Plant communities, diversity and endemism of the Kula Volcano, Manisa, Turkey

Abstract

Volcanoes often harbour specialized plant communities and shelter endemic plant species. Kula Volcano is one of fourteen volcanoes in Turkey. Although this volcano is clearly a landmark of the Aegean region, only few botanical studies analysed the vegetation pattern at the Kula Volcano. None performed a phytosociological classification to delimit different plant communities. We applied a stratified random sampling design according to altitude and aspect and sampled 112 vegetation plots. We classified plant community types using a modified TWINSPAN analysis followed by the determination of diagnostic species based on phi-coefficient fidelity values. Floristic relationships between plant community types were interpreted by ordination and ANOSIM analyses. Further, we used partial correlations of the ordination axes and environmental parameters in order to identify relationships between vegetation zonation and environment. We identified five major plant community types based on 85 diagnostic species. These plant community types were significantly correlated with altitude and aspect. Further, thirteen endemic plant species were found from which one was endangered and one was classified as vulnerable according to IUCN.

Full text

This article was downloaded by: [Deniz IŞIK GÜRSOY]
On: 02 February 2015, At: 10:54
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
Click for updates
Plant Biosystems - An International Journal Dealing
with all Aspects of Plant Biology: Official Journal of the
Societa Botanica Italiana
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tplb20
Plant communities, diversity and endemism of the Kula
Volcano, Manisa, Turkey
D. I¸Ik-Gürsoya, E. Uğurlua & J. Oldelandb
a Department of Biology, Celal Bayar University, Manisa, Turkey
b Biodiversity, Evolution and Ecology of Plants, Biocenter Klein Flottbek and Botanical
Garden, University of Hamburg, Hamburg, Germany
Accepted author version posted online: 06 Jan 2015.Published online: 28 Jan 2015.
To cite this article: D. I¸Ik-Gürsoy, E. Uğurlu & J. Oldeland (2015): Plant communities, diversity and endemism of the Kula
Volcano, Manisa, Turkey, Plant Biosystems - An International Journal Dealing with all Aspects of Plant Biology: Official Journal
of the Societa Botanica Italiana, DOI: 10.1080/11263504.2014.1001000
To link to this article: http://dx.doi.org/10.1080/11263504.2014.1001000
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

ORIGINAL ARTICLE
Plant communities, diversity and endemism of the Kula Volcano,
Manisa, Turkey
D. IS¸ IK-GU
¨RSOY
1
,E.UG
˘URLU
1
, & J. OLDELAND
2
1
Department of Biology, Celal Bayar University, Manisa, Turkey and
2
Biodiversity, Evolution and Ecology of Plants, Biocenter
Klein Flottbek and Botanical Garden, University of Hamburg, Hamburg, Germany
Abstract
Volcanoes often harbour specialized plant communities and shelter endemic plant species. Kula Volcano is one of 14
volcanoes in Turkey. Although this volcano is clearly a landmark of the Aegean region, only few botanical studies analysed the
vegetation pattern at the Kula Volcano. None performed a phytosociological classification to delimit different plant
communities. We applied a stratified random sampling design according to altitude and aspect and sampled 112 vegetation
plots. We classified plant community types using a modified TWINSPAN analysis followed by the determination of
diagnostic species based on
f
coefficient fidelity values. Floristic relationships between plant community types were
interpreted by ordination and ANOSIM analyses. Further, we used partial correlations of the ordination axes and
environmental parameters in order to identify relationships between vegetation zonation and environment. We identified five
major plant community types based on 85 diagnostic species. These plant community types were significantly correlated with
altitude and aspect. Further, 13 endemic plant species were found from which one was endangered and one was classified as
vulnerable according to International Union for Conservation of Nature (IUCN).
Keywords: Biodiversity, conservation, diagnostic species, endemism, modified TWINSPAN, phytosociology
Introduction
Landscapes rich in species, in particular rare,
threatened or endemic species are often considered
as priority areas for nature conservation. These
hotspots of biodiversity, which require protection,
often include rare habitats (Reid 1998). Volcanoes,
with their special geological characteristics, support
such rare habitats for plant life. Thus, volcanoes are
potentially a key landscape feature that may harbour
many rare and endemic species. We need to under-
stand the importance of these geological areas for the
conservation of biodiversity (Vela
´zquez et al. 2003).
Documenting the distribution of vegetation zones
along altitudinal gradients of volcanoes would be a
first major step (Nakashizuka et al. 1993; Cowie &
Holland 2006). However, little work was done
characterizing the uniqueness of volcano vegetation
in general and in particular in Turkey where 14
volcanoes do exist (Siebert et al. 2010).
The Kula volcanic field is the westernmost
inactive volcanic field in Turkey. The volcanic area
of Kula Volcano extends 15 km in the north – south
direction and 40 km in the west – east direction
(Siebert & Simkin 2002). The basalts in the form
of about 80 small cinder cones and associated lava
flows and fields have a total volume of 2.3 km
3
(Richardson-Bunbury 1996). Recently, the Kula
volcanic field was announced by the UNESCO as the
first geopark in Turkey and the 58th in Europe
(Global Geoparks Network 2013).
The first botanical studies and descriptions of the
vegetation at the Kula Volcano were made by O
¨ner
and Oflas (1977) and Ug˘urlu and Secmen (2009).
About 18 endemic species were listed in their study
highlighting the importance of the Kula Volcano as a
potential conservation target in Turkey. However,
the mentioned studies rather reported occurrences of
species than spatially explicit locations or distri-
butions of associated plant communities that would
be required for future conservation work.
This study aims at documenting the vegetation
distribution at Kula Volcano and discusses its
importance for the conservation of endemic vascular
q2015 Societa
`Botanica Italiana
Correspondence: D. Is¸ık-Gu
¨rsoy, Department of Biology, Graduate School of Natural and Applied Sciences, Celal Bayar University, 45140, Manisa, Turkey.
Tel: þ90 236 2013273. Fax: 90 236 2412158. Email: biodeniz-04@hotmail.com
Plant Biosystems, 2015
http://dx.doi.org/10.1080/11263504.2014.1001000
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015

plants. Therefore, we carried out a phytosociological
sampling in order to determine plant communities
and related biodiversity parameters. In addition, we
analysed major trends in plant diversity and relate the
identified plant communities to the abiotics of the
volcano.
Materials and methods
The study area
The Kula Volcano is located in the western part of
Turkey at 388580N and 288520E(Figure 1). The town
of Kula lies 12 km east of the Kula Volcano. The
elevation of the Kula Volcano ranges from 720 m a.s.
l. to 901 m a.s.l. It is located in the B2 square
according to Davis Grid system (Davis 1965).
Mediterranean climate prevails in the whole
Aegean region where the Kula Volcano is located.
The most typical characteristics of this climate are
dry, hot summers and mild, wet winters. According
to the Turkish State Meteorological Service (2013)
the mean annual temperature is 14.38C, with an
average maximum temperature of roughly 35.18Cin
August, dropping to an average minimum tempera-
ture of 21.68C during January; total annual
precipitation is 584.7 mm with most rainfall in
January. The average annual relative humidity is
54% in the town of Kula. Relative humidity ranges
from 45.9% in winter to 17.1% in summer.
The volcanic relief elements found in the area
consist of around 80 cinder cones, locally named
“divlit”, basaltic lava flows, craters and tuff covers
(Kocman 2004; Tokcaer et al. 2005). The tectonic
movements affecting the area had begun from the
middle Miocene until early Holocene and formed all
the volcanic features in the Kula region (Kocman
2004). The geological structure of the study area
mostly contains limeless brown forest soils that
developed on volcanic rocks, e.g. andesite, volcanic
tuff and basalt (Gu
¨nal 2003). Soil depths vary
between 3 and 40 cm depending on the age of the
basaltic lava flows and the slope and aspect of the
volcano (Atalay et al. 1990).
Land use around the Kula Volcano consists
mainly of farming agriculture taking place directly on
the old lava fields. Particularly wheat, barley,
tobacco, sesame, alfalfa, black lentils and chickpeas
are grown in these areas. The volcano itself is used as
a rangeland for goats and sheep. Goats prefer to
browse on the tips of woody shrubs and trees, as well
as the occasional broad-leaved plants, whereas
12345678910
42°
40°
38°
36°
36°34°32°30°28°26°
C
B
A
44°42°40°38°
(a)
(b)
Figure 1. Southern aspect of the Kula Volcano (A). On the eastern slopes are several planted individuals of Pinus pinea L. The lavas flowed
from the volcano spread to the north and west in direction. The study area is placed on B2 square according to the grid system (B) adopted by
Davis (1965).
2D. Is¸ık-Gu
¨rsoy et al.
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015

sheeps prefer to browse on herbs and grasses. The
eastern slopes of the volcano are used by a Pinus pinea
L. plantation.
The Kula Volcano is phytogeographically located
in the Western Anatolian province of the Mediterra-
nean region (Avcı 1993) consisting mainly of
Mediterranean chorotypes. Characteristic plants of
the Mediterranean phytogeographic region found are
Pinus brutia Ten., Quercus coccifera L., Rhus coriaria
L., Jasminum fruticans L., Umbilicus horizontalis
(Guss.) DC. var. horizontalis, Trigonella cretica (L.)
Boiss., Sideritis lanata L., Anthemis auriculata Boiss.
Of these, particularly Pinus brutia is quite effective
and constitutes a forest in the northern aspect of the
volcano. Quercus coccifera, Jasminum fruticans and
Rhus coriaria are very dense in the form of shrub
understory. In addition, Bromus tectorum L., Poa
bulbosa L., Saxifraga cymbalaria L. var. cymbalaria,
Galium verum L. subsp. verum and Trifolium campestre
Schreb. are commonly found herbaceous plants
(Ug˘urlu & Secmen 2009). The different aspects of
the Volcano seem to hold different plant commu-
nities; however, no study did verify this impression.
Vegetation sampling
The study design for selecting the location of the
vegetation plots was inspired by the vertical profile
drawn by O
¨ner and Oflas (1977). The volcano was
separated into four strata according to each altitude,
i.e. A – 720 –740 m a.s.l., B – 740– 840 m a.s.l., C –
840 –901 m a.s.l. and D – 830– 900 m a.s.l., with D
being placed inside the crater of the volcano. Further
four aspect classes, i.e., N, S, W and E were
delineated and thus formed 16 strata in total. In each
stratum, 7 quadratic vegetation plots (in total 112
releve
´s) were placed randomly with a standardized
size of 10 m
2
. The species composition of each plot
was assessed through releve
´s according to the Braun-
Blanquet method (Westhoff & Van Der Maarel
1978). All vascular plant species present in each plot
were listed, and their abundance was recorded
following visual estimates of the respective percen-
tage cover (cover classes: .0– 1%: r;.1– 2.5%: 1;
.2.5– 5%: 1;.5 – 10%: 2a;.10– 25%: 2b;.25 –
50%: 3;.50– 75%: 4;.75 –100%: 5). Plant
specimens collected were identified and deposited
at the herbarium of the Celal Bayar University
Herbarium. Identifications were done according to
the Flora of Turkey (Davis 1965– 1985). Duplicates
were checked for identification at the Ege University
Herbarium (EGE). The releve
´s were entered to
a vegetation database in TURBOVEG format
(Hennekens and Schamine
´e2001). Vegetation
sampling was conducted in May of 2011. For each
releve
´, we determined geographic coordinates and
altitude with a handheld GPS device, aspect with a
compass and slope with an inclinometer. We also
estimated the area covered by rocks and mosses.
Floristic classification and statistical analysis
We applied the modified version of Two-Way
Indicator Species Analysis (TWINSPAN, Hill
1979; Rolec
´ek et al. 2009) to our raw vegetation
table. This algorithm is implemented in the free
software JUICE 7.0 (Tichy
´2002). We used the
default settings of JUICE (pseudo species cut levels:
0%, 2%, 5% and 10%; minimum group size: 5).
In order to identify an optimal clustering depth
within the dendrogram presented by the modified
TWINSPAN algorithm, we applied the OptimClass
procedure (Tichy
´et al. 2010), also implemented in
JUICE. OptimClass was applied in mode 1, where
the total number of significantly faithful species per
cluster depth is plotted against the hierarchical level.
Faithful species are determined based on the p-value
of the Fisher’s exact test with a threshold set to p-
value ¼10
26
(Tichy
´et al. 2010). The clustering
depth with the maximum number of faithful species
was selected, which led to an optimal cluster depth of
five clusters.
To identify the diagnostic species of each plant
community type, we calculated the
f
coefficient of
association. The species were considered as highly
diagnostic when
f
.0.50 and as diagnostic when
f
.0.25, provided that the concentration was
significant according to Fisher’s exact test at
a
¼0.05 (Chytry
´et al. 2002; Chytry
´2007). Because
we calculated and identified five cluster according to
Tichy
´and Chytry
´(2006). The compiled list of
diagnostic species was first sorted according to
decreasing
f
values within the clusters derived from
TWINSPAN, and then columns were manually
rearranged in order to optimize
f
values.
We applied a Principal Coordinates Analysis
(PCoA) to the species by plot matrix using a Bray-
Curtis distance measure in order to visualize the
differences in species composition between the
groups and to identify related environmental
gradients. Further, we conducted a multivariate
Analysis of Similarity (ANOSIM, Clarke 1993)in
order to test for separation between the five clusters
in ordination space. Further, to see whether there is
any corresponding zonation along gradients, we
extracted the first three ordination axes and
correlated these with altitude and slope (in degree)
and common measures of diversity, i.e. species
richness, Shannon diversity and Evenness. The
PCoA, ANOSIM and linear Pearson correlations
were computed using the free software PAST
(Hammer et al. 2001). Scatter plots were made
with the package “ggplot2” (Wickham 2009) in the
free statistical software R (R Core Team 2013).
Plant communities, diversity and endemism 3
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015

Results
Floristic analysis
Total 170 vascular plant species belonging to 123
genera and 42 families were recorded. The families
with highest species numbers were Asteraceae (23),
Caryophyllaceae (19), Lamiaceae (16), Fabaceae
(14) and Poaceae (12). We found 13 endemic taxa
throughout all five vegetation units (Ta b l e I ).
According to the International Union for Conserva-
tion of Nature (IUCN) criteria for red-list species,
three species are classified as endangered and one is
classified as vulnerable. The remaining nine species
are low-risk species with different status (Ekim et al.
2000).
Plant community types and diagnostic species
According to the TWINSPAN classification and
following OptimClass analysis, we obtained five plant
community types. Considering a threshold value of
0.25 for the
f
coefficient, 85 diagnostic species were
selected for the different vegetation units (Table II).
In naming syntaxa, we followed the informal
approach of Alvarez et al. (2012), i.e. the naming of
the syntaxa always includes the name of the most
diagnostic species and the species with the highest
cover. The vegetation units are described as follows:
Group 1: Moenchia mantica –Pinus brutia
community type: This group occurred in 29 plots
and comprised 38 diagnostic species. The average
number of species per plot is 43.48, making this the
most diverse vegetation unit. Moenchia mantica (L.)
Bartl. subsp. mantica and Minuartia hybrida (Vill.)
Schischk. subsp. hybrida (Vill.) Schischk are annual
herbaceous plants with notable high diagnostic
species. Bromus tectorum and Pinus brutia were
among the most dominant species in this vegetation
unit. Six endemic species were found in this
vegetation unit (Table I).
Group 2: Vicia cracca–Pinus brutia com-
munity type: This is the largest cluster encountered
in 33 plots and contained 15 diagnostic species with
Vicia cracca L. subsp. stenophylla Vell., a perennial
herbaceous plant, being the single highest diagnostic
species. The average number of species per plot is
36.12. Dominant species in this group were among
others Pinus brutia,Poa bulbosa and Saxifraga
cymbalaria var. cymbalaria. Other notable but less
dominant tree species are Quercus coccifera and
Q. infectoria Olivier. Seven endemic species were
found in this vegetation unit (Table I).
Group 3: Gypsophila tubulosa–Pistacia
terebinthus community type: This group
occurred in 15 plots with 20 diagnostic species.
The annual Gypsophila tubulosa (Jaub. et Spach)
Boiss. and the perennial Rumex acetosella L. are
highly diagnostic character species for this
group. The average species number per plot is
32.20. Poa bulbosa was the most dominant species,
whereas Pinus brutia and Quercus had the lowest
cover in this group. Pistacia terebinthus L.
subsp. palaestina (Boiss.) Engler was the most
constant species yet with low cover. This vegetation
unit had nine endemic species, which is the highest
number found. Four diagnostic species in this group
are endemics with Gypsophila tubulosa being also the
name of species for this vegetation unit (Table I).
Group 4: Pinus brutia –Milium vernale
community type: This group occurred in 28 plots
and contained 10 diagnostic species of which Pinus
brutia was the only highly diagnostic species. The
Table I. List of endemic plant species found at Kula Volcano.
Endemic taxa Risk categories 12345
Anthemis dipsacea Bornm. EN þþþ
Trigonella cretica (L.) Boiss. LR(NT) þ
Phlomis armeniaca Willd. LR(LC) þþ
Vincetoxicum canascens subsp. (Willd.) Decne subsp.
pedunculata Browicz. VU þþ
Onosma isauricum Boiss. et Heldr. LR(LC) þþþ
Centaurea urvillei DC. subsp. stepposa Wagenitz LR(LC) þþþ
Dianthus anatolicus Boiss. LR(LC) þþ
Gypsophila tubulosa (Jaub. & Spach.) Boiss. LR(LC) þ
Stachys cretica subsp. anatolica LR(LC) þþ
Linaria corifolia Desf. LR(LC) þþ þ
Ziziphora taurica Bieb. subsp. clenioides (Boiss.) Davis LR(CD) þþþ
Linaria genistifolia (L.) Miller subsp. confertiflora
(Boiss.) Davis LR(LC) þþþ
Scrophularia floribunda Boiss. et Bal. LR(NT) þþ
Notes: Threat categories according to Ekim et al. (2000). Species are sorted according to their IUCN risk category. Numbers in table header
refer to vegetation class numbers. A “ þ” indicates occurrence, a grey shade signals diagnostic species for the particular category.
Abbreviations: EN ¼Endangered, VU ¼Vulnerable, LR(LC) ¼Low Risk/Least Concern, LR(CD) ¼Low Risk/Conservation Depen-
dent, LR(NC) ¼Low Risk/Nationally Critical, LR(NT) ¼Low risk/Near Threatened.
4D. Is¸ık-Gu
¨rsoy et al.
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015

Table II. Synoptic table of the vegetation sampling showing the constancy (left) and the fidelity (right) values of the species in the five
groups.
(a) percentage frequency {constancy} (b) fidelity {
f
coefficient £100 }
Group No. 1 2 3 4 5 1 2 3 4 5
No. of releve
´s 29 33 15 28 7 29 33 15 28 7
No. of diagnostic species 38 15 20 10 2 38 15 20 10 2
Diagnostic species of single groups
Group 1: Moenchia mantica–Pinus brutia commutiy type
Moenchia mantica subsp. mantica 86 – – – – 75.8 – – – –
Minuartia hybrida subsp. hybrida 97 – – – – 69.2 – – – –
Logfia arvensis 86 – – – – 62.2 – – – –
Briza maxima 72 30 – – – 53.7 – – – –
Trifolium repens 72 45 – – – 53.4 – – – –
Brachypodium pinnatum 90 – – 57 – 50 – – – –
Verbascum glomeratum 24 – – – – 45 – – – –
Nepeta nuda subsp. albiflora 24 – – – – 45 – – – –
Chrysopogon gryllus 24 – – – – 45 – – – –
Aethionema arabicum 31 – – – – 42.3 – – – –
Erodium cicutorium supsp. cirutorium 31 – – – – 42.3 – – – –
Alkanna tinctoria subsp. tinctoria 21 – – – – 41.6 – – – –
Lupinus micranthus 21 – – – – 41.6 – – – –
Salvia verbenaca 24 – – – – 40.6 – – – –
*Linaria corifolia 41 – – – – 39 – – – –
Legousia speculumveneris 83 61 – 64 – 38.8 – – – –
Anthemis tinctoria var. tinctoria 31 – – – – 38.6 – – – –
Campanula lyrata 24 – – – – 38 – – – –
Mentha pulegium 17 – – – – 37.8 – – – –
Anchusa azurea var. azurea 17 – – – – 37.8 – – – –
Salvia fruticosa 17 – – – – 37.8 – – – –
Myosotis stricta 83 70 – 50 – 37.1 – – – –
Cerastium glomeratum 28 – – – – 36.3 – – – –
Scandix maxrohhyncha 34 – – – – 35.3 – – – –
Taeniatherum caputmedusae subsp. asper 38 – – – – 34 – – – –
Parentucellia latifolia subsp. latifolia 38 – – – – 34 – – – –
Ornithogalum nutans 14 – – – – 33.7 – – – –
Anchusa arvensis 14 – – – – 33.7 – – – –
Picnomon acarna 45 – – – – 29.6 – – – –
Silene odontopetala 10 – – – – 29.1 – – – –
Silene dichotoma subsp. sibthorpiana 10 – – – – 29.1 – – – –
Cnicus benedictus var. kotschyi 10 – – – – 29.1 – – – –
Origanum spyleum 10 – – – – 29.1 – – – –
Jasione heldreichii 10 – – – – 29.1 – – – –
Scrophularia lucida 14 – – – – 28.9 – – – –
Seseli peucedanoides 14 – – – – 28.9 – – – –
Lolium multiflorum 31 – – – – 28.5 – – – –
Tragopogon longirostris var. abbreviatus 14 – – – – 28.2 – – – –
Poa bulbosa 100 97 100 86 – 25.5 – – – –
Group 2: Vicia cracca–Pinus brutia community type
Vicia cracca subsp. stenophylla – 58 – – – – 57.7 – – –
Geranium lucidum – 85 60 50 – – 45.4 – – –
Jurinea consanguinea – 18 – – – – 34.1 – – –
Crupina crupinastrum – 12 – – – – 31.5 – – –
Vicia hybrida – 12 – – – – 31.5 – – –
Silene chlorifolia – 73 53 54 – – 30.2 – – –
Myosotis cadmaea – 15 – – – – 30.2 – – –
Centaurea solstitialis subsp. solstitialis – 15 – – – – 30.2 – – –
Anthemis auriculata – 18 – – – – 30.2 – – –
Muscari comosum – 27 – – – – 29.6 – – –
Quercus coccifera 55 70 47 32 – – 29.5 – – –
Bromus rigidus 45 48 – – – – 29.2 – – –
Silene chlorifolia 41 73 53 54 – – 28.7 – – –
Quercus infertoria 48 73 – 43 43 – 26.1 – – –
Pistacia terebinthus subsp. palaestina 79 100 93 93 – – 25.9 – – –
Plant communities, diversity and endemism 5
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015

average number of species per plot was 30.36. Pinus
brutia,Quercus coccifera,Galium ver um and Bromus
tectorum were also the most dominant species. Eight
endemic species are found in this vegetation unit
(Table I).
Group 5: Pisum sativum –Origanum onites
community type: Group 5 was encountered in only
seven plots. The two highly diagnostic species were
Pisum sativum L. subsp. biflorum (Raf.) Soldans. and
Origanum onites L. The vegetation cover in this group
was generally low, and all plots were completely
without trees. Further, with 10.43, this group had
the lowest average number of species per plot.
No endemic species are found in this vegetation unit.
A PCoA was calculated on the species by plot
table. In ordination space, distant points indicate
dissimilar species composition and abundance
(Figure 2). The first two axes explained 28% of the
TABLE II – continued
(a) percentage frequency {constancy} (b) fidelity {
f
coefficient £100 }
Group No. 1 2 3 4 5 1 2 3 4 5
No. of releve
´s 29 33 15 28 7 29 33 15 28 7
No. of diagnostic species 38 15 20 10 2 38 15 20 10 2
Group 3: Gypsophila tubulosa–Pistacia terebinthus community type
*Gypsophila tubulosa – – 93 – – – – 73.1 – –
Rumex acetosella 52 – 93 – – 22.6 – 67.7 – –
Isatis tinctoria subsp. tomentella 45 45 93 – – – – 57.7 – –
Opopanax hispidus – – 40 – – – – 56 – –
Sideritis lanata – – 47 – – – – 52.1 – –
*Phlomis armeniaca – 30 73 – – – – 52.1 – –
Allium paniculatum subsp. paniculatum – – 47 – – – – 51 – –
*Onosma isauricum 45 33 73 – – – – 46.8 – –
Umbilicus horizontalis var. horizontalis – – 87 57 29 – – 42 12.4 –
Ephedra major – – 47 – – – – 39.6 – –
Orobanche cernua – – 33 – – – – 39.1 – –
Eryngium campestre var. campestre – – 33 – – – – 38.5 – –
Lonicera etrusca var. etrusca – – 20 – – – – 32.7 – –
Sedum amplexicaule 48 39 80 61 – – – 31.5 – –
*Ziziphora taurica subsp. clenioides 0 47 – – – – 31.1 – –
Papaver argemone 0 20 – – – – 29.2 – –
Rhus coriaria 72 70 93 64 29 – – 29.1 – –
Muscari neglectum – – 33 – – – – 28.8 – –
Sisymbrium orientale – – 13 – – – – 27.7 – –
*Scrophularia floribunda – – 20 – – – – 26.4 – –
Group 4: Pinus brutia – Milium vernale community type
Pinus brutia – 45 33 96 – – – – 57.8 –
Milium vernala subsp. vernala 34 55 – 79 – – – – 43.9 –
Senecio vernalis 48 61 – 79 – – – – 42.4 –
*Dianthus anatolicus – – – 43 – – – – 35.8 –
Lamium amplexicaule – – – 29 – – – – 32.7 –
Galium aparine 34 33 – 64 29 – – – 30.9 –
*Trigonella cretica – – – – – – – 29.6 –
Dianthus capitatus – – – – – – – 29.6 –
Alcea pallida – – 40 50 – – – – 29 –
Ceterach officinarum 48 36 53 68 – – – – 27.1 –
Group 5: Pisum sativum–Origanum onites community type
Pisum sativum subsp. biflorum – – – – 43 – – – – 57.8
Origanum onites ––––43––––52
Common diagnostic species two or more groups
Saponaria mesogitana 97 – 93 – – 54.2 – 50.9 – –
Tordylum apulum 83 – 67 – – 52.9 – 35.8 – –
Bromus tectorum 100 79 – 89 43 39 – – 27.9 –
Jasminum fruticans 45 52 – 0 – 32.4 40.9 – – –
Trifolium ar vense var. arvense 66 – 93 – 25.7 – 54.1 – –
Valeriana dioscoridis – 82 – 64 – – 49.6 – 31.1 –
Saxifraga cymbalaria var. cymbalaria 55 73 – 71 – – 33.6 – 32.2 –
Isatis tinctoria subsp. tinctoria 55 70 73 – – – 30.7 34.4 – –
Ferula communis subsp. communis 45 73 93 – – – 25.6 46.2 – –
Notes: Species with no significant fidelity values are not shown. Species in dark grey are highly diagnostic for the respective group; light grey
species are only low diagnostic species (according to a Fisher’s exact test at
a
¼0.05). Diagnostic species (values grey-shaded) were selected
applying a threshold of
f
$0.25. Endemic species are marked with a * in front of the species name.
6D. Is¸ık-Gu
¨rsoy et al.
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015

total variance in species composition. In ordination
space of the first two axes, cluster 5 is separated from
all other clusters in the lower part of the plot. Cluster
4 is a well-defined cluster at the left-hand side of the
plot, whereas cluster 3 is tightly packed in the right-
hand side. Cluster 2 has the biggest variation as
indicated by the 95% ellipses. Cluster 1 is also well
clustered but was situated in the centre of cluster 2.
In order to interpret overlap between vegetation
clusters not only in 2D but in the complete
multivariate space, we performed a multivariate
group comparison with an ANOSIM (Clarke 1993)
analysis using the Bray-Curtis distance (Table III).
The interpretation of Rvalues follows Clarke (1993)
and can be read as follows: 0 – 0.25 means no
separation between clusters; 0.25 – 0.50 means
little separation between clusters; 0.50 – 0.75 good
separation but still overlap and 0.75 – 1.00 shows
clear separation between clusters. The global
ANOSIM-R value was 0.68 ( p,0.001). All clusters
except 1 and 2 and 2 and 3 show Rvalues above 0.5.
Thus, all cluster combinations except cluster 1 –2
and cluster 2– 3 can be characterized as having only
little overlap indicating that the clusters are different
in their species composition.
Environmental gradients
We correlated the first three ordination axes with
environmental and diversity measures. The resulting
correlation table (Table IV) shows that the first two
axes were strongly correlated with diversity
measures. The third PCoA axis was correlated with
the environmental parameters altitude and slope.
Figure 3 revealed a potential altitudinal zonation
beginning from lowest altitude with cluster 1, then
cluster 5 and 2, then 4 and the highest cluster is
cluster 3. However, cluster 3 also had a subcluster at
lower altitudes. The diversity measures themselves
were strongly correlated. Finally, species richness is
Figure 2. Principal Coordinate Analysis (PCoA) of species data using Bray-Curtis distance. Objects are vegetation plots, and distances
quantify similarity in terms of species composition and abundance. Plant community types are displayed in different symbols. Lines show
95% ellipses around the centroid of a respective vegetation class and help to identify centres in ordination space.
Table III. ANOSIM pairwise comparisons between groups.
Cluster 1 2 3 4 5
1 0.00
2 0.36 0.00
3 0.58 0.43 0.00
4 0.75 0.53 0.95 0.00
5 0.99 0.87 0.95 0.98 0.00
Notes: Values are ANOSIM-R values. Global ANOSIM-R was
0.68, p,0.001.
Table IV. Correlation table showing Pearson’s rvalue for PCoA
Axes, environmental parameters (except aspect) and diversity
values.
PCoA 1 PCoA 2 PCoA 3
Altitude 20.36 20.38 0.53
Slope (8)20.31 0.08 0.56
Richness 0.25 0.77 20.11
Shannon 0.51 0.71 20.09
Evenness 0.79 0.02 20.20
Note: Significant Pearson’s correlations i.e. p,0.5 are marked in
light grey.
Plant communities, diversity and endemism 7
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015

negatively correlated with altitude, i.e. we find less
species at the top of the cone.
Regarding the distribution of the vegetation units
within the four major aspect classes, each unit had a
more or less clear centre in aspect. Ta b l e V summarizes
the pattern found. Cluster 1 is scattered over all
aspects, whereas 2 and 3 were clearly related to east and
south aspects. Cluster 4 and 5 were both related to
western and northern aspects. We performed a
x
2
test
of association to examine the relation between aspect
class and cluster type. The null hypothesis was that
there is no relation between aspect class and cluster
type with regard to the number of samples. The
relation between these variables was significant,
x
2
(df ¼12, n¼112) ¼99.133, p,0.001 meaning that
aspect classes were associated with the cluster types.
As a measure of strength of the association, we
calculated Cramers’V, which was 0.55, p,0.001
indicating a very strong association. It should be noted
that the
x
2
test of association is not precise when values
below 5 appear in the contingency Ta b l e V.Fisher’s
exact test would be more appropriate. However,
Fisher’s exact test is very demanding, and it was not
possible to calculate the test for the present table.
As the
x
2
value is very high and highly significant, we
regard this result as plausible. The test was carried out
using PAST version 2.17 (Hammer et al. 2001).
Discussion
Floristic classifications of vegetation are often the
building blocks for many conservation plans and natural
resource management maps (Eminag˘aog˘lu et al. 2006;
Brown et al. 2013). This study compiled 112 vegetation
plots at the Kula Volcano and applied a modified
TWINSPAN classification, and five plant community
types were identified. These matched the observed
zonation according to altitude and aspect. At the feet of
the volcano at 750– 800 m a.s.l., a very species-rich
plant community type was identified that appeared at all
four aspects. The second plant community type had
very high amplitude across the altitudinal gradient but
was mainly bound to the eastern aspects. The third
vegetation unit was found mainly on southern slopes
but appeared at the highest and at intermediate
altitudes. Cluster 4 and 5 were both related to North
West-slopes and were situated in higher altitudes;
however, cluster 5 was sampled in always tree-free areas
resulting in plots with very low species numbers.
The zonation described in this study differs partly
from the first complete vegetation study from O
¨ner
and Oflas (1977). The authors did not apply
phytosociological classification or any statistical
Figure 3. Scatter plot showing positive relationship between altitude and PCoA axis 3. Symbols depict the five different plant community
types.
Table V. Number of plots from plant community types within
four main aspect classes.
Cluster West East South North
16788
222065
301140
4170011
53004
Notes: Grey shade indicates main aspect.
x
2
test shows association
between cluster and aspect classes (df ¼12, n¼112,
x
2
¼99.133, p,0.001) with Cramer’s V (0.55, p¼,0.001)
indicating a strong association.
8D. Is¸ık-Gu
¨rsoy et al.
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015

analysis but carried out a “successional survey”. For
this, the identified six different successional stages
including three stages based only on cryptogams,
then a fourth herbaceous stage and finally a shrubby
and a climax forest state. The latter was dominated
by Quercus infectoria and Pinus brutia. The authors
point out that P. brutia, although being dominant on
the northern slopes, it is not the natural climax
species but Q. infectoria. Our findings agree in
general with the findings of O
¨ner and Oflas (1977).
However, their initial zonation of A, B, C and D
could not be verified by our analysis. This is probably
mainly due to the different methodology in sampling
and classification. While O
¨ner and Oflas only listed
species in four zones and then independently again in
four aspect classes, we were able to identify
vegetation units that clearly span several altitudinal
zonations being restricted to specific aspects of the
Kula Volcano, e.g. cluster 2 and 3. Other studies
applying a plot-based sampling also found inter-
actions between aspect and altitude for example
Vel a
´zquez (1994) at the volcanoes Tla
´loc and Pelado
in Mexcio or Islebe et al. (1995) who analysed high
altitude coniferous forests on volcanoes in Guate-
mala, the latter also applying a phytosociological
approach. Thus, our new classification improves the
previous work of O
¨ner and Oflas (1977) in describing
the vegetation zonation at the Kula Volcano.
Our multivariate analysis further revealed trends
in plant diversity across the environmental gradients.
The first vegetation unit, which was the one at the
lower altitudes surrounding the whole volcano, was
the one with the highest average species richness
(43.48). This rather large group appeared in all
different aspects; however, this cannot explain the
high diversity. The correlation analysis revealed that
plant species richness negatively correlated (Pear-
son’s r¼20.51) with altitude. This agrees with one
of the two most frequently observed pattern of
species richness along altitudinal gradients (Rahbek
1995), i.e. a decrease in richness with increasing
altitude. However, this pattern is often found for
other taxa than plants, i.e. birds or reptiles, whereas
plants often show rather a hump-shaped pattern
(McCain & Grytnes 2010). Potential explanations
for this decrease in richness could be strong winds at
the top of the volcano, source sink dynamics from the
adjacent area at the bottom of the volcano or specific
soil conditions. For example, soil depth at the
northern slopes is deeper than that on southern
slopes, allowing the growth of tall trees like Pinus
brutia. However, water holding capacity of the soils is
generally weak in combination with Mediterranean
climate restricting tree growth on the southern and
western slopes. We advocate further research in order
to disentangle this mechanisms and their role on
forming the plant community types.
Spatially explicit information about endemism is
an important issue for conservation managers.
Ug˘urlu and Secmen (2009) found 18 endemic
species in total, which were mainly located at the
tree-free lava fields surrounding the volcano. With a
much higher number of vegetation samples, we
found 13 endemic species on the slopes of the
volcano excluding the lava fields. The highest
number of endemic species was found in cluster 3
which was found at mid- to high altitudes and south-
facing slopes. All other vegetation units except class 5
also contained endemic species. For such a small and
young volcano with its limited area, 13 or even 18
endemic species can be considered as a rather high
number. Other lava fields worldwide, such as those of
the Eastern Afromontane region, reach much higher
numbers, as reported by Hobohm (2014).
In particular, oceanic volcanic island such as the
Canary Islands are well known for their high rates of
endemism (Hobohm 2000). Irl and Beierkuhnlein
(2011) compiled a unique vegetation plot database
for the Canary island La Palma. They report that
different earlier studies also found different numbers
of endemics, ranging from 41 to 61, depending on
available data and previous knowledge. Heads
(2010) highlighted the importance of volcanic
islands as stepping stones for metapopulations of
palms in the pacific. A similar perspective can be
applied to the Kula Volcano, as three of the occurring
endemic plants are actually listed as endangered.
Still, the vegetation at Kula Volcano is harbouring
these endemic species. Yet, human pressure on
natural resources is ever increasing, and great care
should be taken for the endemic species at this
important and rare habitat. Based on this knowledge,
appropriate actions for conservation management
should be initiated. With regard to the other 12
volcanoes in Turkey, endemism and phytosociology
of volcano vegetation can be valuable field of research
in the future.
Acknowledgements
This work was supported by the [CBU-BAP] under
[grant number 2010-084]; [ERASMUS Exchange
program] under [grant number 40198292-5018].
The authors do not have any personal or institutional
conflicts of interest.
References
Alvarez M, Becker M, Bo¨hme B, Handa C, Josko M, Kamiri HW,
et al. 2012. Floristic classification of the vegetation in small
wetlands of Kenya and Tanzania. Biodiver Ecol 4: 63 – 76,
doi:10.7809/b-e.00060.
Atalay I
˙, Sezer LI
˙, Temucin E, Is¸ık S¸ , Mutluer M. 1990. Ege
Bo¨lu
¨mu
¨’nde toprak olus¸umunu etkileyen fakto¨rler. EU
Cog˘ rafya Dergisi 2: 30 – 47.
Plant communities, diversity and endemism 9
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015

Avcı M. 1993. Tu
¨rkiye’nin Flora Bo¨ lgeleri ve Anadolu Diagonaline
Cog˘ rafi Bir Yaklas¸ım. Tu
¨rk Cog˘ rafya Dergisi 28: 225 – 248.
Brown LR, Du Preez PJ, Bezuidenhout H, Bredenkamp GJ,
Mostert THC, Collins NB. 2013. Guidelines for phytosocio-
logical classifications and descriptions of vegetation in southern
Africa. Koedoe 55(1): 1 –10, doi:10.4102/koedoe.v55i1.1103.
Chytry
´M. 2007. Project Vegetation of the Czech Republic:
Preface and summary of methods. In: Chytry
´M, editor.
Vegetation of the Czech Republic – 1. Grassland and
heathland vegetation [in Czech, with English summary].
Praha: Academia. pp. 35– 52.
Chytry
´M, Tichy
´L, Holt J, Botta-Duka
´t Z. 2002. Determination
of diagnostic species with statistical fidelity measures. J Veg Sci
13: 79–90, doi:10.1111/j.1654-1103.2002.tb02025.x.
Clarke KR. 1993. Non-parametric multivariate analyses of
changes in community structure. Aust Ecol 18(1): 117 – 143,
doi:10.1111/j.1442-9993.1993.tb00438.x.
Cowie RH, Holland BS. 2006. Dispersal is fundamental to
biogeography and the evolution of biodiversity on oceanic
islands. J Biogeog 33(2): 193–198, doi:10.1111/j.1365-2699.
2005.01383.x.
Davis PH. 1965–1985. Flora of Turkey and the East Aegean
lslands. 1– 9. Edinburgh: Edinburgh University Press.
Ekim T, Koyuncu M, Vural M, Duman H, AytacZ, Adıgu
¨zel N.
2000. Tu
¨rkiye Bitkileri Kırmızı Kitabı (Eg˘relti ve Tohumlu
Bitkiler). Ankara: Tu
¨rkiye Tabiatını Koruma Derneg˘i ve
Yu
¨zu
¨ncu
¨yıl U
¨niversitesi. pp. 45– 196.
Eminag˘aog˘lu O
¨, Kutbay HG, Bilgin A, Yalcın E. 2006.
Contribution to the phytosociology and conservation of
tertiary relict species in nor theastern Anatolia (Turkey). Belg
J Bot 139: 124–130.
Global Geoparks Network. 2013. Available: http://www.
globalgeopark.org/. Accessed Sep 2013 09.
Gu
¨nal N. 2003. Yukarı Gediz Havzası’nın Bitki Cog˘ rafyası.
I
˙stanbul: C¸ antay YayınIarı. pp. 34 –35.
Hammer Ø, Harper DAT, Ryan PD. 2001. PAST: Paleontological
statistics software package for education and data analysis.
Palaeontol Electron 4: 1– 9.
Heads M. 2010. The endemic plant families and the palms of New
Caledonia: a biogeographical analysis. J Biogeogr 37(7):
1239– 1250, doi:10.1111/j.1365-2699.2010.02292.x.
Hennekens SM, Schamine
´e JHJ. 2001. TURBOVEG, a compre-
hensive data base management system for vegetation data.
J Veg Sci 12(4): 589 – 591, doi:10.2307/3237010.
Hill MO. 1979. TWINSPAN: A Fortran program for arranging
multivariate data in an ordered two-way table by classification
of the individuals and attributes. Ithaca, NY, USA: Section of
Ecology and Systematics, Cornell University Press.
Hobohm C. 2000. Plant species diversity and endemism on islands
and archipelagos, with special reference to the Macaronesian
Islands. Flora (Jena) 195: 9– 24.
Hobohm C, editor. 2014. Endemism in vascular plants. Plant and
vegetation. Vol. 9. Dordrecht, NL: Springer.
Irl SDH, Beierkuhnlein C, 2011. Distribution of endemic plant
species on an oceanic island-a geospatial analysis of La Palma
(Canary Islands). Proc Environ Sci 7: 170 – 175, doi:10.1016/j.
proenv.2011.07.030.
Islebe GA, Vela
´zquez A, Cleef AM. 1995. High elevation
coniferous vegetation of Guatemala. Vegetatio 116: 7 – 23.
Kocman A. 2004. Yanık U
¨lke’nin Dog˘ al Anıtlar: Kula Yo¨ resi
Volkanik Olus¸umları. Ege Cog˘rafya Dergisi 13: 5– 15.
McCain CM, Grytnes J-A. 2010. Elevational gradients in species
richness. Encyclopedia of life sciences.
Nakashizuka T, Iida S, Suzuki W, Tanimoto T. 1993. Seed
dispersal and vegetation development on a Debris Avalanche
on The Ontake Volcano, Central Japan. J Veg Sci 4(4):
537– 542, doi:10.2307/3236081.
O
¨ner M, Oflas S. 1977. Plant Succession on the Kula Volcano in
Turkey. Vegetatio 34: 55 – 62.
R Development Core Team. 2013. R: A language and environment
for statistical computing. Vienna, Austria: R Foundation for
Statistical Computing.
Rahbek C. 1995. The elevational gradient of species richness:
a uniform pattern? Ecography 18(2): 200 – 205, doi:10.1111/j.
1600-0587.1995.tb00341.x.
Reid WV. 1998. Biodiversity Hotspots. Trends Ecol Evol 13(7):
275– 280, doi:10.1016/S0169-5347(98)01363-9.
Richardson-Bunbury JM. 1996. The Kula volcanic field, western
Turkey: the development of a Holocene alkali basalt province
and the adjacent normal-faulting graben. Geol Mag 133(3):
275– 283, doi:10.1017/S0016756800009018.
Rolec
´ek J, Tichy
ˆL, Zeleny
ˆD, Chytry
ˆM. 2009. Modified
TWINSPAN classification in which the hierarchy respects
cluster heterogeneity. J Veg Sci 20(4): 596 –602, doi:10.1111/j.
1654-1103.2009.01062.x.
Siebert L, Simkin T. 2002. Volcanoes of the World: an Illustrated
Catalog of Holocene Volcanoes and their Eruptions, Smithso-
nian Institution, Global Volcanism Program Digital Infor-
mation Series, GVP-3. Available: http://www.volcano.si.edu/
world/. Accessed Jun 2012 20.
Siebert L, Simkin T, Kimberly P. 2010. Volcanoes of the World.
3rd ed. Berkeley, CA and London: University of California
Press.
Tichy
´L, Chytry
´M. 2006. Statistical Determination of Diagnostic
Species for Site Groups of Unequal Size. J Veg Sci 17:
809– 818.
Tichy
´L, Chytry
´M, Ha
´jek M, Talbot SS, Botta-Duka
´t Z. 2010.
OptimClass: using species-to-cluster fidelity to determine
the optimal partition in classification of ecological commu-
nities. J Veg Sci 21: 287 – 299, doi:10.1111/j.1654-1103.2009.
01143.x.
Tichy
´L. 2002. JUICE, software for vegetation classification. J Veg
Sci 13: 451– 453, doi:10.1111/j.1654-1103.2002.tb02069.x.
Tokcaer M, Agostini S, Savas¸cın MY. 2005. Geotectonic Setting
and Origin of the Youngest Kula Volcanics (Western Anatolia),
with a New Emplacement Model. Turk J Earth Sci 14:
145– 166.
Turkish State Meteorological Service. 2013. Available: http://
www.mgm.gov.tr/. Accessed Jan 2013 15.
Ug˘ urlu E, Secmen O
¨. 2009. Kula Volcano (Turkey). Bull Eur Dry
Grassl Group 3: 23–25.
Vela
´zquez A. 1994. Multivariate analysis of the vegetation of the
volcanoes Tla
´loc and Pelado. Mexico. J Veg Sci 5: 263–270.
Vela
´zquez A, Bocco G, Romero FJ, Vega AP. 2003. A Landscape
Perspective on Biodiversity Conservation. Mt Res Dev 23:
240– 246.
Westhoff V, Van Der Maarel E. 1978. The Braun-Blanquet
approach. In: Whittaker RH, editor. Classification of plant
communities, Vol 5(1). Netherlands: Springer. pp. 287– 399.
Wickham H. 2009. ggplot2: Elegant Graphics for Data Analysis.
New York: Springer.
10 D. Is¸ık-Gu
¨rsoy et al.
Downloaded by [Deniz IIK GÜRSOY] at 10:54 02 February 2015