ThesisPDF Available

Determining potential niche competition regions between Kazdağı fir & (Abies nordmanniana subsp. equi-trojani) & Anatolian black pine (Pinus nigra subsp. pallasiana) and conservation priority areas under climate change by using MaxENT algorithm

Authors:

Abstract and Figures

Kazdagi Fir (Abies nordmanniana subsp. equi-trojani) is an endemic coniferus tree subspecies in Turkey. The species has a narrow distribution and its conservation status defined as EN (Endangered) by IUCN (International Union for Conservation of Nature), due to its decreasing population size. Besides the human activities unfavoring the species viability, it also suffers from a severe competition which has overlapping distribution with Kazdagi fir; Anatolian Black Pine (Pinus nigra subsp. pallasiana). The study presented here, aimed to detect conservation priority areas for Kazdagi fir. In order to achieve this, species distribution modeling approach by using MAXENT algorithm was used to model both Kazdagi fir and Anatolian Black Pine’s potential distributions in 2050 by considering the possible effects of global climate change in near future. Then, a series of overlay analyses were made to be able to detect the areas which are better habitats for Kazdagi fir than Anatolian black pine. The results of the study revealed several regions. Yet, considering the current distribution of the species and its dispersal limits, this study proposes two conservation priority areas for Kazdagi fir; Uludag and Kazdagi (Mt. Ida). The assessed regions are the most important habitats for both species according to both currently and in 2050 climate scenarios. Thus, it is crucial that forestry and conservation practices should be taken into consideration in these areas.
Content may be subject to copyright.
DETERMINING POTENTIAL NICHE COMPETITION REGIONS BETWEEN
KAZDAGI FIR (Abies nordmanniana subsp. equi-trojani) & ANATOLIAN BLACK
PINE (Pinus nigra subsp. pallasiana) AND CONSERVATION PRIORITY AREAS
UNDER CLIMATE CHANGE SCENARIOS BY USING MAXENT
ALGORITHM
A THESIS SUBMITTED TO
THE GRADUATE SCHOOL OF NATURAL AND APPLIED SCIENCES
OF
MIDDLE EAST TECHNICAL UNIVERSITY
BY
NURBAHAR USTA BAYKAL
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR
THE DEGREE OF MASTER OF SCIENCE
IN
BIOLOGY
FEBRUARY 2019
Approval of the thesis:
DETERMINING POTENTIAL NICHE COMPETITION REGIONS
BETWEEN KAZDAGI FIR (ABIES NORDMANNIANA SUBSP. EQUI-
TROJANI) & ANATOLIAN BLACK PINE (PINUS NIGRA SUBSP.
PALLASIANA) AND CONSERVATION PRIORITY AREAS UNDER
CLIMATE CHANGE SCENARIOS BY USING MAXENT ALGORITHM
submitted by NURBAHAR USTA BAYKAL in partial fulfillment of the
requirements for the degree of Master of Science in Biology Department, Middle
East Technical University by,
Prof. Dr. Halil Kalıpçılar
Dean, Graduate School of Natural and Applied Sciences
Prof. Dr. Orhan Adalı
Head of Department, Biology
Prof. Dr. Zeki Kaya
Supervisor, Biology, METU
Examining Committee Members:
Prof. Dr. Musa Doğan
Biology, METU
Prof. Dr. Zeki Kaya
Biology, METU
Prof. Dr. C. Can Bilgin
Biology, METU
Assoc. Prof. Dr. Sertaç Önde
Biology, METU
Assoc. Prof. Dr. Çağatay Tavşanoğlu
Biology, Hacettepe University
Date: 20.02.2019
iv
I hereby declare that all information in this document has been obtained and
presented in accordance with academic rules and ethical conduct. I also declare
that, as required by these rules and conduct, I have fully cited and referenced all
material and results that are not original to this work.
Name, Surname:
Signature:
Nurbahar Usta Baykal
v
ABSTRACT
DETERMINING POTENTIAL NICHE COMPETITION REGIONS
BETWEEN KAZDAGI FIR (ABIES NORDMANNIANA SUBSP. EQUI-
TROJANI) & ANATOLIAN BLACK PINE (PINUS NIGRA SUBSP.
PALLASIANA) AND CONSERVATION PRIORITY AREAS UNDER
CLIMATE CHANGE SCENARIOS BY USING MAXENT ALGORITHM
Usta Baykal, Nurbahar
Master of Science, Biology
Supervisor: Prof. Dr. Zeki Kaya
February 2019, 57 pages
Kazdagi Fir (Abies nordmanniana subsp. equi-trojani) is an endemic coniferus
tree subspecies in Turkey. The species has a narrow distribution and its conservation
status defined as EN (Endangered) by IUCN (International Union for Conservation of
Nature), due to its decreasing population size. Besides the human activities unfavoring
the species viability, it also suffers from a severe competition which has overlapping
distribution with Kazdagi fir; Anatolian Black Pine (Pinus nigra subsp. pallasiana).
The study presented here, aimed to detect conservation priority areas for
Kazdagi fir. In order to achieve this, species distribution modeling approach by using
MAXENT algorithm was used to model both Kazdagi fir and Anatolian Black Pine’s
potential distributions in 2050 by considering the possible effects of global climate
change in near future. Then, a series of overlay analyses were made to be able to detect
the areas which are better habitats for Kazdagi fir than Anatolian Black Pine and which
are better habitats for Anatolian Black Pine than Kazdagi fir.
The results of the study revealed several regions. Yet, considering the current
distribution of the species and its dispersal limits, this study proposes two conservation
vi
priority areas for Kazdagi fir; Uludag and Kazdagi (Mt. Ida) and a conservation line
for Anatolian black pine along Duzce-Bolu-Ankara. The assessed regions are the most
important habitats for both species according to both currently and in 2050 climate
scenarios. Thus, it is crucial that forestr and conservation practises should be taken
into consideration in these areas.
Keywords: Abies nordmannianna subsp. equi-trojani, Climate Change, Maxent,
Conservation, Species Distribution Models
vii
ÖZ
KORUMA ÖNCELİKLİ ALANLARIN BELİRLENMESİ İÇİN KAZDAĞI
GÖKNARI(Abies nordmanniana subsp. equi-trojani) VE KARAÇAM (Pinus
nigra subsp. pallasiana) ARASINDAKİ İKLİM DEĞİŞİKLİĞİNE BAĞLI
OLASI HABİTAT REKABETİ BÖLGELERİNİN MAXENT ALGORİTMASI
KULLANILARAK TESPİTİ
Usta Baykal, Nurbahar
Yüksek Lisans, Biyoloji
Tez Danışmanı: Prof. Dr. Zeki Kaya
Şubat 2019, 57 sayfa
Kazdağı Göknarı (Abies nordmanniana subsp. equi-trojani), Türkiye’de dar
bir dağılım gösteren, Dünya Doğa Koruma Birliği (IUCN) tarafından koruma statüsü
tehlikede (EN) olarak tayin edilmiş, iğne yapraklı bir alt türdür ve populasyon durumu
gittikçe azalan olarak belirlenmiştir. Türün habitatını ve yaşayabilirliğini olumsuz
yönde etkileyen insan aktivitelerinin yanı sıra, bir başka iğneli yapraklı bir alt tür olan
Karaçam (Pinus nigra subsp. pallasiana) ile habitat rekabeti içindedir. Bu rekabet,
son yıllarda iki türün çakışan habitatlarında yapılan ağaçlandırma çalışmalarının
Karaçam’ı destekleyecek yönde gerçekleştirilmesiyle daha da şiddetli hale gelmiştir.
Bu çalışmanın amacı, Kazdağı Göknarı için koruma öncelikli alanların
belirlenmesidir. Bunu gerçekleştirmek için MAXENT algoritması aracılığıyla tür
dağılım modellemesi yaklaşımı kullanılmıştır. Modelleme çalışmalarıyla, iki türün de
2050 yılındaki, iklim değişikliğine bağlı potansiyel habitatları tespit edilmiş, daha
sonra bu haritalar aracılığı ile hangi bölgelerin Kazdağı Göknarı için Karaçam’dan
daha uygun iklimsel koşullar taşıdığı belirlenmiştir. Aynı işlemler Karaçam için tekrar
edilmiştir.
viii
Sonuçlar birden fazla alan sunmaktadır. Fakat türlerin bugünkü dağılımı ve
dispersal limitleri göz önünde bulundurularak Kazdağı Göknarı için Kazdağı ve
Uludağ; Karaçam için Düzce-Bolu-Ankara hattı koruma öncelikli alanlar olarak
belirlenmiştir. Bu alanlar hem günümüzde, hem de 2050 yılı iklim senaryolarında
türler için yüksek dağılım olasılığına sahip alanlardır. Bu doğrultuda bu alanlarda
yapılan ağaçlandırma, yeniden ağaçlandırma ve koruma çalışmalarının türler için
öncelikli alanları zetecek doğrultuda olması; Kazdağı Göknarı’nın yakın gelecekte
neslinin tükenmemesi, Karaçam’ın ise neslinin tehlike altına girmemesi için önem
taşımaktadır.
Anahtar Kelimeler: Abies nordmanniana subsp. equi-trojani, İklim Değişikliği, Tür
Dağılım Modelleri, Doğa Koruma, Maxent
ix
For those who ever tried, ever failed.
Tried again. Failed again. Failed better.
x
ACKNOWLEDGMENTS
Firstly, I would like to thank my supervisor Prof. Dr. Zeki Kaya for his
unconditional support through my studies. I am grateful for his guidance and his
kindness through the very first steps of my academic journey.
I would like to thank Prof. Dr. Can Bilgin and Assoc. Prof. Çağatay Tavşanoğlu
who were always supportive and eager to help me solve any problem I have. I also
would like to thank to my former advisor Assist. Pr. Ayşegül Birand for her patience
and Ecology and Evolutionary Biology Society for their endless inspirations.
I am thankful to Eren Ada. Writing this thesis would be impossible without the
countless hours he spent to teach me the methodology behind every analysis I’ve done.
I believe, my father who cares about keeping a linden tree alive as much as he
cares about his family is my source of nature love and my mother who is the most
dedicated person through her ambitions is my source of ambitions to study. I am
grateful to them for everything they thought to me.
Lastly, I would like to thank to my partner, Melih Baykal. I feel the luckiest
person to share my being in same time and space with him.
xi
TABLE OF CONTENTS
ABSTRACT ................................................................................................................. v
ÖZ vii
ACKNOWLEDGMENTS ........................................................................................... x
TABLE OF CONTENTS ........................................................................................... xi
LIST OF TABLES ................................................................................................... xiii
LIST OF FIGURES ................................................................................................. xiv
LIST OF ABBREVIATIONS .................................................................................... xv
1. INTRODUCTION ................................................................................................ 1
1.1. Kazdagi Fir (Abies nordmanniana subsp. equi-trojani) .................................... 1
1.1.1. Taxonomy of Kazdagi Fir (Abies nordmanniana subsp. equi-trojani) ....... 4
1.1.2. Conservation Actions .................................................................................. 7
1.2. Anatolian Black Pine (Pinus nigra subsp. pallasiana) ..................................... 7
1.3. Species Distribution Models (SDMs) ................................................................ 9
1.3.1. MAXENT Approach................................................................................. 14
1.3.2. SDMtoolbox for ArcGIS .......................................................................... 16
1.4. Justification of Study ....................................................................................... 18
2. METHODS ......................................................................................................... 19
2.1. Species Occurrence Data ................................................................................. 19
2.2. Environmental Data ......................................................................................... 23
2.3. Modeling ......................................................................................................... 23
2.3.1. Variable Correlation ................................................................................. 23
2.3.2. Pseudo-Absence Sampling Bias ............................................................... 25
xii
2.3.3. Occurrence Rarefaction ............................................................................ 25
2.3.4. Modeling .................................................................................................. 27
2.3.5. Overlay Analyses ..................................................................................... 27
3. RESULTS ........................................................................................................... 31
3.1. SDM Results for Kazdagi Fir & Anatolian Black Pine .................................. 31
3.2. Overlay Analyses ............................................................................................ 36
3.2.1. The Logic Behind the Intersection Map ................................................... 38
4. DISCUSSION .................................................................................................... 45
5. CONCLUSIONS ................................................................................................ 49
REFERENCES .......................................................................................................... 53
xiii
LIST OF TABLES
TABLES
Table 1: Abies species distributed in Turkey .............................................................. 5
Table 2: All synonims of Abies nordmanniana subsp. equi-trojani ............................ 6
Table 3: Pinus nigra subspecies .................................................................................. 8
Table 4: Pinus nigra subsp. pallasiana varietes ........................................................... 9
Table 5: Examples of open source species distribution databanks............................ 11
Table 6: Examples of environmental datasets ........................................................... 12
Table 7: Examples of SDM algorithms ..................................................................... 15
Table 8: Cover ratio symbols on the e-Harita ........................................................... 20
Table 9: Tree species and their symbols on the e-Harita .......................................... 22
Table 10: WorldClim climatic model variables......................................................... 24
xiv
LIST OF FIGURES
FIGURES
Figure 1: Abies nordmanniana subsp. equi-trojani. Left: tree, Top Right: shoots,
Down Right: Male cones ............................................................................................. 2
Figure 2: Examples of SDMToolbox for ArcGIS .................................................... 17
Figure 3: A view of the forest stand map of Turkey from the GDF’s e-Harita. ....... 21
Figure 4: Spatially rarefied occurrence data for Kazdagi fir .................................... 26
Figure 5: Spatially rarefied occurrence data for Anatolian Black Pine .................... 26
Figure 6: Current SDM map of Kazdagi fir ............................................................. 33
Figure 7: 2050 SDM map for Kazdagi fir ................................................................ 33
Figure 8: Current SDM map for Anatolian Black Pine ............................................ 35
Figure 9: 2050 SDM map for Anatolian Black Pine ................................................ 35
Figure 10: Areas which have habitat suitability coefficient higher than 0,85 for
Kazdagi fir ................................................................................................................. 37
Figure 11: Areas which have habitat suitability coefficient higher than 0,85 Anatolian
black pine ................................................................................................................... 37
Figure 12: The logic of intersection map .................................................................. 38
Figure 13: Intersection of high probability distribution areas for Anatolian Black pine
and Kazdagi fir .......................................................................................................... 39
Figure 14: The areas which Kazdagi fir have higher probability distribution than
Anatolian Black pine ................................................................................................. 40
Figure 15: The areas which Anatolian black pine have higher probability distribution
than Kazdagi fir ......................................................................................................... 40
Figure 16: Assessed conservation priority areas for Kazdagi fir .............................. 43
Figure 17: Assessed conservation priority areas for Anatolian Black Pine ............. 43
xv
LIST OF ABBREVIATIONS
ABBREVIATIONS
AUC Area Under Curve
SDM Species Distribution MOdels
ROC receıver operatıng characterıstıc
xvi
1
CHAPTER 1
1. INTRODUCTION
1.1. Kazdagi Fir (Abies nordmanniana subsp. equi-trojani)
Kazdagi fir (Abies nordmanniana subsp. equi-trojani) is an endemic
coniferous tree species belonging to Pinaceae family. The species distribution starts
from the Kazdağı which located in the Northwest Anatolia follows the Black Sea cost
until Sinop Province. The scientific name equi-trojani is a reference to the ancient tale
of the Trojan horse, which might well have been constructed of wood from this species
as it grows nearby to the location of ancient Troy.
The species can grow up to 30 meters long. Buds contain high amounts of
resin. The number of buds at the tips of the side shoots variates from 5 to 7. The cones
form on the shoots which are located on top of the tree and lasted from previous year.
They can stand upright on the shoot and can be 15-20 cm long. The outer scales of the
cylindrical cones are longer than the inner scoops and the ends are curled backwards.
In older trees, the crust is thick and cracked. Needle leaves are arranged on long shoots
in comb-like form.
Needle leaves are narrow towards the shoot end. The ends of the leaves are
pointed, and the others are blunt or incised. Leaves are flat and two-sided. The upper
face of the lobe is slightly chiseled, and the lower face has two distinctly silvery white
stoma bands. Needle leaves stay on shoots for a long time such as 7-10 years. When
it falls or broken, it leaves a trail of round, double circle circles on the shoot. The body
shell of the Kazdağı fir is light gray, thin and smooth.
2
Figure 1: Abies nordmanniana subsp. equi-trojani. Left: tree, Top Right: shoots,
Down Right: Male cones
The most important and endangered populations of the species are located on
the Kazdağı. The Kazdağı reaches its highest point in Kartaltepe (1767m), which is
located on the east of Babadağı. The ridges starting from this point, follows a path
through to the north and northeast and forms the Kazdağı. The peaks formed by
different ridges are 30-40 km apart from each other. The Gürgendağı (1434m),
Katrandağı (1300m), Eybekdağı (1295m), Susuzdağı (1000m) ve Eğrikabaağaç
(800m) are main peaks which form Kazdağları. Kazdagi fir (Abies nordmanniana
subsp. equi-trojani) occupies small areas on the northern aspects of these mountain
peaks (Ata 1984). Populations of the species are isolated from each other and do not
represent a continuous distribution.
3
According to the study which carried out by Ata, there are 6 disconnected
populations of Kazdağı fir as listed below:
1. Kalabaktepe foothills, 122 ha, only north aspects, between 1200-1600m
altitude, mixed with Pinus nigra;
2. Gürgendağı, largest population, 2530 ha, only north aspects, between
1000-1434m altitude;
3. Eybekdağı, dispersed population, 1300 ha, only north aspects, between
700-1250 m altitude;
4. Susuzdağ, very dispersed population, 1000 ha, only north aspects, between
700-1295m altitude;
5. Eğrikabaağaç, dispersed population, 410 ha, only north aspects, between
650-800m altitude;
6. Ağıdağı, foothills and river sides, 150 ha,
While total coverage area of Kazdagi fir according to the study of Ata is 5512
ha, more recent studies proposes narrower coverage such as 3600 ha (Aslan, 1986).
The species covers different sized and disconnected areas from 120 ha to 2400 ha on
Kazdağı. Altitude range is defined as between 400-1650m (Aslan, 1986).
Uludağ population of Kazdagi fir, formerly treated as a different subspecies of
Abies nordmanniana named Abies nordmanniana subsp. bornmülleria. It also has a
restricted population on the high altitudes of Uludağ. It is isolated from both Kazdağı
populations and Western Black Sea populations.
Western Black Sea populations were also treated as a different subspecies of
Abies nordmanniana named Abies nordmanniana subsp. nordmanniana. This
populations are healtier and less threatened populations among all. They form
continuous forests as mixed stands with beech and oak.
4
1.1.1. Taxonomy of Kazdagi Fir (Abies nordmanniana subsp. equi-trojani)
The name equi-trojani was first used on an herbarium label (Sintenis, 1883,
pl. exicc. Iter trojanum, No 523) for the specimens collected from a nearby location to
Troja (Çanakkale). The diagnosis of the species came later on from Boissier (1884),
who treated the specimens collected by Sintenis as Abies pectinata DC. var. equi-
trojani Asch. & Sint. ex Boiss. Shortly thereafter, this taxon was included as a variety
of Abies nordmanniana Spach by Guinier (Doğu Karadeniz fir) and Maire (1908): A.
nordmanniana var. equi-trojani (Asch. & Sint. ex Boiss.) Guinier & Maire; and later
as a subspecies of silver fir: Abies alba Mill. subsp. equi-trojani (Asch. & Sint. ex
Boiss.) Asch. & Graebn.
In following years, a shift in the schools of taxonomy from using
morphological traits to internal traits due to the fact that morphological traits vary
depending on the growing environment of the individual; and internal morphology
and palynological privileges would better reflect the hereditary qualities. It became
evident that these characteristics could not be used alone in plant referencing and
classification practices. The examinations based on this shift showed that the species
is a natural hybrid of Abies cephalonica (Greek fir) and Abies bornmülleriana (Uludağ
fir) (Aytuğ, 1959).
Abies equitrojani defined as a distinct species by Ata & Merev (1987), Abies
bornmülleriana accepted as subspecies of Abies nordmannianna in 1990, classified as
subspecies of Abies cephalonica (T.S Lui, 1972) & as Abies nordmanniana subsp.
equitrojani (Govaerts,1975).
As pictured above, the classification of Abies equitrojani had been uncertain
due to its intermediate position between Abies cephalonica Loudon and Abies
nordmanniana geographically and morphologically (Mattfeld, 1925; Aytug, 1959;
Liu, 1971; Yaltırık, 1974; Bagci and Babaç, 2003; Kaya et al., 2008; Kurt et al., 2016).
Based on detailed studies of Abies equi-trojani and Abies cephalonica, Kazdagi fir
appeared to be distant genetically (Scaltsoyiannes et al., 1999; Liepelt et al., 2010) and
5
biochemically (Mitsopoulos and Panetsos, 1987). Later multivariate analyses showed
that Abies equitrojani appeared to be closer to Abies bornmuelleriana than Abies
cephalonica (Bagci and Babaç, 2003). Then, Uludağ fir classified as
Abies nordmanniana subsp. bornmuelleriana (Mattf.) Coode & Cullen, hence
Kazdağı fir become more related to Doğu Karadeniz fir (Abies nordmanniana) than
previosly it is to be thought.
In more recent studies, Abies equitrojani treated as subspecies of Abies
nordmanniana rather than a subspecies of Abies bornmullieriana (Schütt, 1991;
Yaltırık, 1993; Gülbaba et al. 1996; Velioğlu et al., 1998; Kaya et al., 2008; Debreczy
and Rácz, 2011; Kurt et al., 2016, Jasinska et al., 2017). It also is considered as
subspecies of Abies nordmanniana in IUCN Redlist (Knees, S. & Gardner, M. 2011).
Table 1: Abies species distributed in Turkey
NAME
COMMON TURKISH
NAME
Abies cilicica
Toros Göknarı
Abies cilicica subsp.
isaurica
Bozkır Göknarı
Abies nordmanniana
subsp. nordmanniana
Doğu Karadeniz Göknarı
Abies nordmanniana
subsp. equi-trojani
Kazdağı Göknarı
In the most recent study of the taxonomic position of Abies equi-trojani; the
needle characteristics were used to distinguish Abies species and the results indicates
that using single character does not distinguish Abies species but applying several
characacters shows significant differences among other species. Pairwise comparisons
of 39 characters measurements between Abies equi-trojani and other Abies species
6
shows the highest number of characters differs significantly from Abies cephalonica
(30) and Abies alba (25). The closest relationships to Abies equi-trojani were revealed
to be Abies bornmuelleriana and Abies nordmanniana, with 16 characters
significantly varied between (Jasinska et al., 2017).
The current accepted taxonomy of Abies species in Turkey were given in Table
1. Yet, to be able to clarify all the confusions of Abies taxonomy, all synonims for
Abies nordmanniana subsp. equi-trojani were given in the Table 2
(http://www.theplantlist.org).
Table 2: All synonims of Abies nordmanniana subsp. equi-trojani
NAME
Abies alba subsp. equi-trojani (Asch. & Sint. ex Boiss.) Asch. & Graebn.
Abies bornmuelleriana Mattf.
Abies cephalonica var. graeca (Fraas) Tang S.Liu
Abies cephalonica var. greaca (Fraas) T.S. Liu
Abies equi-trojani (Asch. & Sint. ex Boiss.) Mattf.
Abies nordmanniana subsp. bornmuelleriana (Mattf.) Coode & Cullen
Abies nordmanniana var. bornmuelleriana (Mattf.)
Abies nordmanniana var. equi-trojani (Asch. & Sint. ex Boiss.)
Abies olcayana Ata & Merev
Abies pectinata var. equi-trojani Asch. & Sint. ex Boiss.
Abies pectinata var. graeca Fraas
7
1.1.2. Conservation Actions
Kazdagi fir one of the target species in the In-Situ Conservation of Plant
Genetic Diversity Project (28632-TU) supported by World Bank Global Environment
Facility (GEF). Based on the survey and inventory studies performed by a team from
the Ministry of Agriculture, and Ministry of Forestry in 1994, Gürgendağ population
of Kazdağı fir declared as Gene Management Zone (GMZ). The Gürgendağ is the area
where plant species have very rich genetic diversity and are under the threat of
extinction or important from economical aspects. in order to maintain genetic diversity
and evolution in and between populations.
The Kazdağı National Park is an active conservation area since 1993 and
includes small Kazdağı fir populations as one of 77 other endemic species, 29 being
local endemic. Yet, the study which conducted by Satil (2009) reports habitat
degradation caused by the amount of visitor numbers to the National Park, especially
as a result of the annual Sarikiz Festival which is held on the summit in August; the
negative effects are caused through a lack of suitable facilities for large numbers of
visitors.
1.2. Anatolian Black Pine (Pinus nigra subsp. pallasiana)
Anatolian Black pine (Pinus nigra subsp. pallasiana) is a coniferus tree
species from Pinaceae family. It has 5 subspecies native for Europe and distributed
along the Mediterannean Sea costs (Spain, Portugal, Italy, Frane, Grece), middle
Europe (Austria, Macedonia), Black Sea costs (Russia and Crimes), Balkan Peninsula
(Romania & Bulgaria), Anatolia (Turkey), islands (Cyprus, Sicily and Corsica) and
lastly with outliers in Algeria and Morocco.
8
Table 3: Pinus nigra subspecies
NAME
LOCATION
Pinus nigra subsp. salzmannii (Dun)
Franco
Central and southern Spain
Pinus nigra subsp. laricio (Pior.) Marie
Corsica, Calabria and Sicily
Pinus nigra subsp. nigra
Austria, Italy, Balkans and Greece
Pinus nigra subsp. dalmatica (Vis.)
Franco
Croatia and Dinaric Alps
Pinus nigra subsp. pallasiana (Lamb.)
Holmboe
Balkans, southern Carpathians,
Crimea, Turkey, Cyprus and Syria.
Pinus nigra subsp. pallasiana (Lamb.) Holmboe has the widest distribution
among all Pinus nigra subspecies. It has 5 varietes in Turkey (Akkemik et al. 2011).
The species is wide-spread all along Anatolia, Black Sea and Mediterannean regions
(Davis, 1965). It has 4.2 million ha forest coverage yet 1.5 ha of the forrests are
damaged, degraded or fragmented (Ormancılık İstatistikleri 2015). Anatolian Black
Pine generally forms pure stands at 400-1400 m elevations, after 1400 m it forms
mixed stands with Pinus sylvestris, Abies spp., and Quercus spp. at lower elevations.
Anatolian Black Pine differentiated from other Pinus nigra subspecies by its
needle and cone length & the crown shape (Yaltırık, 1993). Seedling’s crowns are
generally conical, as the individual grows it becomes more rounded and flat-topped.
It has a thick, scaly furrowed bark colored dark-grey to black in younger individuals;
as tree gets older the color becomes paler. It can grow up to 20-40 m (Yaltırık, 1988).
9
Table 4: Pinus nigra subsp. pallasiana varietes
NAME
Pinus nigra Arn. subsp. pallasiana (Lamb.) Holmboe var. pallasiana
Pinus nigra Arn. subsp. pallasiana (Lamb.) Holmboe var. şeneriana (Saatçioğlu)
Yaltırık
Pinus nigra Arn. subsp. pallasiana (Lamb.) Holmboe var. yaltirikiana Alptekin
Pinus nigra Arn. subsp. pallasiana (Lamb.) Holmboe var. fastigiata Businsky
Pinus nigra Arn. subsp. pallasiana (Lamb.) Holmboe var. columnaris-pendula
Boydak
1.3. Species Distribution Models (SDMs)
Species distribution models are numerical tools for predicting potential
distribution of species with combined data of observed occurrences of species and
environmental variables within the study area. They are used to gain ecological and
evolutionary insights and to predict distributions across landscapes, sometimes
requiring extrapolation in space and time (Elith et al., 2006). These models help to
visualize the available habitats of species which have different habitat requirements,
both in the past and future climates (Kozak, 2008).
Species distribution models are also known as ecological niche models due to
fact that defining a geographical range (distribution) of a species, also means defining
the ecological niche of the species. Ecological niche can be defined as the combination
of the whole environmental conditions which allows a species to sustain its population
size (Pulliam, 2000). Species distribution models or ecological niche models are
related with the “Fundamental Niche” and “Realized Niche” which are defined by
Hutchinson. According to Hutchinson (1957), while fundamental niche defines a
'multidimensional and high-density' space in which a species can exist without
10
competing with another species; realized niche is a more restrictive niche due to
interspecific interactions of species such as competition.
Widespread use of species distribution models raised the questions that which
niche (fundamental or realized) of the species can be represented by modeling
predicted distribution of species. In their study, Soberon and Peterson (2005), assumed
that the combination of 3 different components form the ecological niche of species;
which are A) abiotic factors (such as soil type and climate aspects), B) biotic factors
(interspecific relationships such as mutualism and prey-predator relations) and M)
parts of the world “accessible” to the species in some ecological sense, without
barriers to movement and colonization. Here, A may be regarded as the geographic
expression of the Fundamental Niche (FN) and intersection of A and B represents the
geographic extent of the Realized Niche (RN) of the species while the intersection of
A and B and M is the right set of biotic and abiotic factors and that is accessible to the
species, and is equivalent to the geographic distribution of the species. Based on the
representation of the components of the potential distribution of species, they discuss
that ‘ecological niche modeling’ algorithms generally produce estimates of the FN (or
at least something more general than just the distribution). There are other researchers
suggest that species distribution models predict species-climate relationships are also
constrained by non-climatic factors, suggesting that the models reflect the realized
niches (Gusian and Zimmermann, 2000; Pearson et al., 2003). It is also proposed that
using fundamental niche and realized niche concepts in species distribution models is
not useful. Niche is redefined by Chase and Leibold (2003) as environmental
conditions which have minimum requirements for survival that allow birth rate in the
population to be equal to or greater than the death rate. Although this niche concept
seems to be more useful for model applications, it is still open to debate for being
relative to the model (Araújo and Guisan, 2006).
One of the fundamental inputs of species distribution models is the locations
of species on earth coordinate system. There are algorithms which use presence-
absence data as well, yet a collecting absence data is both time consuming and
11
unreliable process such as in the case of birds. Since they change their locations
seasonally, and cover a very large distance, it is hard to collected absence data might
be misleading. For plants which sprout once in two years or more, the absence data
also might be unreliable, since the year which data collected might be overlapped with
the hibernation year of the plant. Yet there are varied sources of presence data such as
herbariums, natural history museums and online databanks which make easier to find
necessary data for species distribution models. Elith et al (2008), conducted a study to
compare the models which need presence only data and presence-absence data. They
used 226 species from 6 regions of the world for model comparison presence-only
data to fit models, and independent presence-absence data to evaluate the predictions.
After they compare 16 different models, they found out that presence-only data
requiring models are as predictive as presence-absence data requiring models,
especially in machine-learning algorithms. Table 5 shows the online databanks for
species presence data, covering world-wide geographical range. There are also
available regional databanks of different countries, continents and biogeographical
regions.
Table 5: Examples of open source species distribution databanks
NAME
URL
Global Biodiversity Information Facility
(GBIF)
www.gbif.org
World Information Network on Biodiversity
www.conabio.gob.mx
HerpNET
www.herpnet.org
Ornithological Information System (ORNIS)
www.ornisnet.org
12
The other fundamental input for a species distribution model is environmental
variables which might be climatic variables as well as elevation, land cover, soil type.
Data sets containing these variables can be created by users with the help of
geographical information systems (GIS) programs or they might be available in online
data sets. There are many institutions and organizations that offer data sets over the
internet (Table 6).
The environmental variables used in species distribution models are depend on
the range of the study area. Indirect variables (e.g. elevation) provide more accurate
results while modelling relatively small-scaled areas or topographically complex
areas. On the contrary, direct variables (e.g. pH, temperature) provide more accurate
results when the study area is large because the predictive power of indirect variables
is very low for such areas of low resolution (Guissan & Zimmermann, 2000).
Table 6: Examples of environmental datasets
NAME
DATA CLASS
URL
WORLDCLIM
Climatic variables
CORINE
Land cover data
FAO Soils Portal
Soil type data
ASTGTM
DEM
Species distribution modellings have become a very important component of
conservation biology. It has been used as a tool to assess both land use and
environmental and climate change effects on the distribution of species (Kiensast et
al., 1996; Lischke et al., 1998; Guisan and Theurillat, 2000). Since assessing
biodiversity priority areas one of the key components of conservation biology, the
most popular application of species distribution models is in setting up conservation
priority areas (Margules and Austin, 1994).
13
Besides its prime importance as a research tool in conservation biology,
species distribution modeling recently gained importance as a tool to assess the impact
of accelerated land use and other environmental change on the distribution of
organisms (e.g. climate Lischke et al., 1998; Kienast et al., 1995, 1996, 1998; Guisan
and Theurillat, 2000), to test biogeographic hypotheses (e.g. Mourell and Ezcurra,
1996; Leathwick, 1998) and to improve floristic and faunistic atlases (e.g. Hausser,
1995). A variety of statistical models is currently in use to simulate either the spatial
distribution of terrestrial plant species (e.g. Hill, 1991; Buckland and Elston, 1993;
Carpenter et al., 1993; Lenihan, 1993; Huntley et al., 1995; Shao and Halpin, 1995;
Franklin, 1998; Guisan et al., 1998, 1999), aquatic plants (Lehmann et al., 1997;
Lehmann, 1998), terrestrial animal species (e.g. Pereira and Itami, 1991; Aspinall,
1992; Augustin et al., 1996; Corsi et al., 1999; Mace et al., 1999; Manel et al., 1999;
Mladenoff et al., 1995, 1999), fishes (Lek et al., 1996; Mastrorillo et al., 1997), plant
communities (e.g. Fischer, 1990; Brzeziecki et al., 1993; Zimmermann and Kienast,
1999), vegetation types (e.g. Brown, 1994; Van de Rijt et al., 1996), plant functional
types (e.g. Box, 1981, 1995, 1996), biomes and vegetation units of similar complexity
(Monserud and Leemans, 1992; Prentice et al., 1992; Tchebakova et al., 1993, 1994;
Neilson, 1995), plant biodiversity (e.g. Heikkinen, 1996; Wohlgemuth, 1998), or
animal biodiversity (Owen, 1989; Fraser, 1998).
Setting priority areas is very important in conservation for rare, endemic and
species whose range is known to be declined over the years. Since one of the
challenging threats for all living species is climate change (reference), modeling the
distributions of species under climate change scenarios has become one of the most
widely used tool to assess conservation status of a species (Martínez-Meyer et al.,
2004).
14
1.3.1. MAXENT Approach
Species distribution models requires algorithms to properly process species
observation and environmental data. There are several softwares based on different
algorithms can be used to build SDMs (Table 7), among them MAXENT is one of the
most widely used algorithms (Philips et al., 2006).
MAXENT algorithm is based on the principle of maximum-entropy which
states that probability distribution which best represents the current state of
knowledge is the one with largest entropy, in the context of precisely stated prior data
(Jaynes, 1957). In other words, it takes testable information or precisely stated prior
data about a probability distribution function and considers the set of all possible
probability distributions that would encode the prior data. Application of MAXENT
algorithms to SDMs is a machine learning software named MaxEnt, which takes a
set of environmental (e.g., bioclimatic) grids and georeferenced species occurrence
data (e.g. mediated by GBIF) and build a model to expresses a probability distribution
where each grid cell has a predicted suitability (a value) of habitat conditions for the
subjected species. A higher value of the function at a particular grid cell indicates that
the grid cell is predicted to have more suitable conditions for that species. It has the
advantage of allowing the use of both categorical and continuous variables (Baldwin,
2009).
MaxEnt can generate output data in raw, cumulative and logistic format
(Philips et al., 2008) Maxent's primary output is raw, yet these data are difficult to
interpret because the output values are often too small for each data point. The
cumulative data format gives the probability of finding the species of interest for each
location on a scale. This scale is between 0-1 and this output format is more
understandable when transferred into geographical information system (GIS) (Philips
et al., 2006). Yet, the values are not proportional to each other in cumulative data
format, which causes improper visualization of results in GIS programs. Logistic
15
format more accurately reflects the difference in output values which are between 0-1
scale thus it is more useful over other output formats (Baldwin, 2009).
Table 7: Examples of SDM algorithms
Algorithm
Software
URL
BIOCLIM
DIVA-GIS
www.diva-gis.org
Domain
DIVA-GIS
www.diva-gis.org
GARP
DESKTOPGARP
https://desktop-
garp.software.informer.com/
Generalized Additive
Model (GAM)
GRASP
https://www.unine.ch/cscf/grasp
MAXENT
MAXENT
https://www.cs.princeton.edu/~scha
pire/maxent/
MaxEnt also allows to measure variable importance on predicted distribution.
It can be determined in two ways. First, in the final model MaxEnt provides the
percentage of contribution for each variable. In case of existence of correlation
between two or more variables, results are prone to indicate more importance to them
than actual. Second method is jackknife approach which excludes one variable at a
time when running the model. In so doing, it provides information on the performance
of each variable in the model in terms of how important each variable is at explaining
the species distribution and how much unique information each variable provides
(Baldwin, 2009).
Other important feature of MaxEnt is that it allows to evaluate the model to
determine its relevance. As with any modeling approach, it is important to determine
the fit or accuracy of the model. Model evaluation primarily has been done in two
ways. First method is to calculate area under curve (AUC) of receiver operating
characteristic (ROC) plot. The scale of AUC value is between 0 and 1. Values close
16
to 0.5 indicate a fit no better than that expected by random, while a value of 1.0
indicates a perfect fit (Baldwin, 2009).
1.3.2. SDMtoolbox for ArcGIS
SDMtoolbox (Brown, 2014) is a python-based ArcGIS (ESRI 2011) toolbox
for spatial studies of ecology, evolution and genetics (Brown, 2014; Brown et al.,
2017). The toolbox automates repetitive and complicated and spatial analyses. By
carefully processing of occurrence data, environmental data and model
parameterization it maximizes each model’s discriminatory ability and minimize
overfitting.
SDMtoolbox directly interfaces with Maxent. Besides it provides many of
other analyses are for use on other species distribution modeling methods (see:
Universal Tools). The current version contains a total 78 scripts that are applicable in
fields such as landscape genetics, evolutionary studies and macroecology. For
example;
Species richness calculations,
Corrected weighted endemism calculations,
Generation of least‐cost paths and corridors among shared haplotypes,
Assessment of the significance of spatial randomizations
Enforcement of dispersal limitations of SDMs projected into future climates
17
Also, there are several generalized tools for conversion of GIS data types or
formats and batch processing. For example, it allows to arrange the same coordinate
system format for all type of datasets at the same time. The other batch processing
example is that it allows to mask several data in same extent at the same tam.
Figure 2: Examples of SDMToolbox for ArcGIS
18
1.4. Justification of Study
While Kazdagi fir is an endemic subspecies in Turkey, it is known that it has
shadow pressure on Anatolian Black pine in Kazdagi region (Ata, 1975). On the other
hand, the forestation practices favored Anatolian Black pine plantations following a
wildfire in 1975 in formerly known pure Kazdagi fir stands, turned into Anatolian
Black pine stands, artificially (Asan, 1984). It is also reported that in small habitat
patches, Kazdagi fir naturally took over stand dominance in mixed stands with
Anatolian Black pine (Asan, 1984).
The aim of the study is to resolve both niche competition and conservation priority
competition between species by using an unbiased modeling method. The
determination of the priority areas based on the ecological factors for species will both
ease the future conservation practices and decrease the extinction risks.
19
CHAPTER 2
2. METHODS
2.1. Species Occurrence Data
One of the most favorable features of MAXENT is that it allows to build
species distribution models with presence-only data. Since to prove the absence of a
species in a certain area requires very-long term fieldworks and careful analysis,
presence-only data were used in this study.
There are several methods to collect species occurrence (presence) data such
as observatory fieldworks, herbarium records and museum collections. Current
computational techniques allow to record and share all type of species occurrence data,
including online data. For example; Global Biodiversity Information Facility GBIF
(https://www.gbif.org) is an international network and research infrastructure funded
by the world’s governments and aimed at providing open access to data about all
species presence as coordinate information. There are more than 1.000.000.000
occurrence records on the network. Such network built in Turkey, to collect
biodiversity of Turkey named Nuhun Gemisi (Noah’s Ark), yet the information
sharing of the network is very limited and requires special permissions.
Yet, the data on endemic species is very scarce on GBIF, hence there are only
3 occurrences of Abies nordmanniana subsp. equi-trojani. Eventhough, there are more
than 50 records of Pinus nigra subsp. pallasiana, the geographical distribution of the
data does not cover the actual distribution of the species.
Forest stand maps - the basic unit of forest mapping; a group of trees that are
more or less homogeneous with regard to species composition, density, size, and
sometimes habitat- are other useful tools to collect occurrence data for tree species. In
Turkey, General Directorate of Forestry published an open access web-tool for forest
20
stand map of Turkey named “e-Harita”
(https://www.ogm.gov.tr/Sayfalar/OrmanHaritasi.aspx).
The e-Harita has several information about both the species and the forest
which are available online. It has the most dominant 1 to 3 species information (Table
x), the species’ growth-age information (Table Z) and the cover ratio of the area (Table
T). It uses symbols for all three types of information. It has the land/forest cover class
information represented by colors (Figure 3). It also has the tool to take coordinate
information of any point on the map.
Table 8: Cover ratio symbols on the e-Harita
Symbol
Class
Cover Ratio (%)
0
Forestration
-
ÇB
Severely degraded stands
CR < 10
B
Degraded stands
10 < CR < 40
2
-
40 < CR < 70
3
-
CR > 70
In this study, e-Harita was used to collect occurence data for Pinus nigra subsp.
pallasiana and Abies nordmanniana subsp. equi-trojani. Based on the distribution
areas which are defined on Flora of Turkey, the forest stand maps were scanned
carefully to detect all the stand of both species are present pure or mixed, and then one
or two coordinate information were taken from inside of the stands. For the coordinate
(presence data) to be taken, stands required to have the conditions;
Natural stand no reforestration / regenation / plantation
The most or the second dominant species in the stand
Age status a, ab, b, bc, bd, c, cd, d
Cover ratio > 10
21
Figure 3: A view of the forest stand map of Turkey from the GDF’s e-Harita.
Colors on the left-hand side represent forest/land cover classes on e-Harita,
symbols in the middle of brown outlined divisions (stands) carries the information
about the most dominant species in the stands, their average age and the cover range.
1. Generation, 2. Silvicultural Practices, 3. Industrial plantation, 4. Managed forests,
5. Continious forest, 6. Marshland, 7. Damaged stand, 8. Water, 9. Woodless forest
area, 10. Non-forest woodland, 11. Non-forest woodless area, 12. Settlement area, 13.
Private forest, 14. Private plantations, 15. National park, 16. 2B
22
Table 9: Tree species and their symbols on the e-Harita
CONIFERUS TREES
DECIDUOUS TREES
DECIDUOOUS TREES
Species
Symbol
Species
Symbol
Species
Symbol
Turkish pine
Çz
Beech
Kn
Walnut
Cv
A.Black pine
Çk
Oak
M
Olive
Zy
Scotch pine
Çs
Hornbeam
Gn
Valonia oak
Mp
Fir
G
Alder
Kz
Common oak
Ms
Spruce
L
Poppler
Kv
Sessile oak
Mz
Cedar
S
Chesnut
Ks
Hungarian oak
Mc
Juniper
Ar
Ash
Downy oak
Mt
Stone pine
Çf
Linden
Ih
Aleppo oak
Mm
Cypress
Sr
Maple
Ak
Turkish oak
Ml
Yew
P
Elm
Ka
Evergreen oak
Mr
Aleppo pine
Çh
Boxwood
Ş
Kermes oak
Mk
Maritime pine
Çm
Plane
Çn
Arbutus
Ko
Monterey pine
Çr
Ocalyptus
Ok
Scrub
Ma
Douglas fir
D
Sweetgum
Birch
H
Syrian juniper
An
Hazelnut
Fn
Rhododendron
O
Other conifers
Di
Willow
Others
Dy
23
2.2. Environmental Data
One of the most widely used environmental datasets is WorldClim-Global
Climate Data. WorldClim database offers climatic models which are created with
different modeling techniques and in different resolutions. The data are IPPC5 climate
projections from global climate models (GCMs) for four representative concentration
pathways (RCPs). These data were created based on the global and regional data from
meteorological stations. WorldClim also offers future climate scenarios in 4
resolutions; 10 arc-minutes, 5 arc-minutes, 2.5 arc-minutes, 30 arc-seconds and for
two different years; 2050 and 2070.
Table 10 shows the avaliable climatic variables on WorldClim. In this study,
4 different climatic data set were used; CCSM4, MIROC-ESM, HadGEM2-ES,
MIROC5. The climate data offers 3 RCP values for each scenario, which are 2.6, 4.5
and 8.5, from mild climatic change to severe climatic change, respectively. RCP
values for all climatic variables were selected as 4.5.
2.3. Modeling
2.3.1. Variable Correlation
The climatic variables available are shown in Table 9. Since most of them are
derivatives of elevation, they tend to correlate. Using correlated variables in modeling
studies causes overfit problems in Maxent algorithms which means that the predicted
occurrences of the subjected species mostly occur around the observation points.
To reduce the overfit problem, the most correlated variables among the 19
climatic variables were eliminated from the dataset. One of the widely used methods
to eliminate correlated variables is vifcor analysis in R (https://cran.r-project.org/)
‘usdm’ package (Naimi, 2010). It automatically eliminates the most highly correlated
variables.
24
Since the geographical extent which was used in the modeling for both species
are the same, the climatic data for both species are the same. After the vifcor analysis
of the climatic data, the eliminated variables are also the same for both species, hence,
one vifcor analysis was made. After the vifcor analysis, the variables were used in the
model follows: BIO1, BIO2, BIO3, BIO4, BIO9, BIO12, BIO13, BIO14, BIO15,
BIO19
Table 10: WorldClim climatic model variables
Bioclimatic
Variables
Definitions
BIO1
Annual Mean Temperature
BIO2
Mean Diurnal Range
BIO3
Isothermality (BIO2/BIO7) (* 100)
BIO4
Temperature Seasonality (standard deviation *100)
BIO5
Max Temperature of Warmest Month
BIO6
Min Temperature of Coldest Month
BIO7
Temperature Annual Range (BIO5-BIO6)
BIO8
Mean Temperature of Wettest Quarter
BIO9
Mean Temperature of Driest Quarter
BIO10
Mean Temperature of Warmest Quarter
BIO11
Mean Temperature of Coldest Quarter
BIO12
Annual Precipitation
BIO13
Precipitation of Wettest Month
BIO14
Precipitation of Driest Month
BIO15
Precipitation Seasonality (Coefficient of Variation)
BIO16
Precipitation of Wettest Quarter
BIO17
Precipitation of Driest Quarter
BIO18
Precipitation of Warmest Quarter
BIO19
Precipitation of Coldest Quarter
25
2.3.2. Pseudo-Absence Sampling Bias
One of the other error causes of MAXENT models are pseudo-absence
sampling biases, where some areas in the landscape are sampled more intensively than
others (Phillips et al. 2009). Accumulated occurrences on one sampling area again will
cause the overfit problem mentioned above. To reduce the pseudo-absence sampling
bias, ArcGIS SDMToolBox offers 4 bias reducing tools.
1) Gaussian Kernel Density of Sampling Localities: This tool creates a
Gaussian Kernel Density of sampling localities. Output values of 1
reflect no sampling bias, whereas higher values represent a high
occurrence of sampling bias.
2) Sample by Buffered Local Adaptive Convex-Hull: This tool limits
the selection of background points to an area encompassed by a
buffered regional convex-hull based on species occurrences.
3) Sample by Buffered MCP: This tool limits the selection of
background points to an area encompassed by a buffered minimum-
convex polygon based on observed localities.
4) Sample by Distance from Observation Points: This tool limits the
selection of background points to an area encompassed by a buffered
applied to each observation locality.
In this study, bias selection for both Kazdagi fir and Anatolian Black Pine was
made with Sample by Distance from Observation Points since it reflects the best the
occurrence data collection method. Buffer distance was selected as 1 km for each case.
2.3.3. Occurrence Rarefaction
Most SDM techniques require an unbiased sample. SDMToolbox has Spatially
Rarefy Occurence Data for SDMs (reduce spatial autocorrelation) tool which removes
spatially autocorrelated occurrence points by reducing multiple occurrence records to
a single record within the specified distance. For both species 5 km distance was used
to rarefy occurence data. The distribution of rarefied Kazdagi fir and Karacam are
shown in Figure 4 and Figure 5, respectively.
26
Figure 4: Spatially rarefied occurrence data for Kazdagi fir
Figure 5: Spatially rarefied occurrence data for Anatolian Black Pine
27
2.3.4. Modeling
SDMs of year 2050 projections for Kazdagi fir and Anatolian black pine were
made with ArcGIS SDMToolbox, Spatial Jackknifing Tool.
Both current and future climatic variables downloaded from WORLDCILM
version 1.4, 30 Arcsec resolution. They were extracted by using a mask covering the
area between 26 and 46 longitude and 35 and 43 latitude to include all Turkey.
For Kazdagi fir, number of spatial groups was selected as 3 while for Anatolian
black pine it was 5. Regularization parameters were taken as 1,2,4,5,7,10 for both.
Crossvalidation method was selected to run the model since the number of occurrences
were relatively high. Each model repeated 20 times, and 3 times for each variable. The
modelling process was repeated for the 4 climate scenarios; CCSM4, MIROC-ESM,
HadGEM2-ES and MIROC5. Then the mean of the future predictions of the 4
different climate scenarios were taken to decide the final future prediction result, by
using DIVA-GIS software (http://www.diva-gis.org).
In the first trials of the modeling process, some of the waterbodies in Turkey,
such as Lake Van and Lake Tuz were resulted as suitable habitats. It is expected in
some cases because the data used in modeling are only climatic data, they do not
include the land type information. To be able to reduce the bias occurred in
waterbodies, “Inland Water” dataset was downloaded from the open source web server
of DIVA-GIS (http://www.diva-gis.org/gdata).
2.3.5. Overlay Analyses
A series of numerical analyses were made during the study to assess
conservation priority areas for both Anatolian Black pine and Kazdagi fir.
Model results are raster files, which means every pixel of the data file has a
number value, in this case they refer to the probability of distribution of the subjected
28
species in terms of climatic adaptations. Model results being numerical values gives
the chance to do multiple analyses using them.
The first analysis was done to detect the maximum probability habitat
intersections of Kazdagi fir and Anatolian Black pine. Steps for intersection analysis
are given below:
1. 2050 model results give a probability for each cell a value between 0
and 1. The cells have value 1 represents the highest probability of
presence. To be able to do an intersect analysis, a threshold must be
chosen to overlay both species model results. For this study, 0.85
selected as a threshold to take consideration only the highest
probability areas in further analyses.
Spatial Analyst Tools -> Reclass -> Slice (15 Natural Breaks)
Spatial Analyst Tools -> Extraction -> Select by Attributes
(Cell Value > 0.85)
2. First, the areas which have higher probability then 0.85 were selected
and assigned to a different raster files. Then they converted into
shapefiles which “Intersect” tool of ArcMap 10.2 requires.
Conversion Tools -> Raster to Feature
3. The two shapefiles, one is the probability of distribution of Kazdagi fir
higher than 0.85, the other for Anatolian black pine were intersected.
Analysis Tools -> Overlay -> Intersect
4. Results of intersection analysis were saved as a shapefile.
Export Data
Finaly, the intersection map shows the potential niche competition regions for
both species due to 2050 climate scenarios. Yet to assign conservation priority areas
requires further analysis. Since Kazdagi fir has conservation priority over Anatolian
black pine due to IUCN, the first selection was done for Kazdagi fir. Steps for selection
of conservation priority areas for Kazdagi fir are given below:
29
1. 2050 model results of both species were taken and extracted the areas
which Kazdagi fir has higher probabilty of distribution than Anatolian
black pine.
Spatial Analyst Tools -> Local -> Higher Than Frequency
Analysis
2. The result of analysis, which shows the areas which Kazdagi fir has
higher probability of distribution, is saved as a shapefile to use in further
intersect analysis.
3. First intersection map was used for mask to extract raster values from
model results since they lost the cell value information in the
conversation to shapefile step.
4. The potential niche competition regions map and the areas which have
higher probability of distribution for Kazdagi fir than Anatolian black
pine map was again intersected to assign the conservation priority areas
for Kazdagi fir.
The assessment of conservation priority areas for Anatolian black pine follows
almost the same workflow as Kazdagi fir.
1. 2050 model results of both species were taken and extracted the areas
which Anatolian black pine has higher probability of distribution than
Kazdagi fir.
Spatial Analyst Tools -> Local -> Higher Than Frequency
Analysis
2. The result of analysis, which shows the areas which Anatolian black
pine has higher probability of distribution, is saved as a shapefile to use
in further intersect analysis.
3. First intersection map was used for mask to extract raster values from
model results since they lost the cell value information in the
conversation to shapefile step.
4. The potential niche competition regions map and the areas which have
higher probability of distribution for Anatolian black pine than Kazdagi
fir map were again intersected to assign the conservation priority areas
for Anatolian black pine.
31
CHAPTER 3
3. RESULTS
3.1. SDM Results for Kazdagi Fir & Anatolian Black Pine
Cumulative format is used for the model outputs created in MaxEnt. Each cell
created by the software contains values ranging from 0 to 1 according to the possibility
of the species being present in the cell in question. The lowest probability is indicated
by 0 and the highest probability is indicated by 1. Model for each color group, the
value range for each color group is 0.2. In graphic-wise red colors represent the highest
probability, while blue colors represent the lowest probability of distribution.
SDM results for both species shows available habitat decrease for both species.
This means that climate change will affect both species, negatively. Considering the
potential niche overlaps between species, makes species more vulnerable to future
changes.
The AUC values of models were 0.838 for Kazdagi fir, and 0.814 for Anatolian
black pine.
The probability distribution maps of both species shown in Figure 7 and Figure
8. Kazdagi fir has probability to presence in 2050 more than expected since it is
considered an endemic tree subspecies. Although all the high probability areas than
the actual occurrence areas are already habituated by either subspecies or species of
genus Abies. Besides the AUC value of the model, this fact provides a proof for model
being realistic.
32
Species distribution modeling result of Kazdagi fir (Figure 8) using current
climatic variables shows the habitats which are climatically available for Kazdagi fir.
There are several areas which are not habituated by the species, yet the model predicts
them as available habitats. This is the case for most of the modeling studies and the
need for revisit Hutchinson’s Fundamental and Realized niche theorem gains more
importance.
The modeling results shows the fundamental niche of the species. This means
that it is not necessary for all the “red” areas in the model are habituated by Kazdagi
fir. There might be several reasons for fundamental niche is not realized by the species
as they mentioned before; such as, unavailable habitats patches between two available
habitats.
In case of Kazdagi fir, Mt. Simav is included in the fundamental niche, yet it
is not included in realized niche. The area is not habituated by Kazdagi fir or another
subspecies of firs in Turkey.
Kaçkar Mountains and Toros Mountains are the other available habitats for
Kazdagi fir according to the model results which is convenient because there are fir
species occupying both of them.
In the model result which are projected in 2050 climatic variables (Figure 9),
the available habitats show drastic decrease compared to current model. Yet, the core
habitats of the species, Mt. Kazdağı and Mt. Uludağ remain the same in the future.
Also Mt. Simav, have still availability in the future. But to be able to make further
interpretations a Binomial Distribution Change Analysis was made which shows that
even though the core habitats of the species remains same in the future model, the size
of each will tend to decrease.
33
Figure 6: Current SDM map of Kazdagi fir
Figure 7: 2050 SDM map for Kazdagi fir
34
Anatolian Black pine has wider distribution compared to Kazdagi fir. The
model result also shows the same trend. Additionally, the species has which have
medium availability than Kazdagi fir (yellow-green areas on the map). Also, it can be
seen by comparison of both species’ current distribution maps, Anatolian Black pine
have more suitable habitats in inner regions of Anatolia.
The SDM result map of current climatic variables reveals best fit habitat areas
for Anatolian Black pine (red areas on the map). There is an unpatched, continuous
suitable habitat line from Düzce to Ankara which is a plausible result considering the
realized niche of the species. The natural and healthiest populations of the species are
currently occupying unconnected patches in the area.
2050 model result of Anatolian Black pine partly follows the same trend with
Kazdagi fir. For example, it also retreats from Egeaen Costs to inner regions. It has
high suitability in Mediterranean Costs in current distribution model, yet in the model
for 2050 climatic conditions, it retreats through the Toros mountains.
The future SDM results of both Anatolian Black pine and Kazdagi fir, shows
decrease in suitable areas compared to current distribution models. Yet, to be able to
determine the overlapping suitable habitats of both species requires a series of overlay
analyses.
35
Figure 8: Current SDM map for Anatolian Black Pine
Figure 9: 2050 SDM map for Anatolian Black Pine
36
3.2. Overlay Analyses
Overlay analyses are series of methods to detect overlapping features of different
data sets to be able to hypothesize the ecological, geographical or anthropogenic
reason of the current case.
The aims of overlay analyses in this study are;
1. Determining habitat overlap regions between both species
2. Determining the species which have higher suitability than the other species in
the overlapping habitats.
The modeling process assigns a habitat suitability coefficient in each cell, 30 Arc-
sec in this case since the resolution of climatic variables are 30 Arc-sec, and the
coefficient varies from 0 to 1. To be able to conduct further analysis on modeling
results, a threshold must be determined since the cumulative results include both high,
low and medium suitability areas.
In this study, the threshold for determining high suitability areas was selected
as 0,85. Only the cells which have value greater than 0.85 in 2050 models were used
for both species to determine their potential niche competition regions (Figure 12,
Figure 13). The yellow areas on Figure 12 shows the areas which have habitat suitability
coefficient greater than 0.85 for Kazdagi fir and, the blue areas on Figure 13 shows the areas
which have habitat suitability coefficient greater than 0,85 for Anatolian Black pine.
37
Figure 10: Areas which have habitat suitability coefficient higher than 0,85 for Kazdagi fir
Figure 11: Areas which have habitat suitability coefficient higher than 0,85 Anatolian black
pine
38
3.2.1. The Logic Behind the Intersection Map
In this study, the aim of overlay analyses is to detect overlapping suitable
habitats in their future models for both species. The green areas in Figure 10 represents
the potential future overlaps between Kazdagi fir and Anatolian Black pine. Yellow
areas are the regions which Kazdagi fir have habitat suitability coefficient greater than
0.85 in its 2050 model, and blue areas shows the regions which Anatolian black pine
have habitat suitability coefficient greater than 0.85 in its 2050 model. Following the
yellow and blue colors, the areas where two colors overlap represented as green. Thus,
green areas show the potential future overlapping habitats between two species.
Figure 12: The logic of intersection map
The red areas in Figure 13, shows the results of intersection analysis of high
suitability areas of Kazdagi fir and Anatolian black pine. There are several important
habitats on the intersection map. Firstly, Kazdağı (Mt. Ida) resulted as an important
habitat for both species in the future. Uludağ is another important overlapping habitat
for both species. Importance of these areas comes from the current distribution of both
species. There are several mixed stands of Kazdagi fir and Anatolian Black pine on
Kazdağı and Uludağ. They also declared as national parks several years ago.
39
Figure 13: Intersection of high probability distribution areas for Anatolian Black pine and
Kazdagi fir
Intersection analysis was made to determine to overlapping regions of both
species which have greater habitat suitability coefficient than 0,85. The red areas on
Figure 14 shows the overlapping suitable habitats for both species by using their 2050
model results. This means that the red areas on the map, are potential niche
competition regions for Kazdagi fir and Anatolian Black pine.
It is important to emphasize that the potential niche competition regions
showed in Figure 14 are within the definition of Hutchinson’s fundamental niche. It
does not mean that the red areas will definitely be occupied by both species, it only
means that climatic conditions of the areas based on 2050 model results are suitable
for both species. Considering the potential niche competition regions in Mediterranean
and East Black Sea Costs, the fundamental niche concept becomes more plausible,
since the areas are not populated by both species but different species in the same
genus.
40
Figure 14: The areas which Kazdagi fir have higher probability distribution than Anatolian Black pine
Figure 15: The areas which Anatolian black pine have higher probability distribution than Kazdagi
fir
41
After the determination of potential niche competition regions in future
models, instead of subjective decision making to assess conservation priority areas an
objective overlay analysis was made.
For determination of conservation priority areas, the 2050 model results of
species were used. First, both models were filtered with the potential niche
competition regions map to get the habitat suitability coefficients of species in the
overlapping regions since after the selection of the cells which have a value greater
than 0,85 the results stored as shapefiles (the coefficient information of cells were
lost). Then, a Higher Than Frequency filter was applied for both species to detect
which species has higher habitat suitability coefficient within the potential niche
competition regions.
The yellow areas in Figure 15 shows the areas which have higher habitat
suitability coefficient for Kazdagi fir than Anatolian Black pine. Mt. Kazdağı, Mt.
Uludağ, Mt. Simav, small and unconnected habitat patches on Mt. Kaçkar, very
isolated small patches on Southwestern Turkey and relatively large patches on
Mediterranean costs are the areas which have better habitat suitability for Kazdagi fir
than Anatolian Black pine. The blue areas on Figure 16, shows the areas which have
higher habitat suitability coefficient for Anatolian Black pine than Kazdagi fir. Areas
for Anatolian Black pine are wider than Kazdagi fir, which might be considered as a
reason to give conservation priority for Kazdagi fir, in future studies. Almost all the
Black Sea costs are better habitats for Anatolian Black pine than Kazdagi fir based on
the 2050 model results which might indicate that the effects of global climate change
have the tendency to be less severe for Anatolian Black pine, than Kazdagi fir.
Hutchinson’s fundamental niche definition also applies for Anatolian Black pine
according to results because it also reveals suitable habitats for the species which are
not currently populated by the species. Regions near to Mt. Uludağ, and Mt. Kazdağı
42
also shows better habitat suitability for Anatolian Black pine, but the higher elevations
of the both mountains have better habitat suitability for Kazdagi fir.
These results gave the ability to assess conservation priority areas for the
species, without considering one of them being endemic, and endangered. Both
species considered in equal conditions, the only measurement is the higher cell values
which means higher probability for their future distributions.
The yellow areas on Figure 17, shows the assessed conservation priority areas
for Kazdagi fir. The selection made by based on the current distribution, in other words
the realized niche of the species. Mt. Kazdağı and Mt. Uludağ are already populated
by the species. According to the GDF’s e-Harita, Mt. Simav has large and very
degraded stands of Anatolian Black pine. Future forestation practices might be done
in this area in favor of Kazdagi fir.
The blue areas on Figure 18, shows the assessed conservation priority areas for
Anatolian Black pine. Since there are wider areas which have better habitat suitability
for Kazdagi fir than Anatolian Black pine, the selection was made in favor of the
biggest unpatched suitable habitat. The results show the area between Düzce-Bolu-
Ankara have climatically unpatched suitable habitats for Anatolian Black pine. It is
also plausible in terms of the realized niche of the species. Both Ankara and Bolu have
natural Anatolian Black pine forests. Yet, considering the current land use and
settlement areas in Düzce-Bolu-Ankara line, the future conservation practices for the
species should consider the forest corridors between mentioned provinces.
The analyses were done here, reveals an unbiased conservation priority areas
assessment for the species yet the results follow the same trend if the conservation
priority was given to Kazdagi fir since it has EN status by IUCN while Anatolian
Black pine has LC status. The results also parallel in conservation priority assesment
considering the potential recent speciation events on the Abies genus in Turkey. In that
43
case, Kazdağı would also be assessed for priority area for Kazdagi fir rather than
Anatolian black pine.
Figure 16: Assessed conservation priority areas for Kazdagi fir
Figure 17: Assessed conservation priority areas for Anatolian Black Pine
45
CHAPTER 4
4. DISCUSSION
In this study, species distribution modeling approach was used to analyze the
conservation priority areas for Kazdagi fir, in order to resolve the habitat conflict
favoring Kazdagi fir in certain habitat patches the species is endemic and endangered.
After the careful modeling studies, it is shown that Kazdağı (Mt. Ida) and Uludağ will
still be the most important habitats for the species. Unexpectedly, Simav Mountain
and its surrounding mountains shows high probability of presence, yet on the stand
maps it does not show occurrence. In contrary, the area around Simav Mountain,
defined as mostly degraded stands. There are two possibilities in this case; first the
area might be available for the species yet due to dispersal limits or the unavailable
habitats between its current populations and Simav Mountain, the area is not currently
populated by the species. Second possibility is that the high probability around Simav
Mountains is a model error but considering all the other high probability areas are
currently populated by the closest relatives of the species, the former possibility seems
more prone to be the case. Yet it is almost impossible to know without further field
research in the area.
Both Uludağ and Kazdağı are defined as Key Biodiversity area by Doğa
Derneği – the BirdLife partner in Turkey. There are national parks in the area, several
conservation studies were conducted and currently conducting. It is not surprising, yet
so important to prove these two habitats are very important for an endemic species by
only modeling works. It proves both the power of unbiased, statistical analysis and the
validity of modeling works as conservation evidence.
The reason this study only focus on 2050 climate scenario is to promote
immediate action to conserve Kazdagi fir. 30 years is both long enough time to help
46
the species to form healthy populations in their most adapted habitat and short enough
time to make a species destined to extinction.
One interesting point of this study to take into consideration in further studies,
is that Kazdagi fir have high probability areas in its closest relatives. It might be a hint
of recent speciation events when we consider the recent genetic studies shows
insignificant difference between all species distributed in Turkey.
There are several advantages and disadvantages for the data collection method
used in this study. First advantage is that it does not require fieldwork. Both species
have relatively wide distribution around the country, especially Black pine. Collecting
occurrence data by fieldwork observations would be both time and money consuming.
Second is that the forest stand map has the forestation area information hence the
distinguish between natural and artificial populations can be made easily. In case of
fieldworks, to detect natural and forestated areas would require additional research.
Third, most of the divisions are smaller than the resolution of environmental data
which means that even if the coordinate information from a stand does not fall onto
an individual, it does not cause problem model-wise. As long as the coordinate stays
into the same pixel of environmental data, the results will not be affected since the
environmental value of the pixel will be the same for the occurrence data. Also,
Maxent does not appear to be strongly influenced by moderate spatial error associated
with occurrence data, as location errors up to 5 km appear to have no impact on model
performance (Graham et al, 2008).
The most important disadvantage of using e-Harita to collect occurrence data
is that the loss of the rare occurrences since the map only shows the most dominant 1
to 3 species. If there are few individuals in other stands, the information cannot be
collected. Second disadvantage is that the map is not up-to-date. There is a fast
population decline on Kazdagi fir, so several occurrence data might not be correct.
Third, GDF does not share the map which could be analyzed using GIS software. The
only way to use all the information that they share is to use the map online and there
47
are very limited tools on the web-site comparing to any GIS software. For example;
selecting all the stands which are Kazdagi fir would be 3 steps and 10 minutes analysis
while on the web-site it requires manual search. On the other hand, considering all the
disadvantages of the data collection method and comparing the disadvantages of data
collection by fieldwork observations it is more favorable in terms of time and budget.
At this point, it is crucial to emphasize the fact that the proposed conservation
priority areas are the only regions which Kazdagi fir has a better climatic adaptation
than Anatolian black pine. All the other regions in the extent of the study area are more
habitable for Anatolian black pine. It is also important to emphasize that the aim of
this study is to only release competition stress for both of species in certain habitats to
form healthier populations. In this matter reforestation and forestation works have the
most important role. Uludağ and Kazdağı should be reconsidered in terms of reducing
Anatolian black pine forestation and increasing Kazdagi fir population restoration.
Even though, this study provides a new approach to use species distribution
models in conservation planning, there are several aspects of the study needs to be
improved. First, the taxonomic position of Kazdagi fir needs to be clarified. All current
and previous studies which focused on the phylogeny of Abies species in Turkey, did
not give significant results for populations of Kazdagi fir in Kazdağı as a distinct
taxonomic group. Yet, as IUCN report of the species mentioned, there is a tendency
to consider Kazdagi fir as a distinct species (Abies equi-tojani). Regional genomic
comparisons, morphological comparisons do not support this idea so far. Thus, a
detailed whole genome analysis of at least 5 closely related Abies species is necessary
to enlighten the tangled phylogeny of Abies species in Turkey. To sum up, the
uncertainty of the phylogeny of the species might cause an incorrect distribution map,
hence an overpredicted potential distribution areas for Kazdagi fir.
Secondly, the proposed conservation priority areas for both species are the
regions based on only climatic availability. Land use or land cover status of the areas
were not examined in this study. There are several variables which can be further
48
applied on the proposed conservation priority areas besides than land cover and land
use such as; road maps and water bodies. Using additional variables give the chance
to assess more precise locations for species. Yet, in this study proposing precise
locations would be inconvenient since the resolution of the distribution maps are
relatively low and the occurrence data does not have the maximum certainty. The aim
of this thesis is to provide a new approach to use species distribution models, but it is
insufficient to increase the resolution and certainty model wise.
49
CHAPTER 5
5. CONCLUSIONS
In this study, an unbiased methodology was used to assess conservation
priority areas for Anatolian Black pine and Kazdagi fir. MAXENT approach was used
to model both current and 2050 distributions of the species. Then, the potential niche
competition regions were selected by applying an intersection analysis, using the areas
which have habitat suitability coefficient higher than 0,85. On the intersection areas,
which are defined as the potential niche competition regions, the areas which have
higher habitat suitability for Kazdagi fir, and Anatolian Black pine were selected,
respectively. Among the areas which are better fit for Kazdagi fir than Anatolian Black
pine, Mt. Kazdağı, Mt. Uludağ and Mt. Simav were selected as conservation priority
areas for Kazdagi fir considering the current realized niche of the species. Among the
areas which are better fit for Anatolian Black pine than Kazdagi fir, zce-Bolu-
Ankara line was selected for conservation priority area for Anatolian Black pine
considering both the realized niche of the species and the biggest unpatched suitable
habitat.
There are two main conclusions of this study. Firstly, the species distribution
models which are built even only using climatic variables have certain guidance ability
to review the situation of the species both current and future and they can be used to
get new insights and management approaches in terms of conservation. Secondly, the
assessed conservation priority areas propose suggestions for further forestation
practices in certain important habitats in Turkey such as Kazdağı & Uludağ. The
results do not promote the only one species restricted conservation planning in
assessed conservation priority areas for them, it only suggests that it should be
considered some species have higher probability to habituate in certain areas.
50
This study also indirectly shows that there is an urgent need to produce high
resolution spatial (such as land cover) & distributional (especially for endemic
species) and climatic datasets for stronger modeling studies and for conservation
practices. All global datasets were used in this study have relatively low resolution.
Even though the models built by using global datasets gives a rough picture of the
actual situation and the chance to get new insights from the models, for narrowly
distributed endemic species they become inadequate. To be able to carry the
conservation studies further in such biodiversity rich country, the effort of ecologists
should be invested more into the development of new datasets.
51
53
REFERENCES
Aguinagalde, I., Llorente, F., & Benito, C. Relationships among five populations of
European black pine (Pinus nigra Arn.) using morphometric and isozyme markers.
Silvae genetica, 46(1), 1-5.,1997.
Akkemik, Ü. (Editor). Türkiye’nin Doğal-Egzotik Ağaç ve Çalıları I. İ. Ü. Orman
Genel Müdürlüğü Yayınlar, Ankara. 736 Pages, 2014.
Akkemik, Ü., Çinar Yilmaz, H., Kaya, A., Oral, D. Some Taxonomic Notes on Pinus
L. (Pinaceae) Taxa Naturally Grow in Turkey. Istanbul Üniversitesi Orman
Fakültesi Dergisi, 61(1), 63-78, 2010.
Alptekin, C. Ü. Anadolu karaçamı (Pinus nigra ssp. pallasiana Lamb). İstanbul
Üniversitesi, Orman Fakültesi Dergisi, 36: 132- 154, 1986.
Arbez, M., Bernard-Dagan, C., Filion, C. Intraspecific Variability of Pinus nigra
Monoterpenes-Analyses of First Results. Ann. Sci.For. 31:57-70, 1974.
Bonnet-Masimbert, M., & Bikay-Bikay, V.. Variabilite intranspecifique des isozymes
de la glutamate-oxaloacetate-transaminase chez Pinus nigru Arnold. Inter&t pour la
taxonomie des sous especes. Silvae Genet, 27(2), 71-79., 1978.
Cengel, B., Tayanc¸ Y., Kandemir, G., Velioglu, E., Alan, M., Kaya, Z. Magnitude
and efficiency of genetic diversity captured from seed stands of Pinus nigra (Arnold)
subsp. pallasiana in established seed orchards and plantations. New Forests 43:303
317, 2012.
Cengel, B., Velioglu, E., Tolun, A.A,, Kaya, Z. Pattern and magnitude of genetic
diversity in Pinus nigra Arnold subspecies pallasiana populations from Kazdağı:
implications for in situ conservation. Silvae Genetica 49(6):249256, 2000.
Debreczy, Z. And Racz, I. Conifers Around the World: Conifers of the Temperate
Zones and Adjacent Regions. Dendro Press, 1089 pages, 2012.
Ducci, F., Garavaglia, V., and Monteverdi, M.C. Conserving marginal populations of
forest trees in Euprope. Unasylva 242, Vol 65: 51-55, 2014.
Economou, A. . Growth intercept as an indicator of site quality for planted and natural
stands of Pinus nigra var. pallasiana in Greece. Forest Ecology and Management,
32(2-4), 103-115.,1990.
Enescu, C. M., de Rigo, D., Caudullo, G., Mauri, A., Houston Durrant, T.. Pinus nigra
in Europe: distribution, habitat, usage and threats. In: San-Miguel-Ayanz, J., de Rigo,
D., Caudullo, G., Houston Durrant, T., Mauri, A. (Eds.), European Atlas of Forest
Tree Species. Publ. Off. EU, Luxembourg, pp. e015138+,2016.
Farjon, A. A Handbook of the World's Conifers (2 vols.): Revised and Updated
Edition. Brills, pg 1112., 2010.
54
González-Martínez, S.C., Robledo-Arnuncio, J.J., Collada, C., Díaz,A., Williams, C.
G., Alía, R., Cervera, M.T. Cross-amplification and sequence variation of
microsatellite loci in Eurasian hard pines” Theoretical and Applied Genetics,
109(1), 103-111.,2004.
Işik, K. 1990: Seasonal Course of Height and Needle Growth in Pinus nigra Grown
in Summer-Dry Central Anatolia. Forest Ecol. Man. 35: 261-270, 1990.
Kaya, Z. Ve Temerit, A. Genetic Structure of Marginally Located Pinus nigra var
pallasiana Populations in Central Turkey. Silvae Genetica 43, 5/6 pp 272-275,1994.
Koski, V., Antola, J. National tree breeding and seed production programme for
Turkey: 1994 - 2003. Prepared in cooperation with Research Directorate of Forest
Tree Seeds and Tree Breeding. ENSO Forest Development OY LTD, 1993.
Lewontin, R.C. The genetic basis of evolutionary change.,Columbia University Press,
New York City, New York. 346.,1974,
Matziris, D. I. Variation in Growth and Branching Characters in Black Pine (Pinus
nigra Arnold) of Peloponnesos. Silvae Genetica, 38(3-4), 77-81,1989.
Mayr, E. Populations, species, and evolution, Harvard University Press, Cambridge,
Massachusetts,453,1970.
Naydenov, K. D., Tremblay, F. M., Fenton, N. J., & Alexandrov, A. . Structure of
Pinus nigra Arn. populations in Bulgaria revealed by chloroplast microsatellites and
terpenes analysis: provenance tests. Biochemical Systematics and Ecology, 34(7),
562-574., 2006.
Nei, M. (1977). F‐statistics and analysis of gene diversity in subdivided
populations. Annals of human genetics, 41(2), 225-233 ,1977.
Nei, M. (1978). Estimation of average heterozygosity and genetic distance from a
small number of individuals. Genetics, 89(3), 583-590 1978.
Nikolić, D., & Tucić, N. (1983). Isoenzyme variation within and among populations
of European black pine (Pinus nigra Arnold). Silvae Genet, 32(3-4), 80-89,1983.
Pavia, I., Carvalho, A., Rocha, L., Gaspar, M. J., & Lima-Brito, J. Physical location
of SSR regions and cytogenetic instabilities in Pinus sylvestris chromosomes revealed
by ND-FISH. Journal of genetics, 93(2), 567-571, 2014.
PE, P. R. S. (2012). GenAlEx 6.5: genetic analysis in Excel. Population genetic
software for teaching and researchan update. Bioinformatics, 28, 2537-2539, 2012.
55
Portfaix, C. Exploration of Genetic-Variability of 5 Natural Stands of Corsican Pine
(Pinus nigra ssp. laricio var corsicana) Ann. Sci. For. 46:217-232,1989.
Pritchard, J. K., Stephens, M., & Donnelly, P. Inference of population structure using
multilocus genotype data. Genetics, 155(2), 945-959, 2000.
Rajora, O. P., Rahman, M. H., Buchert, G. P., & Dancik, B. P. Microsatellite DNA
analysis of genetic effects of harvesting in old‐growth eastern white pine (Pinus
strobus) in Ontario, Canada. Molecular ecology, 9(3), 339-348, 2000.
Scaltsoyiannes, A., Rohr, R., Panetsos, K. P., & Tsaktsira, M. Allozyme frequency
distributions in five European populations of black pine (Pinus nigra Arnold).(I)
Estimation of genetic variation within and among populations.(II) Contribution of
isozyme analysis to the taxonomic status of the species. Silvae genetica, 43(1), 20-30,
2000.
Şimşek, Y., Erkuloğlu, Ö.S. Ve Tosun, S. Türkiyede karaçam (Pinus nigra Arn. ssp.
pallasiana (Lamb) Holmboe) Orijin Denemelerinin İlk Sonuçları, OAE Teknik Bülten
No: 247, 1995.
Tolun, A.A, Velioglu, E., Cengel, B., Kaya, Z. Genetic structure of black pine (Pinus
nigra Arn. subsp. pallasiana) populations sampled from the Bolkar Mountains. Silvae
Genetica 49(3):113119, 2000.
Velioğlu, E., Çengel, B., Kaya, Z. Kazdağları Doğal Karaçam (Pinus nigra Arnold
subspecies pallasiana) populasyonlarında genetik çeşitliliğin yapılanması, Forest
Tree Seeds and Tree Breeding Research Institute Directorate, Technical Bulletin No:
1, Ankara-TURKEY, 1999.
Vidakovic, M. Genetics of European Black Pine (Pinus nigra Arn.). In Annales
forestales Vol. 6, No. 3, pp. 57-86, 1974.
Vieira MLC, Santini L, Diniz AL, Munhoz CF , Microsatellite markers: what they
mean and why they are so useful. Genet. Mol. Biol. 39(3):312-328, 2016.
Yaltirik, F. Dendroloji Ders Kitabı. İstanbul Üniversitesi Orman Fakültesi Yayınları,
1988.
56
57
... It has been predicted that the areas populated by the red pine (Pinus brutia) and black pine (P. nigra), which are the dominant tree species along the Mediterranean coastline and are very widely distributed, will expand towards the mountain summits [102][103][104][105]. Both global warming and competition from the dominant species will tend to push species found in the upper forest border, such as the fir (Abies cilicica) and Lebanon cedar, up against the mountain peaks. ...
Article
Full-text available
The effects of climate change on species can influence the delicate balance in ecosystems. For this reason, conservation planning needs to take account of connectivity and the related ecological processes within the framework of climate change. In this study, we focus on the change in the ecological connectivity of the Lebanon cedar (Cedrus libani A. Rich.), which is widely distributed in the Mediterranean, particularly in the Amanus and Taurus Mountains. To this end, we evaluated the changes in spatial units providing connectivity in the potential and future distributions of the species through ecological niche modelling, morphological spatial pattern analysis, and landscape metrics. The results suggest that the species is moving to the northeast. According to the future projections, we predict that the potential habitat suitability of the species will shrink significantly and that, in the case of pessimistic scenarios, the extent of the suitable habitats will decrease, particularly in the western and central Taurus Mountain chains. A comparison of potential and future cores indicates that there will be a slight increase under the RCP 4.5 2050 scenario, whereas core areas will decrease in the RCP 4.5 2070, RCP 8.5 2050, and RCP 8.5 2070 scenarios. In addition, it is predicted that bridges would increase in the RCP 4.5 2070 and RCP 8.5 2050 scenarios but decrease in other scenarios.
... Bu modelleme yaklaşımları arasında Maxent, diğer modelleme yöntemlerine göre küçük örneklem büyüklükleri ile daha iyi performans göstermesi sebebiyle yaygın olarak kullanılmaktadır(Pearson vd., 2007; Tsoar vd., 2007;Phillips ve Dudik, 2008) Bu sebepten dolayı, bu çalışma için, Ekolojik niş modellerinden biri olan, maksimum entropi algoritmasını kullanan MaxEnt modeli kullanılmıştır. MaxEnt algoritması, mevcut bilgi durumunu en iyi temsil eden olasılık dağılımının, kesin olarak belirtilen önceki veriler bağlamında en büyük entropiye sahip olduğunu belirten maksimum entropi ilkesine dayanmaktadır(Jaynes, 1957;Usta Baykal, 2019). Başka bir deyişle, bir olasılık dağılım işlevi hakkında test edilebilir bilgi veya kesin olarak belirtilen önceki verileri alır ve önceki verileri kodlayarak tüm olasılıkları belirler (Usta Baykal, 2019). ...
Thesis
Full-text available
Makine öğrenme tekniği kullanılarak türlerin güncel ve gelecek yayılış alanlarını modellemek günümüzde önemli çalışmalardan biri haline gelmiştir. Ülkemize ait ve Peyzaj Mimarlığı meslek disiplininin en önemli tasarım elemanı olan bitkisel materyalin iklim değişikliğinden nasıl etkileneceğinin analiz edilmesi, bu türlerin bitkilendirme çalışmalarında gelecek kullanımının planlanabilmesi için büyük önem taşımaktadır. Türlerin var olduğu alanları ifade eden noktasal veriler ve bu alanlara ait sayısal biyoiklim verileri kullanılarak oluşturulmuş katmanlar sayesinde farklı iklim senaryolarına göre türün günümüz ve gelecekteki potansiyel yayılış alanları MaxEnt programı ile belirlenebilmektedir. Bu kapsamda tez çalışması 2 ana bölümden oluşmaktadır. İlk bölümde Fabaceae (Baklagiller) familyası’ndan peyzaj tasarımı çalışmalarında en yaygın kullanılan 7 tür seçilerek bu türlere ait var verileri (presence data) ve Worldclim 2.1 versiyonu 2.5 dakika (yaklaşık 16 km2) konumsal çözünürlükteki 19 adet biyoiklimsel değişken kullanılarak türün günümüz koşullarındaki potansiyel yayılış alanı tahmin edilmiştir. İkinci bölümde ise türlerin yayılış alanlarının iklim değişiminden nasıl etkileneceğini belirlemek için ise 6. IPCC raporu temel alınarak oluşturulmuş ve eşleştirilmiş model karşılaştırma Projesi (CMIP6) modellerinden olan IPSL-CM6A-LR iklim değişim modeli kullanılarak türün SSP2 4.5 ve SSP5 8.5’e senaryolarına göre 2041-2060 ve 2081-2100 periyodlarına ait potansiyel yayılış alanı modellenmiş, ayrıca türlere ait üretilen günümüz ve gelecekteki yayılış alanları arasındaki alansal ve konumsal farklar değişim analizi ile ortaya konulmuştur. Günümüz ve gelecek yayılış alanlarının modellenmesinde MaxEnt 3.4.1 versiyonu kullanılmıştır. Fabaceae familyasına ait bazı türlerin günümüz potansiyel yayılış alanı ile gelecekte iklim değişikliğinden nasıl etkileneceği belirlendiği tez çalışmasında bütün türlerin yayılış alanlarında azalma olacağı, Adenocarpus complicatus (L.) GAY ve Ceratonia siliqua L. türlerinin ise SSP5 8.5 2090 yılı senaryosunda Türkiye koşullarında artık doğal olarak yetişemeyeceği tahmin edilmiştir.
... Türün günümüz ve gelecekteki tahmini yayılış alanlarını belirlemek amacıyla ekolojik niş modellerinden biri olan MaxEnt modeli kullanılmıştır. MaxEnt algoritması, mevcut bilgi durumunu en iyi temsil eden olasılık dağılımının, kesin olarak belirtilen önceki veriler bağlamında en büyük entropiye sahip olduğunu belirten maksimum entropi ilkesine dayanmaktadır (Jaynes, 1957;Usta Baykal, 2019). ...
Article
Full-text available
Ülkemize ait ve Peyzaj Mimarlığı meslek disiplininin en önemli tasarım elemanı olan bitkisel materyalin iklim değişikliğinden nasıl etkileneceğinin analiz edilmesi, bu türlerin bitkilendirme çalışmalarında gelecek kullanımının planlanabilmesi açısından önemlidir. Çalışmada ilk olarak Adenocarpus complicatus (L.) Gay’e ait var verileri (presence data) ve WorldClim 2.1 versiyonu 2.5 dakika (yaklaşık 20 km2 ) konumsal çözünürlükteki 19 adet biyoiklimsel değişken kullanılarak türün günümüz koşullarındaki potansiyel yayılış alanı tahmin edilmiştir. İkinci aşamada ise türün yayılış alanlarının iklim değişiminden nasıl etkileneceğini belirlemek için 6. IPCC raporu temel alınarak oluşturulmuş ve eşleştirilmiş model karşılaştırma projesi (CMIP6) modellerinden olan IPSL-CM6A-LR iklim değişim modeli kullanılarak türün SSP2 4.5 ve SSP5 8.5 senaryolarına göre 2041-2060 ve 2081-2100 periyodlarına ait potansiyel yayılış alanı modellenmiştir. Ayrıca türlere ait üretilen günümüz ve gelecekteki yayılış alanları arasındaki alansal ve konumsal farklar değişim analizi ile ortaya konulmuştur. Bulgulara göre günümüz yayılış alanı uygun ve çok uygun olarak değerlendirilen alanlar 63.695 km2 olarak hesaplanmıştır. Sonuçta türün yayılış alanlarının yıllara göre giderek azalacağı, özellikle SSP5 8.5 senaryosuna göre ~2090 yılında Türkiye koşullarında türe rastlanılamayacağı tahmin edilmektedir.
Thesis
Climate change has been expressed as a process with dramatic consequences on ecosystems in recent years. Especially the changes in the distribution of living communities draw attention. Considering this fact, it is aimed to predict the possible future changes of the distribution of black pine, brutian pine and taurus cedar species in the west of the Western Mediterranean Basin. The sustainable use of the mentioned species is important because of their wide distribution and high economic demands. It is necessary to know the ecological properties of these species and their responses to the climate crisis to ensure their sustainability. The MaxEnt method was employed to determine the ecological requirements of the species and model their distribution in current and future climatic conditions. Further, the importance values of the species were calculated and modeled using the random forest method in terms of current and future climatic conditions. Climate data was obtained from the CHELSA. As a result of the MaxEnt modeling, the AUC values for black pine, brutian pine and taurus cedar species were determined as 0.881, 0.806 and 0.902, respectively. The r2 values of the importance value were determined as 0.34, 0.21 and 0.59, respectively. It has been concluded that the distribution of brutian pine will increase according to the SSP1 2.6 and SSP3 7.0 scenarios and partially decrease for SSP5 8.5. It is thought that the results obtained will provide practical and effective benefits to the planning in terms of forestry studies to be performed in the region.
Conference Paper
Full-text available
Increasing energy demands and necessity to reduce the greenhouse gas emissions are key factors driving the understanding climate change effects on woody plants. Given the pace of climate change, the question is raised whether Salix trees and shrubs as bioenergy crops will be able to adapt to the future environmental conditions in Turkey. We select willows as study species because they are expected to expand their range in Turkey under drier and warmer climatic conditions from high latitude and wet areas. And also, selecting willow shrub and tree samples will provide data for the basic bioenergy research in the long run. Thus, our objective is to determine possible climate change observation area in Turkey according to the expected effects of climate change and possible scenarios on Turkish willows as economically valuable bioenergy crops. Two hundred and fourteen willow individuals were collected from all Turkey. The habitats of Turkish willows are generally from riparian areas and high wetlands. One willow map is created by using the locations of Turkish Salix species from field studies. Following this approach, we aim to evaluate and compare the approaches of latitudinal migration by constructing new possible stations in the suitable habitats due to the results of models. We used MAXENT algorithm to have preliminary predictions on the distribution of the species and its range shifts. This study provides the first results of species distribution modeling results of Turkish Salix species. Selection of the localities was done based on the habitat suitability coefficients which gives a probability ranging from 0 to 1. Results showed that Ilgaz Mountains (Kastamonu) and Artvin and Erzurum province (Eastern Black Sea region) are observed to have available habitats for Turkish Salix species area in terms of bioclimatic suitability.
Article
Full-text available
To determine the genetic structure of black pine (Pinus nigra ARNOLD subspecies pallasiana), populations sampled from Kazdǎi (Eybekli, Asar, Katrandaǧ, Kalkim, Gürgendaǧ, Kapidaǧ, Mihlidere). Isozymes from 16 enzyme systems were investigated from haploid female megagametophytes by starch gel electrophoresis. Twenty-nine loci were resolved from the 16 enzyme systems assayed. The results indicated that the mean number of alleles per locus (A) and polymorphisms (P) did not vary significantly in the populations studied. The mean number of alleles per locus (A) was around 1.67 (range, 1.65 to 1.69). Polymorphisms varied between 51.7% in Kapidaǧ and 58.6% in Mihlidere populations. Observed heterozygosity (Hobs) was the highest (0.186) in Asar and the lowest (0.122) in Gürgendaǧ populations. The expected heterozygosities (Hexp) ranged between 0.283 (in Asar) and 0.248 (in Katrandaǧ). There were large differences between Hobs and Hexp. Ninety-four percent of the total observed genetic variation was within populations. Nei's genetic distances also showed that variation among populations is relatively low suggesting that no population differentiation has occurred. From the estimated average genetic distances between populations, it is evident that the genetic distances between population pairs were low, ranging from 0.01 to 0.04. Genetically, most similar population pairs were Eybekli-Asar and Kalkim-Gürgendaǧ, and the least similar ones were Eybekli-Katrandaǧ, Eybekli-Kapidaǧ, Eybekli-Mihlidere, and Mihlidere-Gürgendaǧ. Based on the genetic diversity measurements and genetic distance between populations, Asar (or Eybekli), Mihlidere and Gürgendaǧ populations seem to be forming genetically distinct groups. These populations were, therefore, recommended as potential Gene Management Zones (GMZ) to conserve the genetic resources of Anatolian black pine in the Kazdaǧi Region in Turkey.
Article
Full-text available
In order to determine the magnitude and pattern of genetic diversity among Anatolian Black pine (Pinus nigra Arnold subspecies pallasiana) populations sampled in Bolkar Mountains and to recommend the potential populations which may be suitable for in situ conservation of genetic resources in this species, isoenzymes from 14 enzyme systems were investigated by starch gel electrophoresis. For this reason, open pollinated seed megagametophytes of half-sib families originated from the four populations (Camliyayla, Ulukisla, Cehennemdere and Gulekdere) were used. 24 loci were resolved for the 14 enzyme systems assayed. Polymorphism (P) varied between 41% in Camliyayla and 55% in Ulukisla. The mean number of alleles per locus (A) was around 1.6 and the expected heterozygosity (H(s)) was about 21%. Moreover, only Ulukisla had a unique allele (second allele of GOT2) which gave the distinctive character to the population. Genetic diversity among populations relative to the total genetic diversity (G(ST)) was averaged as 0.070, indicating that only 7% of the total genetic diversity was among populations. Furthermore, Nei's genetic distance values ranged from 0.007 to 0.032 among population, confirming that the diversity among populations was not very high. Based on the results of this study, it is recommended that Ulukisla and Gulekdere populations could be considered for in situ conservation of genetic resources of the species in the Bolkar Mountains.
Article
Full-text available
To test how efficiently plantations and seed orchards captured genetic diversity from natural Anatolian black pine (Pinus nigra Arnold subspecies pallasiana Holmboe) seed stands, seed sources were chosen from 3 different categories (seed stands (SS), seed orchards (SO) and plantations (P)) comprising 4 different breeding zones of the species in Turkey. Twenty-five trees (mother trees) were selected from each SS, SO and P seed sources and were screened with 11 Random Amplified Polymorphic DNA (RAPD) markers. Estimated genetic diversity parameters were found to be generally high in all Anatolian black pine seed sources and the majority of genetic diversity is contained within seed sources (94%). No significant difference in genetic diversity parameters (numbers of effective alleles, % of polymorphic loci and heterozygosity) among seed source categories was found, except for a slight increase in observed heterozygosities in seed orchards. For all seed source categories, observed heterozygosity values were higher (Ho = 0.49 for SS, 0.55 for SO and 0.49 for P) than expected ones (He = 0.40 for SS, 0.39 for SO and 0.38 for P) indicating the excess of heterozygotes. In general, genetic diversity in seed stands has been transferred successfully into seed orchards and plantations. However, the use of seeds from seed orchards can increase the amount of genetic diversity in plantations further. The study also demonstrated that number of plus-tree clones (25–38) used in the establishment of seed orchards was adequate to capture the high level of diversity from natural stands.
Article
We describe a model-based clustering method for using multilocus genotype data to infer population structure and assign individuals to populations. We assume a model in which there are K populations (where K may be unknown), each of which is characterized by a set of allele frequencies at each locus. Individuals in the sample are assigned (probabilistically) to populations, or jointly to two or more populations if their genotypes indicate that they are admixed. Our model does not assume a particular mutation process, and it can be applied to most of the commonly used genetic markers, provided that they are not closely linked. Applications of our method include demonstrating the presence of population structure, assigning individuals to populations, studying hybrid zones, and identifying migrants and admixed individuals. We show that the method can produce highly accurate assignments using modest numbers of loci—e.g., seven microsatellite loci in an example using genotype data from an endangered bird species. The software used for this article is available from http://www.stats.ox.ac.uk/~pritch/home.html.
Article
The first handbook to include detailed information on all 615 conifers, temperate as well as tropical, this encyclopedic work offers users as diverse as ecologists, gardeners, foresters and conservationists the accumulated knowledge of these trees obtained in 30 years of academic research, presented in an easily searchable format. © 2010 by Koninklijke Brill NV, Leiden, The Netherlands. All rights reserved.
Article
Portugal constitutes the southwestern end limit of the distribution area of Scots pine in the world. The population of ‘Ribeira das Negras’ (‘Serra do Gerês’; NW Portugal) has been considered potentially native and peripheral. This species has 24 chromosomes, hardly distinguishable because of their similar size and shape. Cytogenetic studies are scarce and instabilities were previously reported in peripheral Scots pine populations. Here, we intended to cytogenetically characterize individuals from ‘Ribeira das Negras’ using 14 simple sequence repeats (SSRs) and 45S rDNA sequence (pTa71) as probes, by nondenaturing fluorescence in situ hybridization (ND-FISH) and FISH, respectively. Eight SSRs [(AC)10; (AG)10; (AG)12; (AAG)5; (AAC)5; (GATA)4; (GACA)5 and (GGAT)4] and pTa71 showed hybridization. The (AG)10 probe hybridized on all chromosomes and an ideogram was constructed. Each metaphase cell presented cytogenetic instabilities corroborating ‘Ribeira das Negras’ as a peripheral population. As far as we know, this is the first cytogenetic study in Scots pine using SSRs in FISH experiments.