Acta geographica Slovenica, 62-1, 2022, 7–20
FRACTAL ANALYSIS OF
SPATIO-TEMPORAL CHANGES
OF FOREST COVER IN ISTANBUL,
TURKEY
Derya Ozturk
Pressure of rapid urbanization on forests in Turkey.
D
E
R
YA
 O
ZT
U
R
K
62-1_acta49-1.qxd  20.6.2022  10:43  Page 7
Derya Ozturk, Fractal analysis of spatio-temporal changes of forest cover in Istanbul, Turkey
DOI: https://doi.org/10.3986/AGS.10206
UDC: 711.14:630(560)
COBISS: 1.01
Derya Ozturk1
Fractal analysis of spatio-temporal changes of forest cover in Istanbul, Turkey
ABSTRACT: In this study, the spatio-temporal changes in forest cover in Istanbul, one of the provinces
with the most changes in forest areas in Turkey due to the pressure of urbanization and industrialization,
were investigated using fractal analysis. The areal changes and changes in spatial patterns were determined
to assess the spatio-temporal changes in the period 2000–2017. Fragmentation/compactness and hetero-
geneity/homogeneity of forest cover were determined by fractal dimension and lacunarity index, respectively.
The results show that the forest areas have significantly decreased and become more fragmented and het-
erogeneous. In conclusion, this study reveals that fractal analysis can provide considerable information
in the examination and interpretation of spatial changes in forest areas.
KEY WORDS: fractal analysis, fractal dimension, lacunarity index, spatio-temporal changes, forest cover,
land use change
Fraktalna analiza prostorsko-časovnih sprememb gozdnatosti v Carigradu v Turčiji
POVZETEK: V članku s fraktalno analizo raziskujemo prostorsko-časovne spremembe gozdnatosti v Carigradu,
eni od turških pokrajin z največjimi spremembami gozda zaradi pritiska urbanizacije in industrializacije.
Za oceno prostorsko-časovnih sprememb v obdobju 2000–2017 smo ugotavljali spremembe površine gozda
in prostorskih vzorcev gozdnih zemljišč. S fraktalno analizo oziroma indeksom lakunarnosti smo določili
razdrobljenost/kompaktnost ter heterogenost/homogenost gozdne pokrovnosti. Rezultati so pokazali, da
so se gozdne površine znatno zmanjšale, obenem pa postale bolj razdrobljene in raznolike. V sklepu razkri-
vamo, da lahko fraktalna analiza zagotavlja precej informacij pri preučevanju in razlagi prostorskih sprememb
na gozdnih območjih.
KLJUČNE BESEDE: fraktalna analiza, fraktalna dimenzija, indeks lakunarnosti, prostorsko-časovne spre-
membe, gozdna pokrovnost, sprememba rabe zemljišč
The article was submitted for publication on June 4th, 2021.
Uredništvo je prejelo prispevek 4. junija 2021.
8
1 Ondokuz Mayis University, Department of Geomatics Engineering, Samsun, Turkey
dzozturk@gmail.com (https://orcid.org/0000-0002-0684-3127)
62-1_acta49-1.qxd  20.6.2022  10:43  Page 8
Acta geographica Slovenica, 62-1, 2022
9
1 Introduction
Forests are natural resources that provide a wide range of ecosystem services (Nunoo 2008). Forest areas
are being changed into other land use types such as agricultural land, settlements, roads etc. due to urban-
ization and industrialization. A decrease in forest areas has adverse impacts on the forest ecosystem and
the whole ecological environment associated with the forest ecosystem (Chakravarty et al. 2012; Balthazar
et al. 2015). A decrease in forest areas also has a strong negative impact on climate. Other adverse effects
are the increase in the negative effects of hydro-meteorological disasters such as erosion, floods, landslides,
and a decrease in biodiversity (Atmiş et al. 2007; Drăghici et al. 2017).
Since drastic changes in forests affect the overall ecological balance, it is crucial to determine changes,
identify the causes of the changes, and develop effective management strategies (Drăghici et al. 2017). In
this context, information technologies have been used effectively in monitoring forest areas, in recent years.
The term of digital forestry, emerging with information technologies, is defined as science and technolo-
gy that includes acquisition, integration, query, and analysis of digital data and information to ensure forest
sustainability (Zhao et al. 2005; Ianăş and Germain 2018).
Remote sensing and Geographic Information System (GIS) technologies are very important in digi-
tal forestry. Remotely sensed satellite imagery offers significant advantages over traditional ground
survey of forests with the characteristics of fast and inexpensive data acquisition of large areas with a syn-
optic view, monitoring of changes with the repetitive image acquisition and observation of the same location,
easy integration with other GIS layers and maps, provision of data for inaccessible areas, and creation of
historical data (Jovanović et al. 2018). With its spatial data analysis, simulation and decision support sys-
tem facilities, GIS offers significant advantages in forest ecosystem management, forest fires control,
silvicultural activities, and planning forest access roads (Shao and Reynolds 2006; Reddy et al. 2015). 
The rapid developments in GIS have enabled spatial data to be examined with more innovative per-
spectives. In this context, fractal analysis, an effective tool for identifying and analyzing the irregularities
of objects, events, and phenomena, has recently been integrated into GIS to study the spatial pattern of
objects in many subjects such as urbanization (Li et al. 2011; Tannier et al. 2011; Terzi and Kaya 2011; Ozturk
2017; Purevtseren et al. 2018; Man and Chen 2020), transportation (Lu and Tang 2004; Sun et al. 2007;
Dasari and Gupta 2020; Karpinski et al. 2020; Sahitya and Prasad 2020), and geology (Wang et al. 2012;
Pourghasemi et al. 2014; Ni et al. 2017; Sun et al. 2017; Yang et al. 2019). Although fractal analysis has the
potential to improve the accuracy of measurement and identification of forest areas (Lorimer et al. 1994),
very few studies have been conducted on the use of fractal analysis in forest areas. To quantify the land-
scape pattern and dynamics is necessary to monitor and evaluate the ecological consequences of the changes
in spatial usage (Tian et al. 2007; Polenšek and Pirnat 2018; Foški 2019). Fractal dimension and lacunar-
ity index are calculated within the scope of the fractal analysis. For forested areas, deforested areas, and
afforested areas, the fractal dimension shows the degree of fragmentation/compactness and the lacunar-
ity index measures the degree of spatial heterogeneity/homogeneity (Drăghici et al. 2017).
Due to anthropogenic activities, Istanbul is one of the provinces that experienced the most changes
in forest areas in Turkey. In this study, the spatial changes of forest cover during the period 2000–2017 in
Istanbul were examined based on fractal analysis and GIS. Global Forest Change data generated based on
the analysis of Landsat satellite images were used in the analyzes. Areal and spatial changes were deter-
mined in the GIS environment and the spatial pattern of forest cover for the years 2000 and 2017, deforested
and afforested areas in the period 2000–2017 were examined by fractal analysis, and the changes were dis-
cussed.
2 Study Area: Istanbul, Turkey
Istanbul (Figure 1) is located between 28°01’–29°55’ east longitude and 40°28’–41°33’ north latitude. It is
located on a peninsula surrounded by the Black Sea, Marmara Sea, the Bosphorus, and the Golden Horn.
The Bosphorus connects the Black Sea to the Marmara Sea, separates Asia and the European continent,
and divides Istanbul into two regions: the European side and the Anatolian side (Dogan 2013).
Figure 1: Location of Istanbul. p p. 10
62-1_acta49-1.qxd  20.6.2022  10:43  Page 9
Derya Ozturk, Fractal analysis of spatio-temporal changes of forest cover in Istanbul, Turkey
10
TU
RK
EY
IR
A
Q
SY
RI
A
G
EO
RG
IA
IR
A
N
BU
LG
A
RI
A
A
RM
EN
IA
RU
SS
IA
G
RE
EC
E
TR
N
C
SO
U
TH
 C
YP
RU
S
A
ZE
RB
A
IJA
N
45
°0
'E
45
°0
'E
40
°0
'E
40
°0
'E
35
°0
'E
35
°0
'E
30
°0
'E
30
°0
'E
40°0'N
40°0'N
35°0'N
35°0'N
M
ed
ite
rr
an
ea
n 
Se
a
Bl
ac
k 
Se
a
0
15
0
30
0
45
0
75
km
.
Is
ta
nb
ul
Eu
ro
pe
an
 si
de
A
na
to
lia
n 
sid
e
IS
TA
N
BU
L
M
ar
m
ar
a 
Se
a
C
on
te
nt
 b
y:
 D
er
ya
 O
zt
ur
k
M
ap
 b
y:
 D
er
ya
 O
zt
ur
k
©
 2
02
1,
 D
ep
ar
tm
en
t o
f G
eo
m
at
ic
s E
ng
in
ee
rin
g
O
nd
ok
uz
 M
ay
is 
U
ni
ve
rs
ity
62-1_acta49-1.qxd  20.6.2022  10:43  Page 10
According to the data from Turkish Meteorological Service, the average annual precipitation in Istanbul
is 817.4mm, and the average temperature is 14.4°C for the measurement period 1929–2017 (Official Statistics
2019). Istanbul has 39 administrative districts, 14 of which are on the Anatolian side and 25 on the European
side (Dogan 2013). The European side has an area of 3546km2 and the Anatolian side has an area of 1904km2
(Provincial and district surface areas 2018).
Its geographical location has given Istanbul strategic importance and led it to be one of the most pop-
ulous cities in the world (Karaburun et al. 2010). According to Address Based Population Registration of
the Turkish Statistical Institute, the population, which was 11.08 million in 2000, reached 15.46 million
in 2020 (Statistics 2021). In metropolitan cities with rapid population growth, such as Istanbul, it becomes
difficult to protect natural areas. Forest areas are being destroyed in Istanbul due to urbanization and indus-
trialization pressures (Çakir et al. 2008; Karaburun et al. 2010).
3 Data and methods
In this study, spatio-temporal changes in forest cover during 2000–2017 were determined by overlay analy-
sis and cross-tabulation; spatio-temporal changes of spatial pattern characteristics were determined by fractal
analysis. In fractal analysis, fractal dimension and lacunarity index were calculated. In this way, the degree
of fragmentation and dispersion of the forest areas during 2000–2017 was determined, and thus how much
forest areas were affected by deforestation and afforestation. In addition, the fractal dimension and the
lacunarity index were calculated for deforested and afforested areas. All analyzes were performed sepa-
rately for the European side, Anatolian side, and for the whole of Istanbul.
Forest area boundaries for 2000 and 2017 were obtained from the Global Forest Change database with
a pixel size of 1 arc second. Tree cover data are defined as canopy closure for all vegetation with height
more than 5 m. Values are encoded in the range of 0–100 as a percentage on the pixel base (Global Forest
Change 2000–2017 Database 2018). In this study, a threshold of 30% was chosen for the 2000 tree canopy
cover data, and areas with a canopy closure value greater than 30% were identified as »forest cover« and
the other areas were identified as »other areas«. Then, forest cover was coded as 1 and other areas as 0,
and binary images were generated for fractal analysis. Forest cover for the year 2000 was updated with loss
and gain data and the forest cover was determined for the year 2017. The binary images for fractal analy-
sis are shown in Figure 2.
The degrees of fragmentation/compactness and heterogeneity/homogeneity were determined using
the binary images for the forest areas 2000 and 2017 and for the deforested and afforested areas in the peri-
od 2000–2017. The binary coding of the images, the calculations of the areal changes, and the analyzes of
the spatial changes with overlay and cross-tabulation were performed using ArcGIS 10.0 software (Esri,
Redlands, CA), and the fractal analyzes for the fractal dimension and lacunarity index were implement-
ed using ImageJ software (National Institutes of Health, USA) and the FracLac plugin.
3.1 Fractal analysis
In this study, fractal analysis was used to measure the changes in the spatial pattern of forest cover. For
this purpose, fractal dimension was calculated by box-counting algorithm and lacunarity index was cal-
culated by the gliding-box algorithm. In the fractal analyzes, the calculations were made using 5 different
grid locations with a minimum grid size of 3 pixels and a maximum grid size of 45% of the image.
A fractal is a rough or fragmented geometric shape, that can be divided into parts, where approximately
each part is a reduced copy of the whole (Jiang and Brandt 2016; Diaconu et al. 2017). The fractal dimen-
sion is beneficial in determining the irregularity or roughness of man-made and natural objects that do
not conform to classical geometry (Drăghici et al. 2017) and measures the degree of irregularity, complexity,
and fragmentation of a spatial structure (Oliveira et al. 2014; Diaconu et al. 2017). A lower fractal dimen-
sion represents a more compact object, while a higher fractal dimension represents a more complex object
(Hu et al. 2015). The lacunarity index is complementary to the fractal dimension and measures the
Acta geographica Slovenica, 62-1, 2022
11
Figure 2: Binary images for fractal analysis. p p. 12
62-1_acta49-1.qxd  20.6.2022  10:43  Page 11
Derya Ozturk, Fractal analysis of spatio-temporal changes of forest cover in Istanbul, Turkey
12
Eu
ro
pe
an
 si
de
 2
00
0
A
na
to
lia
n 
sid
e 
20
00
W
ho
le
 o
f I
st
an
bu
l 2
00
0
Eu
ro
pe
an
 si
de
 2
01
7
A
na
to
lia
n 
sid
e 
20
17
W
ho
le
 o
f I
st
an
bu
l 2
01
7
Eu
ro
pe
an
 si
de
D
ef
or
es
te
d 
ar
ea
 (2
00
0–
20
17
) 
A
na
to
lia
n 
sid
e
D
ef
or
es
te
d 
ar
ea
 (2
00
0–
20
17
) 
W
ho
le
 o
f I
st
an
bu
l
D
ef
or
es
te
d 
ar
ea
 (2
00
0–
20
17
)
Eu
ro
pe
an
 si
de
A
!o
re
st
ed
 a
re
a 
(2
00
0–
20
17
) 
A
na
to
lia
n 
sid
e
A
!o
re
st
ed
 a
re
a 
(2
00
0–
20
17
) 
W
ho
le
 o
f I
st
an
bu
l
A
!o
re
st
ed
 a
re
a 
(2
00
0–
20
17
)
1:
 F
or
es
t c
ov
er
0:
 O
th
er
 a
re
as
1:
 F
or
es
t c
ov
er
0:
 O
th
er
 a
re
as
1:
 D
ef
or
es
te
d 
ar
ea
s
0:
 O
th
er
 a
re
as
1:
 A
!o
re
st
ed
 a
re
as
0:
 O
th
er
 a
re
as
C
on
te
nt
 b
y:
 D
er
ya
 O
zt
ur
k
M
ap
 b
y:
 D
er
ya
 O
zt
ur
k
So
ur
ce
: G
lo
ba
l F
or
es
t
C
ha
ng
e 
20
00
–2
01
7
©
 2
02
1,
 D
ep
ar
tm
en
t o
f
G
eo
m
at
ic
s E
ng
in
ee
rin
g
O
nd
ok
uz
 M
ay
is
 U
ni
ve
rs
ity
62-1_acta49-1.qxd  20.6.2022  10:43  Page 12
heterogeneity/homogeneity of the spatial distribution of the object (Drăghici et al. 2017). Lacunarity is
concerned with the distribution of gap sizes in a geometric object. A lower lacunarity index value repre-
sents a more homogeneous object, as it expresses gap sizes almost close to each other. In comparison, a higher
lacunarity index value represents a more heterogeneous object since it expresses gap sizes that differ from
each other (Dong 2000).
To calculate the fractal dimension using the box-counting algorithm, the most commonly used method
for calculating the fractal dimension, the object is covered with grids (boxes) of different sizes, and the
number of grids that cover all the object or part of the object is determined. The fractal dimension is cal-
culated using the different grid sizes and the numbers of the filled grids depending on grid size. The calculation
of the fractal dimension using the box-counting method is given in Equation 1 (Morency and Chapleau
2003; Peitgen et al. 2004).
(1)
Where DB is a fractal dimension by box-counting, N is the number of the grid, and S is grid size.
To calculate the lacunarity index using the gliding-box algorithm, the most commonly used method
for calculating the lacunarity index, an r x r-dimensional grid is placed in the upper left corner of the image.
The number of pixels in the grid is determined. To calculate the number of full pixels for each position of
the grid, the grid is systematically moved one column to the right, and when one row is complete, it con-
tinues on the next row (Allain and Cloitre 1991; Dong 2000). The Q(S,r) value is calculated by using Equation 2,
where the total number of positions of the grid is N(r) and the number of grid positions with S full pixels in
the r x r dimensional grid is S(s,r). Using the Q(S,r) and S values, the 1st and 2nd statistical moments (Z1 and
Z2) are calculated as in Equations 3 and 4. The lacunarity index (Λ) is calculated using the 1
st and 2nd sta-
tistical moments according to Equation 5 (Allain and Cloitre 1991):
(2)
(3)
(4)
(5)
4 Results
The values of surface area, fractal dimension, and lacunarity index of forest areas on the European side,
Anatolian side, and the whole of Istanbul in 2000 and 2017 and the amount of the change for the period
2000–2017 are shown in Table 1.
Acta geographica Slovenica, 62-1, 2022
13
Table 1: Forest area (F), fractal dimension (DB) and lacunarity index (Λ(r)) for 2000 and 2017, changes (Δ) in the period 2000–2017.
2000 2017 2000–2017
F (km2) DB Δ(r) F (km
2) DB Δ(r) ΔF (km
2) ΔDB Δ(r)
European Side 1231 1.699 1.127 1151 1.715 1.227 –80 0.016 0.100
Anatolian Side 1018 1.743 0.673 950 1.745 0.700 –68 0.002 0.027
Whole of Istanbul 2249 1.698 0.876 2101 1.708 0.933 –148 0.010 0.057
DB = (logN2 – logN1) / (logS2 – logS1)
Q (S, r) = n (S, r) / N (r)
Z1 = ∑ SQ (S, r)
Z2 = ∑ S
2Q (S, r)
Λ(r) = Z2 / Z1
2
62-1_acta49-1.qxd  20.6.2022  10:43  Page 13
Derya Ozturk, Fractal analysis of spatio-temporal changes of forest cover in Istanbul, Turkey
Table 2 shows the surface area, fractal dimension, and lacunarity index of the deforested and afforested
areas on the European side, Anatolian side, and the whole of Istanbul in the period 2000–2017.
According to Table 1, the forest area on European side was 1231 km2 in 2000, and decreased to 1151 km2
in 2017, while the forest area on the Anatolian side was 1018 km2 in 2000 and decreased to 950 km2 in 2017.
Accordingly, the forest area was 2249 m2 in 2000, and decreased to 2101 km2 in 2017.
According to Table 2, 114 km2 forest area was lost and 34 km2 area changed as forest on the European
side while 78 km2 forest area was lost and 10 km2 area changed as forest on the Anatolian side in the peri-
od 2000–2017. In this context, the loss is higher than the gain as the total loss was 80 km2 on the European
side, 68 km2 on the Anatolian side and accordingly, the total loss for the whole of Istanbul is 148 km2.
The spatial changes of forest areas on the European side, the Anatolian side, and for the whole of Istanbul
in the period 2000–2017 are shown in Figure 2. The green color indicates the forest areas in both 2000
and 2017, the red color indicates the losses (non-forest areas in 2017 while they were in forest in 2000),
and the yellow color indicates the gains (forest areas in 2017 while they were in non-forest in 2000).
The fractal dimension shows the degree of complexity of forest cover, it also explains how and to what
extent forest areas are spatially fragmented by deforestation and afforestation. A higher fractal dimension
indicates a more complex or disperse forest cover form. An increase in the fractal dimension indicates that
the forest cover has become more complex and fragmented, and irregular uses have occurred. Similarly,
the fractal dimension of deforested and afforested areas reveals the complexity of the spatial distributions
of change. The lacunarity index shows the degree of heterogeneity of the spatial distribution of forest, defor-
ested and afforested areas and is complementary to the fractal dimension. A higher lacunarity index indicates
a more heterogeneous forest cover form. An increase in the lacunarity index indicates that the forest cover
has become more heterogeneous. Similarly, the lacunarity index of deforested and afforested areas reveals
the heterogeneity of the spatial distributions of change.
According to Table 1, the fractal dimension of forest areas on the European side was 1.715 in 2017, while
it was 1.699 in 2000. The value of the lacunarity index was 1.227 in 2017, while it was 1.127 in 2000. According
to Table 2, in the period 2000–2017, the fractal dimension was 1.537 and the lacunarity index was 2.126 for
deforested areas, while the fractal dimension was 1.470 and the lacunarity index was 2.961 for afforested areas.
According to Table 1, the fractal dimension of forest areas on the Anatolian side was 1.745 in 2017, while
it was 1.743 in 2000. The value of the lacunarity index was 0.700 in 2017, while it was 0.673 in 2000. According
to Table 2, during 2000–2017, the fractal dimension was 1.585 and the lacunarity index was 1.103 for defor-
ested areas, while the fractal dimension was 1.394 and the lacunarity index was 1.340 for afforested areas.
According to Table 1, the fractal dimension for the whole of Istanbul was 1.698 in 2000 and 1.708 in
2017, while the lacunarity of the forest areas was 0.876 in 2000 and 0.933 in 2017. According to Table 2,
during 2000–2017, the fractal dimension was 1.558 and the lacunarity index was 1.419 for deforested areas,
while the fractal dimension was 1.458 and the lacunarity index was 2.769 for afforested areas.
The fractal dimension increased by 0.016 on the European side and by 0.002 on the Anatolian side.
These increases indicate that the forest areas have become more fragmented. The increase in fragmenta-
tion on the European side is higher than on the Anatolian side. The lacunarity index increased by 0.100
on the European side and by 0.027 on the Anatolian side. These increases indicate that forest areas have
become more heterogeneous. The increase in heterogeneity on the European side is higher than on the
Anatolian side.
14
Table 2: Surface area (F), fractal dimension (DB), and lacunarity index (Λ(r)) of the deforested and afforested area in the period 2000–2017.
Deforested Afforested Total loss 
Fdef (km
2) DB Δ(r) Faf (km
2) DB Δ(r)
(Fdef – Faf)
European Side 114 1.537 2.126 34 1.470 2.961 80
Anatolian Side 78 1.585 1.103 10 1.394 1.340 68
Whole of Istanbul 192 1.558 1.419 44 1.458 2.769 148
Figure 3: The spatial change of forest cover in Istanbul. p
62-1_acta49-1.qxd  20.6.2022  10:43  Page 14
Acta geographica Slovenica, 62-1, 2022
15
29
°3
0'E
29
°3
0'E
29
°0
'E
29
°0
'E
28
°3
0'E
28
°3
0'E
28
°0
'E
28
°0
'E
41°30'N
41°30'N
41°0'N
41°0'N
.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
62-1_acta49-1.qxd  20.6.2022  10:43  Page 15
Derya Ozturk, Fractal analysis of spatio-temporal changes of forest cover in Istanbul, Turkey
5 Discussion
Determining dynamics of the forest landscape is an important issue in terms of sustainable land and for-
est management (Çakir et al. 2008). Particularly in areas under urbanization and industrialization pressure,
determining changes in the landscape in terms of both area and spatial characteristics is essential for mon-
itoring and protecting forest areas in order to prepare effective development plans and sustainable forest
management plans (Çakir and Özdemir 2015; Ren et al. 2019).
Migration from rural to urban areas, which started in the 1950s in Turkey, led to excessive popula-
tion growth in Istanbul in a short period of time, especially after the 1980s (Işık 2005; Çakir et al. 2008).
The natural resources in Istanbul started to be destroyed rapidly due to various reasons resulting from this
rapid population increase. Urbanization and industrialization, which have developed in parallel with the
population increase in Istanbul, have also led to significant changes in forest areas (Atmiş et al. 2007;
Gökburun 2017).
In this study, forest areas were found to decrease by 6.5% on the European side, 6.7% on the Anatolian
side, and 6.6% in the whole of Istanbul during period 2000–2017. These rates show that the decline of 5.4%
reported by Karaburun et al. (2010) for the period 1987–2007 is still increasing, and the decline of forest
areas continues. These changes in forest areas revealed a general downward trend in Istanbul.
In addition, in terms of spatial pattern features, forest fragmentation is the fragmentation of large, con-
tinuous, and compact forest areas (Bogaert et al. 2011), and spatial heterogeneity of forest is based on the
gaps of different sizes (Buajan et al. 2017). Deforestation due to urbanization, road development, agricultural
development, and forest fires generally leads to an increase in forest fragmentation and heterogeneity
(Broadbent et al. 2008). Identification of the fragmentation/compactness and heterogeneity/homogene-
ity is essential in terms of species in the forest ecosystem and landscape characteristics (Bogaert et al. 2011;
Peh et al. 2014).
Considering the results of the fractal analysis of this study, it can be summarized that the fractal dimen-
sion of forest areas on the Anatolian side is higher than on the European side in both 2000 and 2017. This
shows that the forest areas on the Anatolian side are more fragmented and complex than those on the
European side. The Anatolian side has more forest area in percentage than the European side and the per-
centage decrease in forest areas due to urbanization activities on the Anatolian side is higher than on the
European side. This situation results in more fragmented forest areas. The lacunarity index is higher on
the European side than on the Anatolian side in both 2000 and 2017, indicating that forest areas are more
heterogeneous on the European side than on the Anatolian side. Moreover, for the European side, Anatolian
side, and the whole of Istanbul, the fractal dimension and lacunarity index have increased during the peri-
od 2000–2017, which means that complexity and heterogeneity have increased due to deforestation. The
amount of increase is greater on the European side for both fractal dimension and lacunarity index. This
shows that the changes in the form of forest areas on the European side during the 2000–2017 period were
greater than on the Anatolian side.
Fractal analysis has been used in some geographical studies related to forest change, deforestation, and
afforestation. Andronache et al. (2016), Pintilii et al. (2016), Drăghici et al. (2017), Andronache et al. (2019),
and Diaconu et al. (2020) obtained complementary interpretations for forest areas with the integrated use
of fractal dimension and lacunarity index. Andronache et al. (2016) analyzed the change of forest areas
for the period 2000–2012 in the North-Eastern Development Region in Romania, where large forest areas
were converted to different land uses due to local economic pressure. The results of the analysis showed
that the afforested areas were more compact and had more homogeneous arrangement than the defor-
ested areas. Considerable differences were found between the deforested and afforested areas. The
maximum difference was observed in Iaşi county, while the minimum difference was observed in Vaslui coun-
ty. Pintilii et al. (2016) analysed the changes in Suceava-Cârlibaba in Romania for the period 2000–2013 and
found that the deforested areas were more fragmented and heterogeneous. Drăghici et al. (2017) analyzed
the changes in the forest areas in the Northern Carpathian Mountains for the period 2000–2014, and it
was determined that the forest areas turned into a more complex and more heterogeneous structure.
Andronache et al. (2019) performed the analysis of the deforested areas of the Apuseni Mountains for the
period 2000–2014 and determined an increase in fragmentation and heterogeneity with a decrease in the
tree cover area. Diaconu et al. (2020) analyzed the deforestation in the mountain area, the northern and
central groups of the Eastern Carpathians for the period 2000–2017 and determined that deforestation
16
62-1_acta49-1.qxd  20.6.2022  10:43  Page 16
causes the fragmentation of forests. Consistent with these studies, the fragmentation/compactness and het-
erogeneity/homogeneity degree could be evaluated separately using the fractal dimension and lacunarity
index, respectively. When the results of the fractal analysis of deforested and afforested areas are exam-
ined, it is clear that the deforested areas are more fragmented and homogenous than the afforested areas
for the European side, the Anatolian side and the whole of Istanbul. Notable differences were determined
for the values of fractal dimension and lacunarity index of deforested and afforested areas. The difference
between the fractal dimension of deforested and afforested areas is higher on the Anatolian side, while
the difference between the lacunarity index is higher on the European side.
6 Conclusion
Protecting forests is vital for adapting to and mitigating climate change, protecting biodiversity and pre-
venting natural disasters such as erosion, floods and landslides. Therefore, forest areas and changes should
be monitored and necessary measures such as forest conservation and management plans should be taken.
In this study, the spatio-temporal changes of forest cover in Istanbul, Turkey, during 2000–2017 were
examined. In addition to the changes in area, the spatial patterns of both forest areas and areas of change
were investigated using fractal analysis. The results show that forest areas degraded by cutting trees due
to urbanization and industrialization pressures have resulted in a decrease in total forest area and more
fragmented and heterogeneous landscapes. The results of this study can be used to investigate possible
interactions and transitions between forest, deforested and afforested areas, and the results contribute to
the assessment of forest areas under pressure.
The evaluation of spatio-temporal changes of forest cover in Istanbul was carried out using the Global
Forest Change database (2000 and 2017). The results of the study showed that the Global Forest Change
database with a spatial resolution 1 arc second can be effectively used in the quantitative analysis of the
spatial patterns of forest, deforested and afforested areas with the integration of GIS and fractal analysis,
providing new and complementary information in addition to the traditional change detection studies.
The results obtained provide additional information that can be integrated into forest land management
and organization plans. Similar approaches can be adapted to land cover/use classes other than forests,
and new information can be provided through innovative approaches to spatial analysis.
7 References 
Allain, C., Cloitre, M. 1991: Characterizing the lacunarity of random and deterministic fractal sets. Physics
Review A 44-6. DOI: https://doi.org/10.1103/physreva.44.3552
Andronache, I. C., Ahammer, H., Jelinek, H. F., Peptenatu, D., Ciobotaru, A. M., Drăghici, C. C., Pintilii,
R. D., et al. 2016: Fractal analysis for studying the evolution of forests. Chaos, Solitons and Fractals
91. DOI: https://doi.org/10.1016/j.chaos.2016.06.013
Andronache, I., Marin, M., Fischer, R., Ahammer, H., Radulovic, M., Ciobotaru, A. M., Jelinek, H. F., et
al. 2019: Dynamics of forest fragmentation and connectivity using particle and fractal analysis. Scientific
Reports 9. DOI: https://doi.org/10.1038/s41598-019-48277-z
Atmiş, E., Özden, S., Lise, W. 2007: Urbanization pressures on the natural forests in Turkey: An overview.
Urban Forestry and Urban Greening 6-2. DOI: https://doi.org/10.1016/j.ufug.2007.01.002
Balthazar, V., Vanacker, V., Molina, A., Lambin, E. F. 2015: Impacts of forest cover change on ecosystem services
in high Andean mountains. Ecological Indicators 48. DOI: https://doi.org/10.1016/j.ecolind.2014.07.043
Bogaert, J., Barima, Y. S. S., Iyongo Waya Mongo, L., Bamba, I., Mama, A., Toyi, M., Lafortezza, R. 2011. Forest
fragmentation: Causes, ecological impacts and implications for landscape management. Landscape Ecology
in Forest Management and Conservation. Berlin, Heidelberg. DOI: https://doi.org/10.1007/978-3-642-
12754-0_12
Broadbent, E. N., Asner, G. P., Keller, M., Knapp, D. E., Oliveira, P. J. C., Silva, J. N. 2008: Forest fragmentation
and edge effects from deforestation and selective logging in the Brazilian Amazon. Biological
Conservation 141-7. DOI: https://doi.org/10.1016/j.biocon.2008.04.024
Acta geographica Slovenica, 62-1, 2022
17
62-1_acta49-1.qxd  20.6.2022  10:43  Page 17
Derya Ozturk, Fractal analysis of spatio-temporal changes of forest cover in Istanbul, Turkey
Buajan, S., Jinfu, L., Zhongsheng, H., Xueping, F., Muhammad, A. 2017: The effect of light on micro-envi-
ronment and specific leaf area within the gap, subtropical forest, China. Pakistan Journal of Botany 49-1.
Çakir, G., Özdemir, M. 2015: Determination to results of forest management planning applications with land
use changes from 1971 to 2008 in Yalova-Turkey. Journal of the Indian Society of Remote Sensing 43-1.
DOI: https://doi.org/10.1007/s12524-014-0374-8
Çakir, G., Ün, C., Baskent, E. Z., Köse, S., Sivrikaya, F., Keleş, S. 2008: Evaluating urbanization, fragmenta-
tion and land use/land cover change pattern in Istanbul city, Turkey from 1971 to 2002. Land Degradation
and Development 19-6. DOI: https://doi.org/10.1002/ldr.859
Chakravarty, S., Ghosh, S. K., Suresh, C. P., Dey, A. N., Shukla, G. 2012: Deforestation: Causes, effects and
control strategies. Global Perspectives on Sustainable Forest Management. Rijeka. DOI: https://doi.org/
10.5772/33342
Dasari, S., Gupta, S. 2020: Application of fractal analysis in evaluation of urban road networks in small
sized city of India: Case city of Karimnagar. Transportation Research Procedia 48. DOI: https://doi.org/
10.1016/j.trpro.2020.08.227
Diaconu, D. C., Andronache, I., Ahammer, H., Ciobotaru, A. M., Zelenakova, M., Dinescu, R., Pozdnyakov,
A. V., Chupikova, S. A. 2017: Fractal drainage model - A new approach to determinate the complexity
of watershed. Acta Montanistica Slovaca 22-1.
Diaconu, D. C., Papuc, R. M., Peptenatu, D., Andronache, I., Marin, M., Dobrea, R. C., Drăghici, C. C., et
al. 2020: Use of fractal analysis in the evaluation of deforested areas in Romania. Advances in Forest
Management under Global Change. London. DOI: https://doi.org/10.5772/intechopen.91621
Dogan, M. 2013: Geçmişten günümüze İstanbul’da sanayileşme süreci ve son 10 yıllık değişimi. Marmara
Geographical Review 27-1.
Dong, P. 2000: Lacunarity for spatial heterogeneity measurement in GIS. Geographic Information
Sciences 6-1. DOI: https://doi.org/10.1080/10824000009480530
Drăghici, C. C., Andronache, I., Ahammer, H., Peptenatu, D., Pintilii, R. D., Ciobotaru, A. M., Simion, A.
G., et al. 2017: Spatial evolution of forest areas in the northern Carpathian Mountains of Romania.
Acta Montanistica Slovaca 22-2.
Foški, M. 2019: Using the parcel shape index to determine arable land division types. Acta geographica
Slovenica 59-1. DOI: https://doi.org/10.3986/AGS.4574
Global Forest Change 2000–2017 Database. University of Maryland, Department of Geographical Sciences.
College Park, 2018.
Gökburun, İ. 2017: İstanbul’da nüfusun gelişimi ve ilçelere dağılımı (1950–2015). Journal of Anatolian Cultural
Research 1-3.
Hu, S., Tong, L., Frazier, A. E., Liu, Y. 2015: Urban boundary extraction and sprawl analysis using
Landsat images: A case study in Wuhan, China. Habitat International 47. DOI: https://doi.org/10.1016/
j.habitatint.2015.01.017
Ianăş, A. N., Germain, D. 2018: Quantifying landscape changes and fragmentation in a national park in the
Romanian Carpathians. Carpathian Journal of Earth and Environmental Sciences 13-1. DOI: https://doi.org/
10.26471/cjees/2018/013/014
Işık, Ş. 2005: Türkiye’de kentleşme ve kentleşme modelleri. Aegean Geographical Journal 14-1.
Jiang, B., Brandt, S. A. 2016: A fractal perspective on scale in geography. ISPRS International Journal of
Geo-Information 5-6. DOI: https://doi.org/10.3390/ijgi5060095
Jovanović, M. M., Milanović, M. M., Zorn, M. 2018: The use of NDVI and CORINE Land Cover databases
for forest management in Serbia. Acta geographica Slovenica 58-1. DOI: https://doi.org/10.3986/AGS.818
Karaburun, A., Demirci, A., Suen, I. S. 2010: Impacts of urban growth on forest cover in Istanbul (1987–2007).
Environmental Monitoring and Assessment 166-1–4. DOI: https://doi.org/10.1007/s10661-009-1000-z
Karpinski, M., Kuznichenko, S., Kazakova, N., Fraze-Frazenko, O., Jancarczyk, D. 2020: Geospatial
assessment of the territorial road network by fractal method. Future Internet 12-11. DOI: https://doi.org/
10.3390/fi12110201
Li, X., Liu, L., Dong, X. 2011: Quantitative analysis of urban expansion using RS and GIS, A case study in
Lanzhou. Journal of Urban Planning and Development 137-4. DOI: https://doi.org/10.1061/(ASCE)UP.1943-
5444.0000078
Lorimer, N. D., Haight, R. G., Leary, R. A. 1994: The fractal forest: Fractal geometry and applications in
forest science. General Technical Report NC-170. St. Paul. DOI: https://doi.org/10.2737/NC-GTR-170
18
62-1_acta49-1.qxd  20.6.2022  10:43  Page 18
Lu, Y., Tang, J. 2004: Fractal dimension of a transportation network and its relationship with urban growth:
A study of the Dallas-Fort Worth area. Environment and Planning B: Planning and Design 31-6. DOI:
https://doi.org/10.1068/b3163
Man, X., Chen, Y. 2020: Fractal-based modeling and spatial analysis of urban form and growth: A case
study of Shenzhen in China. ISPRS International Journal of Geo-Information 9-11. DOI: https://doi.org/
10.3390/ijgi9110672
Morency, C., Chapleau, R. 2003: Fractal geometry for the characterisation of urban-related states: Greater
Montreal Case. Harmonic and Fractal Image Analysis 2003.
Ni, C., Zhang, S., Chen, Z., Yan, Y., Li, Y. 2017: Mapping the spatial distribution and characteristics of lin-
eaments using fractal and multifractal models: A case study from northeastern Yunnan province, China.
Scientific Reports 7. DOI: https://doi.org/10.1038/s41598-017-11027-0
Nunoo, E. K. 2008: EIA performance standards and thresholds for sustainable forest management in Ghana.
Standards and Thresholds for Impact Assessment. Berlin, Heidelberg. DOI: https://doi.org/10.1007/978-
3-540-31141-6_18
Official Statistics. Turkish Meteorological Service. Ankara, 2019.
Oliveira, M. A. B. de, Brandi, A. C., dos Santos, C. A., Botelho, P. H. H., Cortez, J. L. L., de Godoy, M. F.,
Braile, D. M. 2014: Comparison of fractal dimension and Shannon entropy in myocytes from rats treat-
ed with histidine-tryptophan-glutamate and histidine-tryptophan cetoglutarate. Brazilian Journal of
Cardiovascular Surgery 29-2. DOI: https://doi.org/10.5935/1678-9741.20140052
Ozturk, D. 2017: Assessment of urban sprawl using Shannon’s entropy and fractal analysis: A case study
of Atakum, Ilkadim and Canik (Samsun, Turkey). Journal of Environmental Engineering and
Landscape Management 25-3. DOI: https://doi.org/10.3846/16486897.2016.1233881
Peh, K. S. H., Lin, Y., Luke, S. H., Foster, W. A., Turner, E. C. 2014: Forest fragmentation and ecosystem
function. Global Forest Fragmentation. Oxfordshire. DOI: https://doi.org/10.1079/9781780642031.0096
Peitgen, H. O., Jürgens, H., Saupe, D. 2004: Chaos and Fractals. New Frontiers of Science. New York. DOI:
https://doi.org/10.1007/b97624
Pintilii, R. D., Andronache, I. C., Simion, A. G., Draghici, C. C., Peptenatu, D., Ciobotaru, A. M., Dobrea,
R. C., Papuc, R. M. 2016: Determining forest fund evolution by fractal analysis (Suceava-Romania).
Urbanism. Architecture. Constructions 7-1.
Polenšek, M., Pirnat, J. 2018: Forest patch connectivity: The case of the Kranj–Sora Basin, Slovenia. Acta
geographica Slovenica 58-1. DOI: https://doi.org/10.3986/AGS.3001
Pourghasemi, H. R., Moradi, H. R., Fatemi Aghda, S. M., Sezer, E. A., Goli Jirandeh, A., Pradhan, B. 2014.
Assessment of fractal dimension and geometrical characteristics of the landslides identified in North
of Tehran, Iran. Environmental Earth Sciences 71-8. DOI: https://doi.org/10.1007/s12665-013-2753-9
Provincial and district surface areas. Turkish General Directorate of Maps. Ankara, 2018.
Purevtseren, M., Tsegmid, B., Indra, M., Sugar, M. 2018: The fractal geometry of urban land use: The case
of Ulaanbaatar city, Mongolia. Land 7-2. DOI: https://doi.org/10.3390/land7020067
Reddy, C. S., Jha, C. S., Diwakar, P. G., Dadhwal, V. K. 2015: Nationwide classification of forest types of
India using remote sensing and GIS. Environmental Monitoring and Assessment 187-12. DOI:
https://doi.org/10.1007/s10661-015-4990-8
Ren, C., Chen, L., Wang, Z., Zhang, B., Xi, Y., Lu, C. 2019: Spatio–temporal changes of forests in north-
east China: Insights from Landsat images and geospatial analysis. Forests 10-11. DOI: https://doi.org/
10.3390/f10110937
Sahitya, K. S., Prasad, C. S. R. K. 2020: Fractal modelling of an urban road network using Geographical
Information Systems (GIS). World Review of Intermodal Transportation Research 9-4. DOI: https://doi.org/
10.1504/WRITR.2020.111078
Shao, G., Reynolds, K. M. 2006: Introduction. Computer Applications in Sustainable Forest Management.
Including Perspectives on Collaboration and Integration. Dordrecht. DOI: https://doi.org/10.1007/978-
1-4020-4387-1_1
Statistics. Turkish Statistical Institute. Ankara, 2021.
Sun, T., Wu, K., Chen, L., Liu, W., Wang, Y., Zhang, C. 2017: Joint application of fractal analysis and weights-
of-evidence method for revealing the geological controls on regional-scale tungsten mineralization in
Southern Jiangxi Province, China. Minerals 7-12. DOI: https://doi.org/10.3390/min7120243
Acta geographica Slovenica, 62-1, 2022
19
62-1_acta49-1.qxd  20.6.2022  10:43  Page 19
Derya Ozturk, Fractal analysis of spatio-temporal changes of forest cover in Istanbul, Turkey
Sun, Z., Jia, P., Kato, H., Hayashi, Y. 2007: Distributive continuous fractal analysis for urban transporta-
tion network. Journal of the Eastern Asia Society for Transportation Studies 7. DOI: https://doi.org/
10.11175/easts.7.1519
Tannier, C., Thomas, I., Vuidel, G., Frankhauser, P. 2011: A fractal approach to identifying urban bound-
aries. Geographical Analysis 43-2. DOI: https://doi.org/10.1111/j.1538-4632.2011.00814.x
Terzi, F., Kaya, H. S. 2011: Dynamic spatial analysis of urban sprawl through fractal geometry: The case
of Istanbul. Environment and Planning B: Planning and Design 38-1. DOI: https://doi.org/10.1068/b35096
Tian, G., Yang, Z., Xie, Y. 2007: Detecting spatiotemporal dynamic landscape patterns using remote sens-
ing and the lacunarity index: A case study of Haikou City, China. Environment and Planning B: Planning
and Design 34-3. DOI: https://doi.org/10.1068/b3155
Wang, G., Carranza, E. J. M., Zuo, R., Hao, Y., Du, Y., Pang, Z., Sun, Y., Qu, J. 2012: Mapping of district-
scale potential targets using fractal models. Journal of Geochemical Exploration 122. DOI: https://doi.org/
10.1016/j.gexplo.2012.06.013
Yang, B., Yuan, J., Duan, L., Liu, Q. 2019: Using GIS and fractal theory to evaluate degree of fault complexity
and water yield. Mine Water and the Environment 38-2. DOI: https://doi.org/10.1007/s10230-018-0563-8
Zhao, G., Shao, G., Reynolds, K. M., Wimberly, M. C., Warner, T., Moser, J. W., Rennolls, K., et al. 2005:
Digital forestry: A white paper. Journal of Forestry 103-1.
20
62-1_acta49-1.qxd  20.6.2022  10:43  Page 20