346], Fort Collins, CO,
US, 222 pp. https://doi.org/10.2737/RMRS-GTR-346
Fischer HS (2015) On the combination of species cover values from
dierent vegetation layers. Applied Vegetation Science 18: 169–170.
https://doi.org/10.1111/avsc.12130
Fischer MA, Oswald K, Adler W (2008) Exkursionsora für Österreich,
Liechtenstein und Südtirol. 3. Auage. Biologiezentrum der Oberös-
terreichischen Landesmuseen, Linz, AT.
Gillet F (1988) L’approche synusiale intégrée des phytoeoenoses
forestières. Application aux forêts du Jura. Colloques Phytoso-
ciologiques 14: 81–92.
Gillet F, Julve P (2018) e integrated synusial approach to vegetation
classication and analysis. Phytocoenologia 48: 141–152. https://doi.
org/10.1127/phyto/2017/0164
Grabherr G, Reiter K, Willner W (2003) Towards objectivity in vegeta-
tion classication: the example of the Austrian forests. Plant Ecology
169: 21–34. https://doi.org/10.1023/A:1026280428467
Guarino R, Willner W, Pignatti S, Attorre F, Loidi JJ (2018) Spatio-tem-
poral variations in the application of the Braun-Blanquet approach
in Europe. Phytocoenologia 48: 239–250. https://doi.org/10.1127/
phyto/2017/0181
Hill M (1979) TWINSPAN – A FORTRAN program for arranging mul-
tivariate data in an ordered two-way table by classication of the in-
dividuals and attributes. Cornell University, Ithaca, NY, US.
Jakucs P (1961) Die phytozönologischen Verhältnisse der Flaumeichen-Bus-
chwälder Südostmitteleuropas. Akadémiai Kiadó, Budapest, HU.
Keith DA, Ferrer-Paris JR, Nicholson E, Bishop MJ, Polidoro BA,
Ramirez-Llodra E, Tozer MG, Nel JL, Mac Nally R, … Kingsford RT
(2022) A function-based typology for Earth’s ecosystems. Nature 610:
513–518. https://doi.org/10.1038/s41586-022-05318-4
Leuschner C, Ellenberg H (2017) Ecology of Central European Forests.
Vegetation Ecology of Central Europe, Volume I. Springer, Cham,
CH. https://doi.org/10.1007/978-3-319-43042-3
Loidi J (2020) e concept of vegetation class and order in phytosocio-
logical syntaxonomy. Vegetation Classication and Survey 1: 163–
167. https://doi.org/10.3897/VCS/2020/59977
Loidi J, Navarro-Sánchez G, Vynokurov D (2022) Climatic denitions of
the world’s terrestrial biomes. Vegetation Classication and Survey 3:
231–271. https://doi.org/10.3897/VCS.86102
Mayer H (1974) Wälder des Ostalpenraumes. Gustav Fischer, Stuttgart, DE.
Moor M (1978) Die Klasse der Eschen-Buchenwälder (Fraxino-
Fagetea). Phytocoenologia 4: 433–445. https://doi.org/10.1127/
phyto/4/1978/433
Mucina L (2019) Biome: evolution of a crucial ecological and bio-
geographical concept. New Phytologist 222: 97–114. https://doi.
org/10.1111/nph.15609
Mucina L, Grabherr G, Wallnöfer S [Eds] (1993) Die Panzenge-
sellschaen Österreichs, Teil III: Wälder und Gebüsche. Gustav
Fischer, Jena, DE.
Mucina L, Bültmann H, Dierßen K, eurillat J-P, Raus T, Čarni A,
Šumberová K, Willner W, Dengler J, … Tichý L (2016) Vegetation
of Europe: Hierarchical oristic classication system of vascular
plant, bryophyte, lichen, and algal communities. Applied Vegetation
Science 19(Suppl. 1): 3–264. https://doi.org/10.1111/avsc.12257
Oberdorfer E (1957) Süddeutsche Panzengesellschaen. Gustav Fisch-
er, Jena, DE.
Oberdorfer E [Ed.] (1992) Süddeutsche Panzengesellschaen, 2.
Auage, Teil IV, Wälder und Gebüsche. Gustav Fischer, Jena, DE.
Pignatti S, Oberdorfer E, Schaminée JHJ, Westho V (1995) On the con-
cept of vegetation class in phytosociology. Journal of Vegetation Sci-
ence 6: 143–152. https://doi.org/10.2307/3236265
Schütt P, Lang UM (2014) Corylus avellana. In: Rolo A, Weisgerber H,
Lang UM, Stimm B (Eds) Enzyklopädie der Holzgewächse: Hand-
buch und Atlas der Dendrologie. Wiley-VCH, Weinheim, DE.
Tichý L (2002) JUICE, soware for vegetation classification. Journal of Veg-
etation Science 13: 451–453. https://doi.org/10.1111/j.1654-1103.2002.
tb02069.x
Tichý L, Chytrý M (2006) Statistical determination of diagnostic spe-
cies for site groups of unequal size. Journal of Vegetation Science 17:
809–818. https://doi.org/10.1111/j.1654-1103.2006.tb02504.x
Tüxen R (1952) Hecken und Gebüsche. Mitteilungen der Geographi-
schen Gesellscha in Hamburg 50: 85–117.
Walter H (1976) Die ökologischen Systeme der Erde. Prinzipien ihrer
Gliederung mit Beispielen. Gustav Fischer, Stuttgart, DE.
Weber HE (1997) Hecken und Gebüsche in den Kulturlandschaen
Europas. Panzensoziologische Dokumentation als Basis für Schutz-
maßnahmen. Berichte der Reinhold-Tüxen-Gesellscha 9: 75–106.
Weber HE (1999) Rhamno-Prunetea (H2A). Schlehen- und Trauben-
holunder-Gebüsche. In: Dierschke H (Ed.) Synopsis der Panzenge-
sellschaen Deutschlands 5. Göttingen, DE.

Wolfgang Willner: How to classify forests?26
Westho V, van der Maarel E (1978) e Braun-Blanquet Approach. In:
Whittaker RH (Ed.) Classication of Plant Communities. Dr. W. Junk,
e Hague, NL, 287–399. https://doi.org/10.1007/978-94-009-9183-5_9
Willner W (2002) Syntaxonomische Revision der südmitteleu-
ropäischen Buchenwälder. Phytocoenologia 32: 337–453. https://doi.
org/10.1127/0340-269X/2002/0032-0337
Willner W (2017) Classication of the Austrian forest and shrub com-
munities: concepts, methods and experiences. Documents Phytoso-
ciologiques 6: 44–50.
Willner W, Grabherr G (2007) Die Wälder und Gebüsche Österreichs.
Ein Bestimmungswerk mit Tabellen. Spektrum Akademischer Ver-
lag, Heidelberg, DE.
Willner W, Faber-Langendoen D (2021) Braun-Blanquet meets
EcoVeg: a formation and division level classication of European
phytosociological units. Vegetation Classication and Survey 2: 275–
291. https://doi.org/10.3897/VCS/2021/71299
Willner W, Berg C, Heiselmayer P (2012) Austrian Vegetation Database.
In: Dengler J, Oldeland J, Jansen F, Chytrý M, Ewald J, Finckh M,
Glöckler F, Lopez-Gonzalez G, Peet RK, Schaminée JHJ (Eds) Veg-
etation databases for the 21st century. Biodiversity & Ecology 4: 333.
https://doi.org/10.7809/b-e.00125
Willner W, Jiménez-Alfaro B, Agrillo E, Biurrun I, Campos JA, Čarni A,
Casella L, Csiky J, Ćušterevska R, … Chytrý M (2017) Classication
of European beech forests: a Gordian Knot? Applied Vegetation Sci-
ence 20: 494–512. https://doi.org/10.1111/avsc.12299
Willner W, Wessely J, Gattringer A, Moser D, Záveská E, Dullinger S,
Schönswetter P, Hülber K (2023) Postglacial range formation of tem-
perate forest understorey herbs – insights from a spatiotemporally
explicit modelling approach. Global Ecology and Biogeography 32:
1046–1058. https://doi.org/10.1111/geb.13677
Záveská E, Kirschner P, Frajman B, Wessely J, Willner W, Gattringer
A, Hülber K, Lazić D, Dobeš C, Schönswetter P (2021) Evidence
for glacial refugia of the forest understorey species Helleborus niger
(Ranunculaceae) in the southern as well as in the northern Limestone
Alps. Frontiers in Plant Science 12: 683043. https://doi.org/10.3389/
fpls.2021.683043
E-mail and ORCID
Wolfgang Willner (Corresponding author, wolfgang.willner@univie.ac.at), ORCID: https://orcid.org/0000-0003-1591-8386
Supplementary material
Supplementary material 1
Full synoptic table
Link: https://doi.org/10.3897/VCS.117703.suppl1