Environ Monit Assess (2022) 194:846

https://doi.org/10.1007/s10661-022-10509-7

Land cover change detection in the Turkish coastal zone

based on 28‑year (1990–2018) Corine data
Tuncay Kuleli · Seyma Bayazıt

Received: 21 April 2022 / Accepted: 16 September 2022 / Published online: 1 October 2022
© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2022

Abstract In this research, land cover changes in 4.5% in 2034 and to 5.0% in 2050. It is seen that more
the coastal region of Turkey were analyzed using the artificial surfaces will be needed in the Turkish coastal
Corine dataset between 1990 and 2018. Changes in region due to the increasing population and number of
each period were analyzed by using the rate of change tourists. It is important to evaluate and investigate the
and the annual rate of change, the transition matrix coastal areas where more artificial areas are expected
of net changes, and the transition probability matrix. to be needed within the scope of the coastal area man-
In order to predict land cover change trends and to agement plans to be prepared at the national scale.
determine future land cover change probabilities, a
combination of Markov and cellular automata models Keywords Corine · Land use · Land cover · Coastal
was used. It was determined that the highest increase zone · Turkey
in each study period was in artificial surfaces and the
highest decrease was in the forest and the semi-natural
area. The total forest areas were converted to the other Introduction
land cover type in the first study period amounting to
2479.32 ­km2. Also, it was decreased and transformed The coastal zone is one of the rarest regions in the
by about 62.47 k­ m2, 118.82 ­km2, and 203.09 k­ m2 of world where complex socio-economic and ecologi-
the forest area that existed in the second, third, and cal relations are in a sensitive balance. The coastal
fourth periods respectively. The results indicate that zone has always been important to humanity which
the probabilities of the increasing area will be covered consists of the intersection of land, air, and marine
by artificial surfaces and agricultural areas. It is esti- environment, and it is also important in terms of hav-
mated that the rate of 57% in 1990, the initial year of ing different types of land cover, natural resources,
the forest areas, will decrease to 53.4% in 2034 and and rich ecosystems (Costanza et al., 1997). Human
to 53% in 2050. Also, the rate of 2.1% in 1990, the demands and uses in coastal regions are varied and
initial year of the artificial surfaces, will increase to consist of many sectors including urbanization, lei-
sure and recreation, tourism, transport, agriculture,
and fishing (Mimura, 2008). Coastal areas have
T. Kuleli (*) · S. Bayazıt always been used intensively for socio-economic ben-
Bodrum Maritime Vocational School, Mugla Sitki
efits including commercial, industrial, and residential
Kocman University, Bodrum, Mugla, Turkey
e-mail: [email protected] (Neumann et al., 2015).
The level of impact of human activities in the coastal
S. Bayazıt
e-mail: [email protected] zone is different, subject to local or regional conditions.

Vol.: (0123456789)
13
 846 Page 2 of 24 Environ Monit Assess (2022) 194:846

The main factors that cause an impact on coastal zones intervals, and the analysis interested the contiguous
are demographic change, socio-economic develop- periods but omits the next process of land cover tran-
ment, and land use/land cover policy. Within these sitions (Liu & Zhu, 2010; Zhou et al., 2008). While
factors, population growth and economic development transition matrix is commonly used for statistical
are considered the greatest possible cause of pressure analysis of land cover changes, the most commonly
on coastal regions (Chilar, 2000; Clark, 1982; Jaiswal used techniques for modelling land cover change
et al., 1999; Joshi et al., 2011; Rawat et al., 2013; Yuan are Markov chain analysis and cellular automata
et al., 2005). Turkey is a good example where the popu- (CA) analysis (Choukiker & Dohare, 2021). The
lation is concentrated in coastal areas. Coastal environ- CA-Markov analysis is one of the most used model-
ments are one of the most dynamic, vulnerable, easy, ling methods among land cover modelling methods,
and fast changeable areas of the Earth’s surface, and, which measures both temporal and spatial changes
therefore, changes in the coastal zone including shore- (Regmi et al., 2014; Weng, 2002). CA-Markov analy-
line/coastline changes, water quality changes, socio- sis integrates the Markov chain and cellular automata
economic changes, and land cover changes should be to forecast the land cover change trends and attributes
closely managed and monitored. Coastal zone vulner- over time (Nouri et al., 2014). In addition, the CA-
ability can increase due to changes in the land cover, Markov method is one of the land cover planning sup-
which conducts to corruption of the coastal and marine port tools for the analysis of the spatial distribution
ecosystems, socio-economic system, and physical infra- and temporal changes of land cover (Hua, 2017). In
structure (Huang et al., 2018). Because the economic order to perform land cover change analyses accord-
advantage acquired from the usage of coastal zone ing to accepted methods, land cover data obtained at
resources and land cover change has caused major prob- regular intervals and platforms where these data will
lems including coastline change, coastal erosion, habitat be processed are required. Geographical information
loss, population growth, artificialization, and pollution systems (GIS) and remote sensing techniques which
(Tian et al., 2018; Zhu et al., 2012). The present conse- have an important role in providing useful data in
quences in the issue of social, economic, environmen- spatial and temporal analyses have been extensively
tal, and climatic have made the management of coastal used as powerful tools to analyze the characteris-
land cover changes a basic issue, and also land cover tics and trends of land cover changes and to provide
changes have been important in management and plan- detailed information on land cover (Campbell &
ning activities, particularly for preserving the coastal Randolph, 2011; Wang et al., 2011; Han et al., 2009;
zone and sectoral planning (Niculescu & Lam, 2019). Deng et al., 2009; Sidle et al., 2006; Longley et al.,
Land cover changes can be determined by statis- 2005; Fromard et al., 2004 and Antrop, 2004).
tical assessment by comparing temporal or different When the literature on land cover change is exam-
dated land cover maps. There are several major land ined, it is seen that the change analyses are gener-
cover change detection methods including image dif- ally made by the transition matrix, Markov, and CA-
ferencing, image ratioing, post-classification com- Markov methods and by processing the data obtained
parison, artificial neural networks, image fusion, and from the remote sensing technique in GIS software
decision tree (Choukiker & Dohare, 2021). Besides (Choukiker & Dohare, 2021). In addition to the
these, one of the change analysis methods the transi- remote sensing data, the Coordination of Informa-
tion matrix is a widely used technique, which gives tion on the Environment (Corine) land cover data-
a cross-tabulation matrix including change direc- set in five periods covering the years 1990 to 2018
tions and quantities and allows the determination is also an important data source that can be used for
of differences between systematic and random land land cover change analysis. For example, Colea et al.
cover transitions (Isabel et al., 2011; Lu et al., 2004). (2018) have researched how the UK landscape has
Transition cross-tabulation matrix-based land cover changed at the scale of the Corine land cover minimum
changes analyses mainly determine total losses and mapping unit between 2006 and 2012; Feranec et al.
gains without the geographic locations and trans- (2002) conducted a landscape change analysis of Slo-
formative changes of land cover transitions (Johnson vakia using the Corine database methodology; Feranec
& Zuleta, 2013; Li et al., 2009). The maps or tables et al. (2010) determined land cover changes in Euro-
of the analyses are tested or ordered at separate pean landscapes between 1990 and 2000 using Corine

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 3 of 24 846

data; Gallego and Bamps (2008) used the Corine data- similar or different reasons. Determining where the land
set for land cover area estimation in a research. Even, cover changes are the most in the entire Turkish coastal
the “Corine land cover 2000 coastline” dataset has region and determining the direction of the land cover
been created using the Corine land cover 2000 class, change are the new results of this research. Finally, the
which contains complete coastline features for all the results of land cover change trend and analysis have been
European countries that have a coastline, with detailed presented as maps, graphs, and tables for evaluation and
descriptions of the environment and type of coastal discussion.
zones (URL-1, 2021). However, this data set does not
cover the Turkish coastal region.
In a region like Turkey which has more than 8333 km
coastline, it is very difficult to obtain remote sensing Research area
data for monitoring and analyzing land cover changes
in the entire coastal region. However, the Corine data- Turkish coastal zone as a research area has 8333 km
set is quite useful for this analysis. Turkey is one of the of coastline with 20.34% in the Black Sea, 11.20% in
countries with a very long and different type of coastal Marmara, 33.66% in the Aegean Sea, and 20.07% in
zone among the countries with a coast on the Mediterra- the Mediterranean (Kuleli, 2015). The total land area
nean. Therefore, land cover in the coastal zone may differ is 783,562 ­km2 (GCoM (General Command of Map-
from region to region. Corine data set for the years 1990, ping), 2021). These four coastal regions show distinct
2000, 2006, 2012, and 2018 covers all of Turkey includ- geographical features. Mountain ranges run parallel
ing the coastal zone. In this research, whether there is a to the coast along the Black Sea coast, especially in
land cover change in the coastal region of Turkey, and the eastern part, limiting the size of the coastal area
if so, in which direction this land cover change was ana- to extreme minimums on one hand and bringing a
lyzed using the Corine dataset from these five different marked influence on the climatology of the region on
dates. Also, land cover changes in each period were ini- the other. The western Mediterranean coast has geo-
tially analyzed by using the rate of change and the annual logical features similar to the eastern Black Sea coast,
rate of change, the transition matrix of net changes, and e.g., high mountain ranges running in close proxim-
the transition probability matrix. In order to predict land ity to the shoreline. However, the climatology of this
cover change trends over a given period, and to determine coast, which is basically the Mediterranean, is far dif-
future land cover change probabilities, a combination ferent. Along the eastern Black Sea and the western
of Markov and cellular automata (CA-Markov) mod- Mediterranean coast, the width of the coastal area is
els were used. A transition matrix of net changes (area very narrow (in the order of a few hundred meters),
as ­km2) and transition probability matrix was created thus rendering the area unsuitable for many coastal
using the Markov model, and transition probability was uses including urbanization. According to the 2020
calculated using the CA–Markov model. As stated in the census, the total population of Turkey is 83,614,362
“Discussion” section, although the previous research on persons (TurkStat (Turkish Statistical Institute),
land cover change analyses was conducted locally on the 2021). Turkey has 81 administrative district-city of
Turkish coasts, in this study, a first, new, and nearly up- which 28 have a coastline, and each city has several
to-date land cover change analysis was carried out cover- sub-districts. In this study, coastal regions are defined
ing all Turkish coasts and using the specified methods. at the sub-districts level with a coastline, which cor-
There are some local and regional studies, but there is responds to approximately 98,993 ­km2 and 12.6% of
no study in which land cover trend analysis is performed the total country land (Fig. 1).
for the entire Turkish coastal region. In this study, land
cover trend analysis using transition matrix, CA–Markov
model, and Corine data was performed for the first time Materials and methods
for the entire Turkish coastal region, thus contributing
to filling this gap in the literature. In addition, by con- In this research, the Coordination of Information on
ducting land cover trend analysis for the entire Turkish the Environment (Corine) data set and vector type of
coastal region, it could be revealed whether the land coastal zone boundary data were used to determine
cover changes are local, regional, or country-wide due to five-period land cover changes between 1990 and

Vol.: (0123456789)
13
 846 Page 4 of 24 Environ Monit Assess (2022) 194:846

Fig. 1  Research area. Turkish coastal zone map (Kuleli, 2015). The coastal districts shown in dark gray were accepted as coastal
zone

2018 in the Turkish coastal areas. The rate of change Corine Land Cover 2018 (CLC2018) data Version
in each period was determined according to the 2020_20u1 (release date, 2018)
Corine dataset periods.
Evolution of Corine land cover data set (EC, 1994;
Büttner & Kosztra, 2017; EEA, 2007)
Corine 100 m land cover dataset from the European
Environment Agency (EEA) Geometric accuracy satellite images: 1990 ≤ 50 m;
2000, 2006, 2012 ≤ 25 m; 2018 ≤ 10 m.
The Corine program was initiated in the European Geometric accuracy Corine data: 1990 100 m;
Union in 1985 as a prototype project working on 2000, 2006, 2012, 2018 better than 100 m.
many different environmental issues. The Corine Thematic accuracy ≥ 85%
database is an inventory of land cover in 5 base Change mapping boundary displacement: mini-
classes, 15 subclasses, and 44 classes, and presented mum 100 m; all changes > 5 ha must be mapped.
at a resolution of 100 m (EEA (European Environ- CLC mapping minimum mapping unit: 25 ha for
mental Agency), 2007). This database is freely avail- all periods.
able over the internet for most European countries CLC mapping minimum width: 100 m for all
and Turkey (https://​land.​coper​nicus.​eu/​pan-​europ​ean/​ periods.
corine-​land-​cover). In this research, the following The CLC nomenclature has three levels: the first
Corine dataset was used: level indicates the major categories of land cover on
Corine Land Cover 1990 (CLC1990) data Version the planet; the second level is for use on a scale of
17 (12/2013, last update 08 Apr. 2014), 1:500,000 and 1:1,000,000; the third level is for the
Corine Land Cover 2000 (CLC2000) data Version project on a scale of 1:100,000 (Table 1) (EC (Euro-
17 (12/2013, last update 08 Apr. 2014), pean Commission), 1994). In this study, the first level
Corine Land Cover 2006 (CLC2006) data Version of CLC was used.
17 (12/2013, last update 08 Apr. 2014), A shapefile containing the complete administrative
Corine Land Cover 2012 (CLC2012) data Version map of Turkey obtained from Global Administrative
18.5.1 (release date, 19 Sep. 2016), and Areas in vector format was used to extract the coastal

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 5 of 24 846

region from the CLC raster dataset (GAA (Global Losses-gain-net changes of land cover in the
Administrative Areas), 2021). research area were analyzed by using Eqs. (3) and
The workflow chart of the methodology is shown (4) and the transition matrix (Zhu, 2010, as cited in
in Fig. 2. Xu et al., 2019) (Table 2). The transition matrix is a
Vector and raster data need to be optimized for widely used technique for determining land cover
analysis. For data optimization, the transforma- changes. It is used to analyze the quantitative attributes
tions of spatial coordinates, preparing vector type and direction of land cover changes between two dif-
of coastal zone mask map, and CLC raster data set ferent time periods (Adedeji et al., 2015; Areendran
clipping methods were used. Projections and trans- et al., 2013; Falahatkar et al., 2011; Hegazy & Kaloop,
formations; CLC raster dataset and vector mask map 2015; Kumar et al., 2014; Müller-Hansen et al., 2017;
must have the same spatial reference for clipping the Petit et al., 2001; Pontius et al., 2004; Rawat & Kumar,
coastal region as a study area. Therefore, the CLC 2015; Tang et al., 2007; Vázquez-Quintero et al., 2016;
raster dataset was transformed from Corine geospa- Wijanarto, 2006).
tial coordinate to Country coordinate by using GIS Rows in the transition matrix (Table 2) describe the
software. Preparing vector type of coastal zone mask land cover type at the time t, and columns describe
map; coastal districts were extracted from a shape- the land cover type at the time t+1. Pij describes the
file containing the complete districts map of Turkey. region transformed from land cover type i to land
Clipping CLC raster data; according to the coastal cover j during t–t+1 and S is the total research area
zone mask map, the CLC dataset was clipped to the (Xu et al., 2019).
extent of the study area and stored in geodatabase
format. Thus, the Corine dataset containing only the
ΔPij = Pij − Pji (3)
study area was prepared for change analysis, and the
presence of land cover in each period was prepared as | |
|ΔPij |
a table. 𝜀Pij = | | (4)
|ΔP |+|ΔP |+ ⋯ +|ΔP |
Land cover changes in each period were ini- | i1 i2 in |
tially analyzed in three ways. These are the rate of
where ΔPij is the difference between the land cover
change in each period, the annual rate of change in
area converted from land cover -type i to land cover
each period, and losses-gain-net changes. The rate
-type j and the area converted from land cover -type j
of change in each period was determined according
to land cover -type i during t–t+1; when ΔPij > 0, the
to the Corine dataset periods. The rate of change in
final change is from land cover -type i to land cover
each period as a percentage was calculated according
-type j, and the amount of conversion is ΔPij; when
to Eq. (1) (Barasa, 2011; Peng, 2008)
ΔPij < 0, the final change is from land cover -type j
At+1 − At into land cover -type i, and the amount of conversion
CRi = ( ) × 100 (1) is ΔPij; εPij is the proportion of conversion between
At
land cover -type i and land cover -type j to the total
where CR is the rate of change, i is the period, t is change of land cover -type i or j (Xu et al., 2019).
the first year of the period, t+1 is the second year of In order to predict land cover change trends over
the period, and A is the area of the land cover class a given period, and to determine future land cover
as ­km2. change probabilities, Markov and cellular automata
The annual rate of change in each period as a per- (CA-Markov) models were used. A transition matrix
centage was calculated according to Eq. (2) (Aldwaik of net changes (area as ­ km2) and transition prob-
& Pontius, 2012; Liu et al., 2002) ability matrix were created using the Markov model,
and transition probability was calculated using the
At t 1 −t
CAi = ( ) t+1 − 1) × 100 (2) CA–Markov model.
At+1 The Markov model forecasts the future condi-
where CA is the annual rate of change, i is the period, tion of land cover changes starting from period one
t is the first year, t+1 is the second year, and A is the to period two (Logsdon et al., 1996). This transition
area of the land cover class as ­km2. from one condition to another is represented by the

Vol.: (0123456789)
13
 846 Page 6 of 24 Environ Monit Assess (2022) 194:846

Table 1  Corine land cover nomenclature. Nomenclature consists of three levels. There are 5 different land cover types at the first
level, 15 at the second level, and 44 different land cover types at the third level
Level 1 Level 2 Level 3

1. Artificial surfaces 1.1. Urban fabric Continuous urban fabric

Discontinuous urban fabric
1.2. Industrial, commercial, and transport units Industrial or commercial units
Road and rail networks and associated land
Port areas
Airports
1.3. Mine, dump, and construction sites Mineral extraction sites
Dump sites
Construction sites
1.4. Artificial, non-agricultural vegetated areas Green urban areas
Sport and leisure facilities
2. Agricultural areas 2.1. Arable land Non-irrigated arable land
Permanently irrigated land
Rice fields
2.2. Permanent crops Vineyards
Fruit trees and berry plantations
Olive groves
2.3. Pastures Pastures
2.4. Heterogeneous agricultural areas Annual crops associated with permanent crops
Complex cultivation patterns
Land principally occupied by agriculture, with
significant areas of natural vegetation
Agro-forestry areas
3. Forest and semi-natural areas 3.1. Forests Broad-leaved forest
Coniferous forest
Mixed forest
3.2. Scrub and/or herbaceous vegetation associa- Natural grasslands
tions
Moors and heathland
Sclerophyllous vegetation
Transitional woodland-shrub
3.3. Open spaces with little or no vegetation Beaches, dunes, sands
Bare rocks
Sparsely vegetated areas
Burnt areas
Glaciers and perpetual snow
4. Wetlands 4.1. Inland wetlands Inland marshes
Peat bogs
4.2. Maritime wetlands Salt marshes
Salines
Intertidal flats
5. Water bodies 5.1. Inland waters Water courses
Water bodies
5.2. Marine waters Coastal lagoons
Estuaries

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 7 of 24 846

Table 1  (continued)
Level 1 Level 2 Level 3

Sea and ocean

No data No data No data
Unclassified Unclassified land surface Unclassified land surface
Unclassified Unclassified water bodies Unclassified water bodies
Unclassified Unclassified Unclassified

transition probability described below (Adegbola, where S is the group of finite cell conditions, N is
2021): the number of neighborhood cells, t and tt+1 are ear-
lier and later times, and f is the transformation rule of
⎡ P11 P12 ⋯ P1n ⎤
land cover area.
⎢P P22 … P2n ⎥
P = Pij = ⎢ 21 (5) Using CLC datasets for two different time peri-
⋯ ⎥⎥
⎢ ⋯ ods (1990 and 2000, 2000 and 2006, 2006 and 2012,
⋯ ⋯
⎣ Pn1 Pn2 Pnn ⎦
2012 and 2018, 1990 and 2018) for the research area,
⋯

where P is the probability from one condition to the Markov model was used to analyze, simulate, and
another condition (condition i to j) (Sathees et al., predict land cover types. After the model implemen-
2014; Guan et al., 2008; Jianping et al., 2005, as cited tation, the transition area matrix, transition probabil-
in Adegbola, 2021). In this case, Eq. (5) should fulfil ity matrix, and a set of conditional probability maps
the following two conditions:: could be obtained. CA–Markov model would use the
transition areas table/matrix and the conditional prob-
∑
n
ability maps to forecast land cover change over the
Pij = 1 0 ≤ Pij ≤ 1 (6) given land cover period in the Markov model analy-
j=1
sis. The model is finalized using a 5 by 5 contiguity
Two major steps in the Markov model are acquir- filter to create a spatially clear proximity-weighting
ing a primary matrix and a transition probability factor to change the condition of cells based on their
matrix (Pij). Therefore, the Markov prediction model neighbors. Finally, tables, graphics, and maps were
is described as expressed in Eq. (7). prepared by the analysis results.

Pn = P(n−1) Pij = P(0) Pnij (7)

Results
where Pn is the state probability and P(0) is the pri-
mary matrix (Sathees et al., 2014; Guan et al., 2008;
In this study, land cover change was determined from
Jianping et al., 2005, as cited in Adegbola et al.,
a comparison of CLC data for 1990, 2000, 2006,
2021).
2012, and 2018. The total coastal land cover area was
CA is a discrete and spatial land cover change
98,993.98 ­km2. CLC Level 1 (five classes) land cover
simulation model that describes the dynamics of
types were extracted in the study area for these peri-
land cover changes over time by using transition
ods. According to Level 1, at each period, forest and
rules, cells, cell size, cell neighborhoods, and time.
semi-natural areas were the dominant land cover type
(Memarian et al., 2012; Santé et al., 2010; Guan
between 54 and 57% of the study area and followed
et al., 2011; Yang & Li, 2007, as cited in Munthali
by agricultural areas (39–40%), artificial surfaces
et al., 2020). In general, the CA model is described in
(2–4%), water bodies (1%), and wetlands (0.46%)
the following equation (Liping et al., 2018; Mondal
(Figs. 3 and 4).
et al., 2016; Singh et al., 2015, as cited in Munthali
When the net land cover changes were evaluated
et al., 2020)
between 1990 and 2000, 2000 and 2006, 2006 and
( )
S (t, t + 1) = f ( St , N) (8) 2012, and 2012 and 2018, it was found that the high-
est increase in each period was in artificial surfaces

Vol.: (0123456789)
13
 846 Page 8 of 24 Environ Monit Assess (2022) 194:846

Vol:. (1234567890)
13
 Environ Monit Assess (2022) 194:846 Page 9 of 24 846

◂Fig. 2  The workflow chart of the methodology Aegean regions, while a significant increase was not
observed in the Black Sea coasts (Fig. 6d). The net
changes in water bodies increased the most in the
and the highest decrease was in the forest and the Mediterranean, Aegean, and Marmara regions, while
semi-natural area (Table 3). The results presented in a significant increase was observed on the Black Sea
Table 4 for each period show that the artificial sur- coasts (Fig. 6e). The geographical distribution of net
faces have increased net by 1237.14 ­m2 in the first land cover changes in the coastal zone. The highest
period, 114.43 ­m2 in the second period, 198.10 in the increase in artificial surfaces and the highest decrease
third period, 331.54 in the fourth period, and 1881.21 in forest and semi-natural areas are in the Marmara,
in the overall period (between 1990 and 2018). The Aegean, and Mediterranean coasts (a, c). The highest
change results revealed a considerable reduction of decrease in agricultural areas and wetlands is in the
the forest and the semi-natural area for each period. Aegean and Mediterranean coasts (b, d). In Fig. 6f,
The total forest and the semi-natural area converted in gain and losses were shown on a real scale.
the first study period amounts to 2479.32 ­km2. Since land cover changes in all Turkish coasts
Also, it was decreased and transformed by about were analyzed in this study, the study area covers a
62.47 ­km2, 118.82 ­km2, and 203.09 ­km2 of the forest very large region. It was necessary to zoom in very
and semi-natural area in the second, third, and fourth closely to see the areas with significant changes on
periods, respectively. The highest rate of artificial sur- the map. Therefore, the details of changes in major
faces and agricultural areas increase was found dur- coastal cities under the pressure of artificialization
ing the first period of change (1990 to 2000). How- in the Aegean and Mediterranean regions are shown
ever, the highest rate of forest and semi-natural areas in Fig. 7. Significant changes are clearly shown in
decrease was also observed during the first period of Fig. 7 which presented Izmir coastal city in Aegean
change (1990 to 2000). Region (A), Kuşadası coastal settlement (B), Bod-
The results of the rate of changes and the annual rum tourism center (C), Marmaris and Fethiye
rate of changes are presented in Fig. 5 for each tourism centers (D), Antalya coastal province and
period. Between 1990, the first date of the analy- tourism center (E), and Mersin coastal and port
sis, and 2018, the last date of the analysis, the rate city (F) in Mediterranean Region. It is understood
of change in artificial surfaces increased by 89.92%. that the direction of change in important coastal
During the same period, water bodies increased by cities in Fig. 7, whose changes are closely exam-
11.56%, wetlands by 8.61%, and agricultural areas by ined, increases in artificial surfaces and decreases
2.11%, while forest and semi-natural areas decreased in forest and semi-natural areas, and agricultural
by − 5.08%. When the four periods are compared, areas. Mostly changed coastal cities sample maps.
the highest increase in artificial surfaces, agricultural The details of changes in major coastal cities under
areas, wetlands, and water bodies and the highest the pressure of artificialization in the Aegean and
decrease in forest and semi-natural areas were seen Mediterranean regions. İzmir coastal city in Aegean
in the first period. The same situation was observed Region (A), Kuşadası coastal settlement (B), Bod-
for the annual rate of change. According to the find- rum tourism center (C), Marmaris and Fethiye tour-
ings in Table 3, the geographical distribution of net ism centers (D), Antalya coastal province and tour-
land cover changes in the coastal zone is shown in ism center (E), and Mersin coastal and port city
Fig. 6. As can be seen in Fig. 6a, the highest increase (F) in Mediterranean Region. A transition matrix
in net changes of artificial surfaces between 1990 analysis was carried out to determine which land
and 2018 was detected in the Marmara, Aegean, use type transformed into another (Table 4). The
and Mediterranean regions. However, the highest transition matrix table which shows the land cover
losses in forest and semi-natural areas and agricul- transformation lists land cover classes of earlier
tural areas were also detected in these regions and the data in the columns and the last data in the rows.
same period (Fig. 6c). On the other hand, the high- Based on this matrix, it is found that mainly large
est increase in net changes in agricultural areas was forest and semi-natural areas have been transformed
in the Black Sea region (Fig. 6b) The net changes in into agricultural areas, and then, agricultural areas
wetlands highest increased in the Mediterranean and have been transformed into artificial surfaces. In the

Vol.: (0123456789)
13
 846 Page 10 of 24 Environ Monit Assess (2022) 194:846

Table 2  A sample of land t+1 A1 A2 … An Total

cover transition matrix (Xu
et al., 2019) t

A1 P11 P21 … Pn1 P11+P21+…+Pn1

A2 P12 P22 … Pn2 P11+P22+…+Pn2
An P1n P2n … Pnn P1n+P2n+…+Pnn
Total P11+P12+…+P1n P21+P22+…+P2n … Pn1+Pn2+…+Pnn S

first period, 2148.76 k­ m2 of forest and semi-natural land cover transformation direction is from the for-
areas were transformed into agricultural areas. In est and semi-natural areas to agricultural areas and
the same period, 960.13 ­km2 of agricultural areas from agricultural areas to artificial surfaces. Although
have been transformed into artificial surfaces. In a decrease was detected in the amount of chang-
Fig. 8, it can be clearly seen that forest and semi- ing field in the second, third and fourth periods, the
natural areas firstly transform into agricultural direction of land cover change was found to be the
areas, and agricultural areas transform into artificial same as in the first period. The transition probability
surfaces. matrix which presented the replacement of one land
In order to visually show which land use type has cover with another land cover is given in Table 5. The
transformed into another, especially the transitions transition probability matrix measures the percent of
between artificial surfaces, forest and semi-natural land cover area change in a given year and provides
areas, and agricultural areas are given visually in a transition matrix showing the proportions of pix-
Fig. 8. According to Table 5, as seen in Fig. 8, the els changing from one land cover to another and the

Fig. 3  Land cover types for the years 1990, 2000, 2006, 2012, and 2018 as ­km2. In each period, the land cover with the most area is
forest and semi-natural areas, agricultural areas, and artificial surfaces, respectively

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 11 of 24 846

Fig. 4  Land cover maps for the years 1990, 2000, 2006, 2012, and 2018

matrix gives the area change table which presents the artificial surfaces and agricultural areas. Figure 9
change in the land cover from 1990 to 2018 in all 5 shows the change in each land cover. It is estimated
classes. According to Table 5, the highest probabili- that the rate of 57% in 1990, the initial year of the for-
ties of change in agricultural areas (0.0981) were con- est and semi-natural areas, will decrease to 53.4% in
verted to artificial surfaces, while the lowest probabil- 2034 and to 53% in 2050. In addition to that, it is esti-
ities were in the water bodies (0.0006). The forest and mated that the rate of 2.1% in 1990, the initial year of
semi-natural areas have been replaced by other land the artificial surfaces, will increase to 4.5% in 2034
covers such as agricultural areas (0.0387) and artifi- and to 5.0% in 2050. It is estimated that there will not
cial surfaces (0.0289). be a significant change in agricultural areas (39.4% in
In order to predict likely future trends and changes 1990, 40.4% in 2034, and 40.3% in 2050).
in the study area, cellular automata, based on the As a result, when the net CLC changes were evalu-
change in land cover between 2018 and 2034, and ated among the periods, it was determined that the
between 2018 and 2050, was used in transitional highest increase in each period was in artificial sur-
potential Ca_Markov modelling to forecast future faces and the highest decrease was in the forest and
changes (Fig. 9). The result indicates that the prob- the semi-natural area. The highest increase in artifi-
abilities of the increasing area will be covered by cial surfaces and the highest decrease in forest and

Vol.: (0123456789)
13
 846 Page 12 of 24 Environ Monit Assess (2022) 194:846

Table 3  Losses, gain, and Land cover type Losses Gain Net change
net changes of land cover
types between 1990 and Between 1990 and 2000
2000; 2000 and 2006; 2006
Artificial surfaces (ArS) − 631.14 1868.28 1237.14
and 2012; 2012 and 2018;
1990 and 2018 as ­km2 Agricultural areas (AgA) − 12,621.32 13,742.63 1121.31
Forest and semi-natural areas (FSA) − 13,734.07 11,254.75 − 2479.32
Wetlands (WL) − 132.16 178.86 46.70
Water bodies (WB) − 138.65 212.82 74.17
Total − 27,257.34 27,257.34 0.00
Between 2000 and 2006
Artificial surfaces (ArS) − 38.99 153.42 114.43
Agricultural areas (AgA) − 104.86 62.30 − 42.56
Forest and semi-natural areas (FSA) − 1079.81 1017.34 − 62.47
Wetlands (WL) − 7.34 0.31 − 7.03
Water bodies (WB) − 10.82 8.45 − 2.37
Total − 1241.82 1241.82 0.00
Between 2006 and 2012
Artificial surfaces (ArS) − 116.21 314.31 198.10
Agricultural areas (AgA) − 320.56 212.20 − 108.36
Forest and semi-natural areas (FSA) − 1236.04 1117.22 − 118.82
Wetlands (WL) − 1.63 1.37 − 0.26
Water bodies (WB) − 6.51 35.85 29.34
Total − 1680.95 1680.95 0.00
Between 2012 and 2018
Artificial surfaces (ArS) − 91.08 422.62 331.54
Agricultural areas (AgA) − 257.41 107.62 − 149.79
Forest and semi-natural areas (FSA) − 728.40 525.31 − 203.09
Wetlands (WL) − 2.09 0.00 − 2.09
Water bodies (WB) − 10.00 33.43 23.43
Total − 1088.98 1088.98 0.00
Between 1990 and 2018
Artificial surfaces (ArS) − 628.59 2509.80 1881.21
Agricultural areas (AgA) − 12,968.33 13,788.93 820.60
Forest and semi-natural areas (FSA) − 15,110.86 12,247.16 − 2863.70
Wetlands (WL) − 138.18 175.50 37.32
Water bodies (WB) − 158.03 282.60 124.57
Total -29,003.99 29,003.99 0.00

semi-natural areas were on the Marmara, Aegean, the forest and semi-natural areas, (b) agricultural
and Mediterranean coasts, while the highest decrease areas were gained firstly from the forest and semi-nat-
in agricultural areas and wetlands was found on the ural areas, (c) forest and semi-natural areas were lost
Aegean and Mediterranean coasts. It is seen that firstly to agricultural areas and secondly to artificial
CLC changes were relatively least in the Black Sea surfaces, (d) wetlands were gained the most firstly
coastal zone. The change results revealed a consider- from the forest and semi-natural areas and secondly
able reduction of the forest and the semi-natural area from agricultural areas, (e) water bodies were gained
for each study period. Transition matrix analysis was the most firstly from agricultural areas and secondly
showed that (a) artificial surfaces were gained the from the forest and semi-natural areas, and were lost
most firstly from agricultural areas and secondly from to artificial surfaces.

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 13 of 24 846

Table 4  Transition (area) matrix in k­ m2 from 1990 to 2018. According to the Corine Level l, “Contributions to Net Change in:”
between 1990 and 2000; 2000 and 2006; 2006 and 2012; 2012 and 2018; 1990 and 2018
Land cover type Artificial surfaces Agricultural areas Forest and Wetlands Water bodies
semi-natural
areas

Between 1990 and 2000 LC at t + 1 (2000) contributions to net change in:

Artificial surfaces 0 − 960.13 − 257.06 − 0.18 − 19.77
(ArS)
Agricultural areas 960.13 0 − 2148.76 3.27 64.05
(AgA)
LC at Forest and semi-natu- 257.06 2148.76 0 42.14 31.36
ral areas (FSA)
t (1990) Wetlands (WL) 0.18 − 3.27 − 42.14 0 − 1.47
Water bodies (WB) 19.77 − 64.05 − 31.36 1.47 0
Sub-total 1237.14 1121.31 − 2479.32 46.7 74.17
Between 2000 and 2006 LC at t + 1 (2006) contributions to net change in:
Artificial surfaces 0 − 58.92 − 48.48 0 − 7.03
(ArS)
Agricultural areas 58.92 0 − 14.18 − 5.41 3.23
(AgA)
LC at Forest and semi-natu- 48.48 14.18 0 − 1.45 1.26
ral areas (FSA)
t (2000) Wetlands (WL) 0 5.41 1.45 0 0.17
Water bodies (WB) 7.03 − 3.23 − 1.26 − 0.17 0
Sub-total 114.43 − 42.56 − 62.47 − 7.03 − 2.37
Between 2006 and 2012 LC at t + 1 (2012) contributions to net change in:
Artificial surfaces 0 − 114.31 − 83.11 − 1.1 0.42
(ArS)
Agricultural areas 114.31 0 − 23.13 0.4 16.78
(AgA)
LC at Forest and semi-natu- 83.11 23.13 0 0.24 12.34
ral areas (FSA)
t (2006) Wetlands (WL) 1.1 − 0.4 − 0.24 0 − 0.2
Water bodies (WB) − 0.42 − 16.78 − 12.34 0.2 0
Sub-total 198.1 − 108.36 − 118.82 − 0.26 29.34
Between 2012 and 2018 LC at t + 1 (2018) contributions to net change in:
Artificial surfaces 0 − 162.32 − 160.33 − 0.98 − 7.91
(ArS)
Agricultural areas 162.32 0 − 37.61 − 0.74 25.82
(AgA)
LC at Forest and semi-natu- 160.33 37.61 0 0 5.15
ral areas (FSA)
t (2012) Wetlands (WL) 0.98 0.74 0 0 0.37
Water bodies (WB) 7.91 − 25.82 − 5.15 − 0.37 0
Sub-total 331.54 − 149.79 − 203.09 − 2.09 23.43
Between 1990 and 2018 LC at t + 1 (2018) contributions to net change in:
Artificial surfaces 0 − 1285.04 − 555.72 − 1.38 − 39.07
(ArS)
Agricultural areas 1285.04 0 − 2209.03 − 2.9 106.29
(AgA)

Vol.: (0123456789)
13
 846 Page 14 of 24 Environ Monit Assess (2022) 194:846

Table 4  (continued)
Land cover type Artificial surfaces Agricultural areas Forest and Wetlands Water bodies
semi-natural
areas

LC at Forest and semi-natu- 555.72 2209.03 0 40.6 58.35

ral areas (FSA)
t (1990) Wetlands (WL) 1.38 2.9 − 40.6 0 −1
Water bodies (WB) 39.07 − 106.29 − 58.35 1 0
Total 1881.21 820.6 − 2863.70 37.32 124.57

Discussion and to decrease for forest areas in the Mediterranean

and Aegean Regions. According to the Corine data,
The purpose of the research and analysis in this study over the past 28 years, from 1990 to 2018, it is esti-
has been to determine whether there is any change in mated that the highest increase in each period was in
land cover in the research area for a 28-year period. artificial surfaces and the highest decrease was in the
Land cover changes in the coastal region of Turkey forest and the semi-natural area.
were analyzed using the Corine dataset from these There may be many reasons for the decrease in
five different dates. Also, land cover changes in each forest areas and agricultural areas and the increase in
period were initially analyzed by using the rate of artificial areas, e.g., population growth, rapid urbani-
change and the annual rate of change, the transition zation, forest fires, floods, desertification, infrastruc-
matrix of net changes, and the transition probability ture needs of coastal-dependent sectors, facilities
matrix. In order to predict land cover change trends required for tourism activities, land use policies, land
over a given period, and to determine future land planning policies, legal regulations, and various natu-
cover change probabilities, a combination of Markov ral and man-made causes. However, in this study, due
and cellular automata (CA-Markov) models was used. to limitations of the available data about those pos-
A transition matrix of net changes (area as ­km2) and sible reasons, the relationship between the decrease in
transition probability matrix were created using the forest areas, decrease in agricultural areas, increase in
Markov model, and transition probability was calcu- artificial areas, and population growth-tourism activi-
lated using the CA–Markov model. ties was evaluated.
As a result of the analysis, it has been concluded Although there are several reasons that cause land
that land cover tends to increase for artificial areas cover changes, it is seen that tourism is the most

Fig. 5  The rate of changes and the annual rate of changes. In each period, artificial surfaces increased and forest and semi-natural
areas decreased

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 15 of 24 846

Fig. 6  According to the findings in Table 3, the geographical distribution of net land cover changes in the coastal zone

influential factor affecting the land cover change in golf, water sports, concert area, marinas, ports, and
the coastal zone of Turkey. The policies and incen- diving centers (Ceylan & Yakut, 2021).
tives followed also support this. Aegean and Mediter- Touristic activities and structures concentrated
ranean coasts of Turkey are the regions with the high- in coastal areas at the country level especially in
est tourism activities and these regions have many the Mediterranean and Aegean can be interpreted
artificial touristic infrastructures, such as hotels, as a result of the tourism policies followed from the
resorts, summer houses, transport, socio-cultural, planned period to the present. From the first to the ­8th

Vol.: (0123456789)
13
 846 Page 16 of 24 Environ Monit Assess (2022) 194:846

Fig. 7  The details of changes in major coastal cities under the ris and Fethiye tourism centers (D), Antalya coastal province
pressure of artificialization in the Aegean and Mediterranean and tourism center (E), and Mersin coastal and port city (F) in
regions. Izmir coastal city in Aegean Region (A), Kuşadası Mediterranean Region
coastal settlement (B), Bodrum tourism center (C), Marma-

Development Plans, touristic capacity increase and the results of this study, the Black Sea coastal zone
mass tourism strategies were followed and invest- has been less used for tourism development thus is
ment for accommodation facilities, transportation less spoiled when comparing the Aegean and Medi-
infrastructures, marinas, yacht ports, and cruise ports terranean coasts. It is thought that the biggest reason
seemed to be encouraged especially in the Mediter- for the increase in artificial areas is the increase in the
ranean and Aegean coastline in Turkey (Strategy and population and the number of tourists. The CLC type
Budget Presidency (SBP), 2022). Ceylan and Yakut that covers the most area in all periods of the study
(2021)’s study also emphasized that touristic facilities was forest and semi-natural areas (FSA), agricultural
and marine tourism structures are concentrated in the areas (AgA), artificial surfaces (ArS), water bodies
Aegean and Mediterranean regions, especially in the (WB), and wetlands (WL), respectively. The highest
provinces of Antalya, Mugla, Izmir, and Aydın prov- decrease was in the FSA. On the other hand, the high-
inces. Also, the Turkish coastal zone has been a sig- est increase was in ArS, AgA, WB, and WL, respec-
nificant increase in the coastal population in the last tively. The FSA has been the most transformed area in
50 years (Kuleli, 2015). Although the Black Sea is each period. The biggest transformation in FSA was
an important contributor to the regional economy via the first period between 1990 and 2000. These results
fishing and energy industries, tourism is not included were not a surprise because this period (between
in the dominant industries in the littoral countries in 1990 and 2000) is remarkable as it is when tourism
the Black Sea (Bat et al., 2018). Thus, as it is seen in investments were the highest in the South Aegean and

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 17 of 24 846

Fig. 8  Transition (area) matrix for ArS, AgA, and FSA in ­km2 between artificial surfaces, forest and semi-natural areas, and
from 1990 to 2018. In order to visually show which land use agricultural areas are given visually
type has transformed into another, especially the transitions

Mediterranean coast of Turkey. In accordance with İzmir, and Muğla (Ministry of Culture and Tourism
the Tourism Incentive Law, which came into force in (MoCT), 2022). More than 70 percent of about 25
1982, a total of 231 Tourism Center/Culture-Tourism thousand yacht mooring capacities in marinas, 19
Conservation and Development Zones have been out of a total 25 cruise ports are located in the Medi-
declared, 165 of which are located in coastal cities terranean and Aegean coasts of Turkey (Ministry of
and mostly those centers are concentrated in the Med- Transport and Infrastructure (MoTI), 2021). The most
iterranean and Aegean coasts, especially in Antalya, FSA loss and the highest ArS gain occurred in the

Vol.: (0123456789)
13
 846 Page 18 of 24 Environ Monit Assess (2022) 194:846

Table 5  Transition probability matrix from 1990 to 2018. According to the Corine Level l, “Contributions probability to Net
Change in:” between 1990 and 2000; 2000 and 2006; 2006 and 2012; 2012 and 2018; 1990 and 2018
Land cover type Artificial surfaces Agricultural Forest and Wetlands Water
areas semi-natural bodies
areas

Between 1990 and 2000 LC at t + 1 (2000)

Artificial surfaces (ArS) 0.8678 0.0981 0.0289 0.0011 0.004
Agricultural areas (AgA) 0.0215 0.9375 0.0387 0.0006 0.0016
LC at Forest and semi-natural areas 0.0035 0.0509 0.9443 0.0006 0.0006
(FSA)
t (1990) Wetlands (WL) 0.0051 0.0525 0.0193 0.8994 0.0237
Water bodies (WB) 0.0198 0.0224 0.0176 0.0104 0.9298
Between 2000 and 2006 LC at t + 1 (2006)
Artificial surfaces (ArS) 0.9971 0.0007 0.0015 0 0.0007
Agricultural areas (AgA) 0.0015 0.9982 0.0001 0 0.0001
LC at Forest and semi-natural areas 0.001 0.0004 0.9986 0 0
(FSA)
t (2000) Wetlands (WL) 0 0.0119 0.003 0.9847 0.0004
Water bodies (WB) 0.0082 0 0.0011 0 0.9907
Between 2006 and 2012 LC at t + 1 (2012)
Artificial surfaces (ArS) 0.9918 0.0031 0.0031 0 0.0019
Agricultural areas (AgA) 0.0031 0.9962 0.0002 0 0.0004
LC at Forest and semi-natural areas 0.0017 0.0006 0.9974 0 0.0002
(FSA)
t (2006) Wetlands (WL) 0.0023 0.0011 0 0.9966 0
Water bodies (WB) 0.0053 0 0.0001 0.0002 0.9943
Between 2012 and 2018 LC at t + 1 (2018)
Artificial surfaces (ArS) 0.9968 0.0002 0.0029 0 0.0001
Agricultural areas (AgA) 0.0041 0.9951 0.0001 0 0.0007
LC at Forest and semi-natural areas 0.0032 0.0008 0.9959 0 0.0001
(FSA)
t (2012) Wetlands (WL) 0.0021 0.0016 0 0.9956 0.0008
Water bodies (WB) 0.0071 0.0009 0.0004 0 0.9915
Between 1990 and 2018 LC at t + 1 (2018)
Artificial surfaces (ArS) 0.9457 0.0431 0.0092 0.0005 0.0016
Agricultural areas (AgA) 0.0119 0.9716 0.0153 0.0002 0.0009
LC at Forest and semi-natural areas 0.0021 0.0206 0.9767 0.0002 0.0003
(FSA)
t (1990) Wetlands (WL) 0.0013 0.0255 0.0069 0.9554 0.0108
Water bodies (WB) 0.0139 0.0065 0.0065 0.0046 0.9684

Aegean and Mediterranean coasts. Also in the same When the results of previous studies in the Turk-
period, the coastal population increased by 24%, and ish coastal zone and Mediterranean coastal countries
the number of tourists increased by 93%. It has been are compared with the findings and predictions in
concluded that land cover changes tend to increase in this research, it is seen that there are similar results
the direction of artificial areas, decrease in the direc- and trends related to land use/land cover changes. In
tion of the forest, and semi-natural areas in the Turk- the Spanish coastal zone, studies showed that areas
ish coastal zone. close to the coasts transformed from natural areas to

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 19 of 24 846

Fig. 9  Ca_Markov model analysis results. According to the model, it is seen that the increase in artificial surfaces and the decrease
in forest and semi-natural areas will continue

artificial areas because of basically urban and tour- the Italian and Greece coastal zone indicated similar
istic usage (Bellot et al., 2007; Martínez-Fernández trends and predictions for land use and land cover
et al., 2009; Symeonakis et al., 2007). There are changes. For example, in Sardinia (Italy) coastal
some other significant researches related to land areas, Feranec et al. (2010) pointed out issues such
use/land cover change in coastal areas in France, as urbanization, extensive agricultural usage, and
Greece, and Italy. Research on the French coastal deforestation in the period between 1990 and 2000.
areas draws attention to artificialization as a result of According to a research conducted simultaneously in
excessive and unplanned urbanization. According to Rome (Italy) and Athens (Greece), using CLC maps
Robert et al. (2019), artificial areas in Biguglia have continuously increase in built-up areas as a result of
increased annually by 4.7 percent while the increase the consumption of agricultural and forest uses were
in Cote Bleue was found as 7.5 percent. Research on found in coastal areas between the years 1975 and

Vol.: (0123456789)
13
 846 Page 20 of 24 Environ Monit Assess (2022) 194:846

2006 (Salvati et al., 2014). Results of some studies studies show that similar land use/land cover changes are
for the Turkish coastal zone support other research observed in the coastal regions of Mediterranean coun-
about land cover/land use changes in different Medi- tries, including Turkey. Almost all of the artificial areas
terranean coasts that all indicated similar results. on the Aegean and Mediterranean coasts of Turkey con-
Research for Candarli Bay in the Turkish coastal zone sist of urban areas and tourism facilities. Especially after
found a noteworthy increase in urban areas of about 1960, there has been an increasing population migra-
32.22 percent which was respectfully due to build- tion from the inner regions of Turkey to the Aegean and
ing up the secondary house for the summer season Mediterranean coasts. In the 1970s and 1980s, much of
(Kesgin & Nurlu, 2009). Atik et al. (2010) investi- the South Aegean and West Mediterranean regions were
gated land use changes in the Turkish Coastal Zone declared tourism areas. Thus, over time, nearly all dif-
and whether it is related to coastal tourism develop- ferent types of land covers in these coastal areas were
ment. Their study showed that agricultural areas with converted to tourism land use.
816 hectares and natural coastal forest with 457 hec-
tares were degraded due to touristic buildings such
as accommodation and touristic facilities. Another Conclusions
research conducted in the Cesme coastal zone, Tur-
key, for 48 years (1957–2005), showed that natural In this study, we researched, analyzed, and estimated
and agricultural areas decreased, while artificial areas the cause of whether there was any change in land
increased (Hepcan et al., 2011). cover in the Turkish coastal zone for the 28-year
Of course, this research has some limitations, but period. The land cover change trend seen in the Turk-
these limitations did not constitute a major obstacle ish coasts shows that artificial surface has increased
to the achievement of the objectives of this study. The due to tourism and urbanization. It is thought that the
most important limitation of the research is that the biggest reason for the increase in artificial areas is the
research area covers a very large area. Because the increase in the population and the number of tourists.
research area is very large: Coastal areas are generally rich in natural resources
and valuable to many industries. However, the use
– Level 1 (5 classes) has been used instead of Level of land cover diversity in coastal areas by only one
3 (44 classes) or Level 2 (15 classes), which had or two sectors/land use, such as tourism and urban
more details. If Level 3 or Level 2 is to be used, development, may increase socio-economic and eco-
coastal cities under the pressure of tourism and logical vulnerability. Insufficient or incorrect land
urbanization with the most change identified in use policies and implementations not only impose the
this research can be analyzed individually. economic potential of coastal areas on a single sec-
– It was necessary to zoom in a lot to see the areas tor but also endanger the sustainability of coastal-
with significant changes on the map. For this rea- dependent sectors, e.g., fisheries, sponge production,
son, the details of these changes could not be visu- marine transportation, ecosystem services, wetland,
ally evident in the maps that had to be prepared in sandy beaches, public access to coastal areas. To
limited sizes in the article. Nevertheless, it can be reduce these risks, international sustainability stand-
a guiding study to investigate the areas with signif- ards such as the Global Sustainable Council (GSTC),
icant changes detected in this study separately and Green Destinations (GD), and The European Tourism
in more detail (for example, Level 3–44 classes). Indicator System (ETIS) in line with The UN Sus-
– Especially CA-Markov change prediction maps tainable Development Goals can be used. For exam-
took a lot of time to prepare and required a lot of ple, GSTC’s destination management organization,
processing power. Therefore, estimates of change planning regulations, and tourism awareness and edu-
are given as graphs rather than maps. cation standards (GSTC (Global Sustainable Tourism
Council), 2019); GD’s land use and planning stand-
These limitations did not constitute a major obstacle ards (GD (Green Destinations), 2021); ETIS’s land-
to the achievement of the objectives of this study. Anal- scape and biodiversity protection standards (ETIS
ysis results could be given with visual maps, graphics, (European Tourism Indicator System), 2016). It is
and tables. Both the findings of this study and previous recommended that policies and roadmaps that can be

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 21 of 24 846

followed in coastal areas should be created by inter- Campbell, J. B., & Randolph, H. W. (2011). Introduction to
national sustainability standards to eliminate prob- remote sensing (pp. 335–375). The Guilford Press.
Bellot, J., Bonet, A., Pena, J., & Sanchez, J. R. (2007).
lems such as over-urbanization and over-tourism at Human impacts on land cover and water balances in a
the national, regional, and local levels. coastal Mediterranean county. Environmental Man-
agement, 39(3), 412–422. https://​doi.​org/​10.​1007/​
s00267-​005-​0317-9
Data availability Corine dataset from Copernicus Land Moni- Ceylan, M. A., Yakut, M. E. (2021). A geographic view on
toring Service downloadable from https://​land.​coper​nicus.​eu/​ the touristic accommodation type and distribution in
pan-​europ​ean/​corine-​land-​cover Turkey. Turkish Academic Research Review, 6 (2), 697–
Administrative map of Turkey from Global Administrative 724. Retrieved from https://​dergi​park.​org.​tr/​tr/​pub/​tarr/​
Areas downloadable from https://​gadm.​org/​downl​oad_​count​ry_​ issue/​62824/​931416
v3.​html Chilar, J. (2000). Land cover mapping of large areas from
Census and tourist data from Turkey Statistical Institute satellites: Status and research priorities. International
downloadable from https://​www.​tuik.​gov.​tr/​Home/​Index Journal of Remote Sensing, 21(67), 1093–1114. https://​
Surface Area of City and District from General Command doi.​org/​10.​1080/​01431​16002​10092
of Mapping downloadable from https://​www.​harita.​gov.​tr/​urun/​ Choukiker, S.K., Dohare, D. (2021). A literature review
il-​ve-​ilce-​yuzol​cumle​ri/​176 on land use land cover changes detection using remote
sensing and GIS. International Journal for Research in
Declarations Applied Science & Engineering Technology (IJRASET).
Volume 9 Issue III. https://​doi.​org/​10.​22214/​ijras​et.​2021.​
Competing interests The authors declare no competing 33349
interests.The authors have no financial or proprietary interests Clark, D. (1982). Urban geography: An ıntroductory guide
in any material discussed in this article. (p. 231). Croom Helm.
Colea, B., Smithc, G., & Balzter, H. (2018). Acceleration
and fragmentation of Corıne land cover changes in the
United Kingdom from 2006–2012 detected by Coper-
nicus IMAGE2012 satellite data. International Journal
References
of Applied Earth Observation and Geoinformation, 73,
107–122. https://​doi.​org/​10.​1016/j.​jag.​2018.​06.​003
Adedeji, O. H., Tope-Ajayi, O. O., & Abegunde, O. L. (2015). Costanza, R., d’Arge, R., de Groot, R., et al. (1997). The
Assessing and predicting changes in the status of Gambari value of the world’s ecosystem services and natural
Forest Reserve, Nigeria using remote sensing and GIS capital. Nature, 387, 253–260. https://​doi.​org/​10.​1038/​
techniques. Journal of Geographic Information System, 38725​3a0
7(3), 301–318. https://​doi.​org/​10.​4236/​jgis.​2015.​73024 Deng, J. S., Wang, K., Hong, Y., & Qi, J. G. (2009). Spatio-
Adegbola, P. A., Adewumi, J. R., Obiora-Okeke, O. A. (2021). temporal dynamics and evolution of land use change
Application of Markov chain model and ArcGIS in land and landscape pattern in response to rapid urbanization.
use projection of Ala River catchment, Akure, Nigeria. Landscape and Urban Planning, 92, 187–198.
Nigerian Journal of Technological Development, Vol. 18, EC (European Commission). (1994). Corine land cover Tech-
No.1. https://​doi.​org/​10.​4314/​njtd.​v18i1.5 nical guide. Office for Official Publications of European
Aldwaik, S. Z., & Pontius, R. G. (2012). Intensity analysis Communities. ISBN 92–826–2578–8, 120p.
to unify measurements of size and stationarity of land Büttner, G., Kosztra, B. (2017). CLC2018 Technical Guide-
changes by interval, category, and transition. Landscape lines. European Environmental Agency. Retrived July
and Urban Planning, 106, 103–114. 23, 2022, from (https://​land.​coper​nicus.​eu/​user-​corner/​
Antrop, M. (2004). Landscape change and urbanization pro- techn​ical-​libra​ry/​clc20​18tec​hnica​lguid​elines_​final.​pdf)
cess in Europe. Landscape and Urban Planning, 67, 9–26. EEA (European Environmental Agency). (2007). CLC2006
https://​doi.​org/​10.​1016/​S0169-​2046(03)​00026-4 technical guidelines. Technical report 17/2007. Euro-
Areendran, G., Rao, P., Raj, K., Mazumdar, S., & Puri, K. pean Environmental Agency, 66p.
(2013). Land use/land cover change dynamics analysis ETIS (European Tourism Indicator System). (2016). The
in mining areas of Singrauli district in Madhya Pradesh, European Tourism Indicator System (ETIS) toolkit 2016.
India. Tropical Ecology, 54(2), 239–250. Retrived September 15, 2021, from https://​ec.​europa.​eu/​
Atik, M., Altan, T., & Artar, M. (2010). Land use changes docsr​oom/​docum​ents/​21749
in relation to coastal tourism developments in Turkish Falahatkar, S., Soffianian, A. R., Khajeddin, S. J., Ziaee, H. R.,
Mediterranean. Polish Journal of Environmental Stud- & Nadoushan, M. H. (2011). Integration of remote sens-
ies, 19(1), 21–33. ing data and GIS for prediction of land cover map. Inter-
Barasa, B., Majaliwa, J. G. M., Lwasa, S., Obando, J., & national Journal of Geomatıcs and Geosciences, 2(4),
Bamutaze, Y. (2011). Magnitude and transition potential 847–864.
of land-use/cover changes in the trans-boundary river Feranec, J., Šúri, M., Cebecauer, T., Oťaheľ, J. (2002). Meth-
Sio catchment using remote sensing and GIS. Annals of odological aspects of landscape changes detection and
GIS, 17(1), 73–80. https://​doi.​org/​10.​1080/​19475​683.​ analysis in Slovakia applying the CORINE land cover
2011.​558023

Vol.: (0123456789)
13
 846 Page 22 of 24 Environ Monit Assess (2022) 194:846

databases. Geografický časopis, 54(3), 271–288. ISSN Isabel, P., Cunha, M., & Pereira, L. (2011). Remote sensing
0016–7193. based indicators of changes in a mountain rural landscape
Feranec, J., Jaffrain, G., Soukup, T., & Hazeu, G. (2010). of northeast Portugal. Applied Geography, 31, 871–880.
Determining changes and flows in European landscapes Jaiswal, R. K., Saxena, R., & Mukherjee, S. (1999). Application
1990–2000 using CORINE land cover data. Applied of remote sensing technology for land use/land cover change
Geography, 30(1), 19–35. https://​doi.​org/​10.​1016/j.​ analysis. Journal of the Indian Society of Remote Sensing.,
apgeog.​2009.​07.​003 27(2), 123–128. https://​doi.​org/​10.​1007/​BF029​90808
Fromard, F., Vega, C., & Proisy, C. (2004). Half a century of Jianping, L., Bai, Z., & Feng, G. (2005). RS-and-GIS-sup-
dynamic coastal affecting mangrove shoreline of French ported forecast of grassland degradation in southwest
Guiana: A case study based on remote sensing data analy- Songnen plain by Markov model. Geo-Spatial Informa-
ses and field surveys. Marine Geology., 208, 265–280. tion Science, 8, 104–106.
GAA (Global Administrative Areas). (2021). Retrieved Johnson, B. G., Zuleta, G. A. (2013). Land-use land-cover
December 1, 2021, from http://​www.​gadm.​org change and ecosystem loss in the Espinal ecoregion,
Gallego, J., & Bamps, C. (2008). Using Corıne land cover and Argentina. Agriculture Ecosystems & Environment, 181,
the point survey Lucas for area estimation. International 31–40.
Journal of Applied Earth Observation and Geoinforma- Joshi, R. R., Warthe, M., Dwivedi, S., Vijay, R., & Chakrabarty,
tion, 10, 467–475. T. (2011). Monitoring changes in land use land cover of
GCoM (General Command of Mapping). (2021). Surface area Yamuna riverbed in Delhi: A multi temporal analysis. Inter-
of city and district. Retrived September 7, 2021, from national Journal of Remote Sensing., 32(24), 9547–9558.
https://​www.​harita.​gov.​tr/​urun/​il-​ve-​ilce-​yuzol​cumle​ri/​176 https://​doi.​org/​10.​1080/​01431​161.​2011.​565377
GD (Green Destinations). (2021). Green Destinations Standard Kesgin, B., & Nurlu, E. (2009). Land cover changes on the
V2 Recognized by GSTC. Retrived September 3, 2021, coastal zone of Candarli Bay, Turkey using remotely sensed
from https://​green​desti​natio​ns.​org/​wp-​conte​nt/​uploa​ds/​ data. Environmental Monitoring and Assessment, 157(1–
2021/​04/​GD-​Stand​ard-​V2-​2021-​GSTC-​Recog​nised.​pdf 4), 89–96. https://​doi.​org/​10.​1007/​s10661-​008-​0517-x
GSTC (Global Sustainable Tourism Council). (2019). GSTC Kuleli, T. (2015). The socio-economic significance of Turkish
destination criteria with performance indicators and coastal environment for sustainable development. Environ-
SDGs. Retrived November 5, 2019, from https://​www.​ mental Monitoring and Assessment, 187, 231. https://​doi.​
gstco​uncil.​org/​wp-​conte​nt/​uploa​ds/​GSTC-​Desti​nation-​ org/​10.​1007/​s10661-​015-​4477-7
Crite​ria-​v2.0.​pdf Kumar, S., Radhakrishnan, N., & Mathew, S. (2014). Land use
Guan, D., Gao, W., Watari, K., & Fukahori, H. (2008). Land change modelling using a Markov model and remote sens-
use change of Kitakyushu based on landscape ecology and ing. Geomatics, Natural Hazards and Risk, 5(2), 145–156.
Markov model. Journal of Geographical Sciences, 18(4), https://​doi.​org/​10.​1080/​19475​705.​2013.​795502
455–468. Levent, B. A. T., Öztekin, A., Şahin, F., Arıcı, E., & Sandıkçı,
Guan, D. J., Li, H. F., Inohae, T., Su, W. C., Nagaie, T., & U. (2018). An overview of the Black Sea pollution in Tur-
Hokao, K. (2011). Modeling urban land use change by the key. Mediterranean Fisheries and Aquaculture Research,
integration of cellular automaton and Markov model. Eco- 1(2), 67–86.
logical Modelling, 222, 3761–3772. Li, S. N., Wang, G. X., Deng, W., Hu, Y. M., & Hu, W. W.
Han, J., Hayashi, Y., Cao, X., & Imura, H. (2009). Application (2009). Influence of hydrology process on wetland land-
of an integrated system dynamics and cellular automata scape pattern: A case study in the Yellow River Delta.
model for urban growth assessment: A case study of Ecological Engineering, 35, 1719–1726.
Shanghai, China. Landscape and Urban Planning, 91, Liping, C., Yujun, S., & Saeed, S. (2018). Monitoring and
133–141. predicting land use and land cover changes using remote
Hegazy, I. R., & Kaloop, M. R. (2015). Monitoring urban sensing and GIS techniques—A case study of a hilly area,
growth and land use change detection with GIS and Jiangle, China. PLoS ONE 13(7), e0200493. https://​doi.​
remote sensing techniques in Daqahlia governorate Egypt. org/​10.​1371/​journ​al.​pone.​02004​93
International Journal of Sustainable Built Environment, Liu, J. Y., Liu, M. L., Deng, X. Z., Zhuang, D. F., Zhang, Z.
4, 117–124. https://​doi.​org/​10.​1016/j.​ijsbe.​2015.​02.​005 X., & Luo, D. (2002). The land use and land cover change
Hepcan, C. C., Turan, I. A., & Ozkan, M. B. (2011). Monitoring database and its relative studies in China. Journal of Geo-
land use change in the Cesme coastal zone, Turkey using graphical Sciences, 12, 275–282.
aerial photographs and satellite imaging. Land Degradation Liu, R., & Zhu, D. L. (2010). Methods for detecting land use
& Development, 22, 326–333. https://​doi.​org/​10.​1002/​ldr.​ changes based on the land use transition matrix. Resource
997 Science, 32, 1544–1550.
Hua, A. (2017). Application of Ca-Markov model and land use/ Logsdon, M. G., Bell, J. E., & Westerlund, V. F. (1996). Prob-
land cover changes in Malacca River Watershed, Malaysia. ability mapping of land use change: A GIS interface for
Applied Ecology and Environmental Research, 15, 605–622. visualizing transition probability. Computers, Environ-
Huang, F., Huang, B., Huang, J., & Li, Sh. (2018). Measuring ment and Urban Systems, 20, 389–398.
land change in coastal zone around a rapidly urbanised bay. Longley, P. A., Goodchild, M. F., Maguire, D. J., & Rhind, D.
International Journal of Environmental Research and Pub- W. (2005). Geographic ınformation systems and science
lic Health, 15, 1059. https://​doi.​org/​10.​3390/​ijerp​h1506​1059 (pp. 364–382). England: John Wiley & Sons Ltd.

Vol:. (1234567890)
13
Environ Monit Assess (2022) 194:846 Page 23 of 24 846

Lu, D., Mausel, P., Brondízio, E., & Moran, E. (2004). Change Peng, J. (2008). Rural land use change during 1986–2002 in
detection techniques. International Journal of Remote Lijiang, China based in remote sensing and GIS data. Sen-
Sensing, 25, 2365–2407. sors, 8, 8201–8223.
Martínez-Fernández, J., Esteve, M. A., Carreño, M. F., Palazón, Petit, C., Scudder, T., & Lambin, E. (2001). Quantifying pro-
J. A. (2009). Dynamics of land use change in the Mediter- cesses of land-cover change by remote sensing: Reset-
ranean : Implications for sustainability, land use planning tlement and rapid land-cover changes in south-eastern
and nature conservation, in Land Use Policy (Nova Sci- Zambia. International Journal of Remote Sensing, 22(17),
ence Publishers), 101–145. ISBN: 978–1–60741–435–3. 3435–3456. https://​doi.​org/​10.​1080/​01431​16001​00068​81
Memarian, H., Balasundram, K. S., Talib, B. J., Sung, T. B. S. Pontius, R. G., McEachern, M., & Shusas, E. (2004). Detect-
C., Sood, M. A., & Abbaspour, K. (2012). Validation of ing important categorical land changes while accounting
CA-Markov for simulation of land-use and cover change for persistence. Agriculture, Ecosystems and Environment,
in the langat basin Malaysia. Journal of Geographic Infor- 101(2–3), 251–268. https://​doi.​org/​10.​1016/j.​agee.​2003.​
mation System, 4, 542–554. 09.​008
Mimura, N. (2008). Conclusions: The rapidly changing environ- Rawat, J. S., & Kumar, M. (2015). Monitoring land use/cover
ment of the Asia and Pacific Region and its ımplications for change using remote sensing and GIS techniques: A case
sustainability of the coastal zones. In: Mimura, N. (ed.), Asia- study of Hawalbagh block, district Almora, Uttarakhand,
Pacific Coasts and Their Management. States of Environment. India. The Egyptian Journal of Remote Sensing and Space
Springer, Amsterdam. pp. 345–358. https://​doi.​org/​10.​1007/​ Sciences, 18(1), 77–84. https://​doi.​org/​10.​1016/j.​ejrs.​
978-1-​4020-​3625-5_6 2015.​02.​002
Ministry of Culture and Tourism (MoCT). (2022). Culture- Rawat, J. S., Biswas, V., & Kumar, M. (2013). Changes in land
tourism conservation and development zones and tourism use/cover using geospatial techniques: A case study of
centers. Republic of Turkey Ministry of Culture and Tour- Ramnagar town area, district Nainital, Uttarakhand, India.
ism. Retrieved November 9, 2021, from https://​yigm.​ktb.​ Egyptian Journal of Remote Sensing and Space Sciences,
gov.​tr/​TR-​9669/​ktkgb-​ve-​turizm-​merke​zleri.​html 16, 111–117. https://​doi.​org/​10.​1016/j.​ejrs.​2013.​04.​002
Ministry of Transport and Infrastructure (MoTI). (2021). Ship- Regmi, R., Saha, S., & Balla, M. (2014). Geospatial analysis of
yards and coastal structures statistics. Republic of Tür- land use land cover change predictive modeling at Phewa
kiye Ministry of Transport and Infrastructure. Retrived Lake Watershed of Nepal. International Journal of Cur-
November 9, 2021, from https://​tkygm​istat​istik​leri.​uab.​ rent Engineering and Technology, 4, 2617–2627.
gov.​tr/ Robert, S., Fox, D., Boulay, G., Grandclément, A., Garrido,
Mondal, M. S., Sharma, N., Garg, P. K., & Kappas, M. (2016). M., Pasqualini, V., Prévost, A., Schleyer-Lindenmann,
Statistical independence test and validation of CA Markov A., & Trémélo, M. L. (2019). A framework to analyse
land use land cover (LULC) prediction results. The Egyp- urban sprawl in the French Mediterranean coastal zone.
tian Journal of Remote Sensing and Space Science, 19(2), Regional Environmental Change, 19, 559–572. https://​doi.​
259–272. org/​10.​1007/​s10113-​018-​1425-4
Müller-Hansen, F., Cardoso, M. F., Dalla-Nora, E. L., Donges, Salvati, L., Smiraglia, D., Bajocco, S., & Munafò, M. (2014).
J. F., Heitzig, J., Kurths, J., & Thonicke, K. (2017). A Land use changes in two Mediterranean coastal regions:
matrix clustering method to explore patterns of land-cover Do urban areas matter? World Academy of Science, Engi-
transitions in satellite-derived maps of the Brazilian Ama- neering and Technology. International Journal of Envi-
zon, Nonlin. Processes Geophys, 24, 113–123. https://​doi.​ ronmental, Ecological, Geological and Mining Engineer-
org/​10.​5194/​npg-​24-​113-​2017 ing, 8(9), 562–566.
Munthali, M. G., Mustak, S., Adeola, A., Botai, J., Singh, S. Santé, I., García, A. M., Miranda, D., & Crecente, R. (2010).
K., & Davis, N. (2020). Modelling land use and land cover Cellular automata models for the simulation of real-world
dynamics of Dedza district of Malawi using hybrid cellu- urban processes: A review and analysis. Landsc. Urban
lar automata and Markov model. Remote Sensing Applica- Plann., 96, 108–122.
tions: Society and Environment., 17, 100276. https://​doi.​ Sathees, K., Nisha, R., & Samson, M. (2014). Land use change
org/​10.​1016/j.​rsase.​2019.​100276 modelling using a Markov model and remote sensing.
Neumann, B., Vafeidis, A. T., Zimmermann, J., & Nicholls, R. Geomatics, Natural Hazards and Risk, 5(2), 145–156.
(2015). Future coastal population growth and exposure to Strategy and Budget Presidency (SBP). (2022). Development
sea-level rise and coastal flooding - A global assessment. plans (from ­1st five year plans to the ­11th ). Presidency of
A Global Assessment. PLOS ONE, 10(6), e0131375. the Republic of Turkey. Strategy and Budget Presidency.
https://​doi.​org/​10.​1371/​journ​al.​pone.​01313​75 Retrieved November 9, 2021, from https://​www.​sbb.​gov.​
Niculescu, S., & Lam, C. (2019). Geographic object-based tr/​kalki​nma-​planl​ari/
ımage analysis of changes in land cover in the coastal Sidle, R. C., Ziegler, A. D., Negishi, J. N., Nik, A. R., Siew,
zones of the Red River Delta (Vietnam). Journal of Envi- R., & Turkelboom, F. (2006). Erosion processes in steep
ronmental Protection, 10, 413–430. https://​doi.​org/​10.​ terrain -Truths, myths, and uncertainties related to forest
4236/​jep.​2019.​103024 management in Southeast Asia. Forest Ecology and Man-
Nouri, J., Gharagozlou, A., Arjmandi, R., Faryadi, S., & Adl, agement., 224, 199–225.
M. (2014). Predicting urban land use changes using a CA– Singh, S. K., Mustak, S., Srivastava, P. K., Szabó, S., & Islam,
Markov model. Arabian Journal for Science and Engi- T. (2015). Predicting spatial and decadal LULC changes
neering, 39, 5565–5573. through cellular automata Markov chain models using

Vol.: (0123456789)
13
 846 Page 24 of 24 Environ Monit Assess (2022) 194:846

earth observation datasets and geo-information. Environ- change over Banten Bay. Jurnal Ilmiah Geomatika, 12(1),
mental Processes, 2, 61–78. 11–21.
Symeonakis, E., Calvo-Cases, A., & Arnau-Rosalen, E. (2007). Xu, P., Guo, Y., & Fu, B. (2019). Regional ımpacts of climate
Land use change and land degradation in Southeastern and land cover on ecosystemwater retention services in
Mediterranean Spain. Environmental Management, 40, the Upper Yangtze River Basin. Sustainability, 11, 5300.
80–94. https://​doi.​org/​10.​1007/​s00267-​004-​0059-0 https://​doi.​org/​10.​3390/​su111​95300
Tang, J., Wang, L., & Yao, Z. (2007). Spatio-temporal urban Yang, Q. S., & Li, X. (2007). Integration of multi-agent sys-
landscape change analysis using the Markov chain model tems with cellular automata for simulating urban land
and a modified genetic algorithm. International Journal expansion. Scientia Geographica Sinica, 27(2007),
of Remote Sensing, 28(15), 3255–3271. https://​doi.​org/​10.​ 542–548.
1080/​01431​16060​09627​49 Yuan, F., Sawaya, K. E., Loeffelholz, B. C., & Bauer, M. E.
Tian, H., Lindenmayer, D. B., Wong, G. T. W., Mao, Z., (2005). Land cover classification and change analysis of
Huang, Y., & Xue, X. (2018). A methodological frame- the twin cities (Minnesota) metropolitan area by multi
work for coastal development assessment: A case study of temporal Landsat remote sensing. Remote Sensing of
Fujian Province, China. The Science of the Total Environ- Environment, 98, 317–328. https://​doi.​org/​10.​1016/j.​rse.​
ment, 615, 572–580. https://​doi.​org/​10.​1016/j.​scito​tenv.​ 2005.​08.​006
2017.​09.​309 Zhou, Q., Li, B., & Kurban, A. (2008). Trajectory analysis of
TurkStat (Turkish Statistical Institute). (2021). Address based land cover change in arid environment of China. Interna-
population registration system results, TurkStat. Retrived tional Journal of Remote Sensing, 29, 1093–1107.
December 21, 2021, from http://​www.​tuik.​gov.​tr. Zhu, D. L. (2010). Methods for detecting land use changes
URL-1. Retrieved December 15, 2021, from https://​www.​eea.​ based on the land use transition matrix. Resourc Science,
europa.​eu/​data-​and-​maps/​data/​corine-​land-​cover-​2000-​ 32, 1544–1550.
coast​line Zhu, G., Xu, X., Ma, Z., Xu, L., & Porter, J. H. (2012). Spa-
Vázquez-Quintero, G., Solís-Moreno, R., Pompa-García, M., tial dynamics and zoning of coastal land-use change along
Villarreal-Guerrero, F., Pinedo-Alvarez, C., & Pinedo- Bohai Bay, China, during 1979–2008. Journal of Coastal
Alvarez, A. (2016). Detection and projection of forest Research, 28, 1186–1196.
changes by using the Markov chain model and cellular
automata. Sustainability, 8(3), 236. https://​doi.​org/​10.​ Publisher’s Note Springer Nature remains neutral with regard
3390/​su803​0236 to jurisdictional claims in published maps and institutional
Wang, H., Li, X., Long, H., Qiao, Y., & Li, Y. (2011). Develop- affiliations.
ment and application of a simulation model for changes in
land-use patterns under drought scenarios. Computers and
Springer Nature or its licensor holds exclusive rights to this
Geosciences, 37, 831–843.
article under a publishing agreement with the author(s) or other
Weng, Q. (2002). Land use change analysis in the Zhujiang
rightsholder(s); author self-archiving of the accepted manuscript
Delta of China using satellite remote sensing, GIS and
version of this article is solely governed by the terms of such
stochastic modelling. Journal of Environmental Manage-
publishing agreement and applicable law.
ment, 64, 273–284.
Wijanarto, A. B. (2006). Application of Markov change detec-
tion technique for detecting landsat etm derived land cover

Vol:. (1234567890)
13