Scientific Journal of King Faisal University (Basic and Applied Sciences) In Press

Land Use/ Land Cover Change Detection in Baer and al- Bassit Region,
Latakia, Syria during the Period of 1972- 2018
Ola Ali Merhej(1) , Mahmoud Ali(1), Ali Thabeet(2), and Younes Idriss(3)
(1) Forestry and Ecology Department, Agriculture Faculty
Tishreen University, Syria.
(2) Natural resources Department, Agriculture Faculty,
Aleppo University- Syria
(3) General Organization of Remote Sensing
Damascus, Syria
Received 24 March 2019 - Accepted 10 December 2019
https://doi.org/10.37575/b/sci/2070

ABSTRACT
Monitoring land use/ land cover (LULC) changes is important for assessing the dynamics between land cover
types and understanding the anthropogenic impact on these changes, and remote sensing techniques represent
important tools to achieve this goal.
This paper aimed at mapping and analyzing LULC changes in Baer and al-Bassit region, Latakia governorate,
Syria. For this goal 15 Landsat multi-temporal images during the period of 1972-2018 were used, and each image
was classified using (Maximum Likelihood Algorithm) supervised classification into four categories of land use
named: forests, agricultural land, water and urban areas. Accuracy assessment of all images was performed and the
average values of the overall accuracy of the classification was (89%) and the average value of the Kappa index
was (0.85).
The area of each land use category was calculated in each LULC map and the trend of change among these
categories was analyzed during the study period using linear regression analysis. Forest category was the only
group that decreased by 21.8% between 1972 and 2018, compared to an increase in all other categories (0.6, 4.3
and 16.8 % for water, urban and agriculture, respectively) for the same period. This indicates conversion of forests
into agricultural land and urban areas. The results of this study can be used as an efficient tool to manage and
improve Baer and al-Bassit forests regarding physiographical and human characteristics and could facilitate the
creation of a database for LULC changes in this region.
Key Words: Linear regression, LULC dynamic, Remote Sensing, Supervised classification, Syria.

INTRODUCT3ION biodiversity loss and climate changes which

Land cover refers to the physical attributes in turn cause losses of related environmental
of the earth’s surface, including those made services to such a degree that more than a
by human activities e.g., settlements. While half of the world’s forest cover has been lost
land use refers to the manner by which (FAO, 2010).
land has been used by people, usually with Many activities of local people and
an emphasis on the functional role of land government organizations i.e. urban
for economic activities (Ramachandra and extension, expansion of agricultural land,
Kumar, 2004). The land use/ land cover forest fires, and so on have resulted in the
patterns of a region are a result of natural and loss of extensive amounts of forest cover
socio-economic factors and their utilization and other land use. LULC change analysis
by humans in time and space (Rawat and is a vital step in managing natural resources,
Kumar, 2015). monitoring environmental changes,
It is increasingly acknowledged that LULC protecting and planning for earth resources
changes have become a key subject that (Torahi and Rai, 2011).
urgently needs to be addressed in the Geographic Information Systems (GIS) and
study of global environmental changes remote sensing techniques are widely used
(Hu et al., 2019), because of major current to analyze the distribution, patterns and
environmental issues such as desertification, trends of the LULC processes. These tools
35
Land Use/ Land Cover Change Detection in Baer and al- Bassit Region,... O. A. Merhej et al.

have proven to be very valuable to obtain period of 1972- 2018; and 2) to detect the
exact and coherent information according to deforestation rate and the changes that have
the spatial reality (Monjardín et al., 2017). taken place during the study period.
Another important tool used by planners to
control the trends of forest cover changes is MATERIALS AND METHODS
regression relations. According to the fact that Baer and al-Bassit region extends between
environmental science deals with different longitudes (35o 47’ 49.2”, 36o 15’ 57.44” E)
phenomena; therefore, in ecological issues, and latitudes (35o 57’ 0.6”, 35o 35’ 42.7” N)
multiple regression has much importance and encompasses an area of 73000 hectares.
(Jahanifar et al., 2018). It lies in the north western part of coastal
In the last few decades, many specialists mountains in in the north and north-west
(Ramachandran and Reddy, 2017; Batar et Latakia, Syria (Figure 1). The study area
al., 2017; Teka et al., 2018) have studied is characterized by brown Mediterranean
LULC changes explicitly using LULC soils, and a Mediterranean climate type with
maps with the application of GIS and a rainy winter and a long dry summer. The
remote sensing data. In Syria, a time series altitude above sea level is between 300 and
of Landsat images of Tartous Governorate 1400 meters, and the average rainfall ranges
was used to study the relationship between between 800 and 1200 mm/year.
the applied agricultural policy and land use The area is characterized by forest cover
changes in this region. The study showed covering more than 70% of the area, which
that there is an expansion in the cultivation is the best and most important forests
of citrus and greenhouses as a result of the in terms of area and diversity in Syria
agricultural plan implemented in the same consisting mainly of conifers, oak and
time period. It also confirmed the possibility many other species. Villages are located in
of determining the spatial relationship the western and southern parts of the study
between LULC changes and agricultural area, and the majority of people living there
policies through remote sensing applications depend on agriculture in their livelihoods,
(Ibrahim et al., 2014). Al-Fares (2013) also so agricultural land is mostly concentrated
relied on Landsat-MSS, TM, and ETM+ around the villages.
images; and Terra: Aster for mapping LULC,
and mapping irrigated agricultural projects
for the years of 1975, 1987, 2005 and 2007
in the Euphrates River Basin in Syria. The
results showed that the area of irrigated
agricultural projects in the Euphrates basin
has increased by 148% between 1975 and
2007.
Although the Baer and al-Bassit region is
the most important forestry areas of Syria, it
suffers from different human encroachments.
Therefore, detecting forest land cover
changes can provide enough knowledge
to design a forest management strategy.
This study attempts to identify the spatio- Figure 1. Location of the study area, Baer and
temporal patterns of land cover changes in al- Bassit region, Latakia governorate, Syria
the last five decades using multi-temporal
Landsat images. The main objectives of The Landsat data are available freely on
this research are: 1) to accurately map the NASA’s Web site (USGS, 2018), and we
extent of the different LULC classes in the directly downloaded 15 images for the period
36
Scientific Journal of King Faisal University (Basic and Applied Sciences) In Press

of 1972-2018 from the website. Because the using Erdas Imagine 2015 software.
changes that have occurred in this region Google Earth imageries for different years,
are so frequent, we have chosen an image a set of topographic maps of North Latakia
of every 2 or 3 years. The selected acquired at 1:25000, and an image of Corona, (KH-
date was in the summer (June to October) 4A) with a 9-foot, black -and- white spatial
to avoid changes due to different seasons resolution captured in 1969 were used to
and to obtain cloud-free images as much as assess the accuracy of Landsat images
possible. Table (1) shows the resolution and classification.
other specifications of these images. The overall goal of image classification
Image pre-processing is very important procedures is to automatically classify all
for LULC analysis since different sensors pixels in an image into categories or features
were used, so Geometric and Radiometric of land cover, and typically, multi-spectral
corrections were applied as in Ali et al. data are used to make a classification. In
(2018), and because of the differences in fact, the spectral pattern within the data for
spatial resolution between MSS and other each pixel is used as a numerical basis for
images, the resampling process was applied classification (Lillesand et al., 2008).

Table (1) Characteristics of Landsat scenes used in the classification.

SENSOR_ID Date Acquired Spatial Resolution (m) Spectral resolution Used bands
L1 MSS 22/10/1972 80 4 4 to 7
L5 TM 04/07/1984 30 8 1 to 5 + 7
L5 TM 15/09/1987 30 8 1 to 5 + 7
L5 TM 20/09/1989 30 8 1 to 5 + 7
L5 TM 12/09/1992 30 8 1 to 5 + 7
L5 TM 15/10/1998 30 8 1 to 5 + 7
L5 TM 04/10/2000 30 8 1 to 5 + 7
L5 TM 20/06/2002 30 8 1 to 5 + 7
L5 TM 16/09/2005 30 8 1 to 5 + 7
L5 TM 21/08/2007 30 8 1 to 5 + 7
L5 TM 11/11/2008 30 8 1 to 5 + 7
L5 TM 31/07/2011 30 8 1 to 5 + 7
OLI_TIRS 08/10/2013 30 11 2 to 7
OLI_TIRS 12/09/2015 30 11 2 to 7
OLI_TIRS 19/10/2017 30 11 2 to 7
OLI_TIRS 06/10/2018 30 11 2 to 7

To determine the classification of land use/ is based primarily on pre-determination of

land cover, the supervised classification training areas and samples, and Figure (2)
method was applied in the Erdas Imagine 2015 summarizes the three basic steps involved in
software using the reflective bands in each a typical supervised classification procedure
image (Table 1). Supervised classification (Lillesand et al., 2008).

37
Land Use/ Land Cover Change Detection in Baer and al- Bassit Region,... O. A. Merhej et al.

Figure 2. Basic steps in supervised classification (Lillesand et al., 2008).

The training samples were taken from each is through a Corona image and topographic
image individually, and a spectral signature maps for 1972 image, 1980s and 1990s
file of each image was created using a images, Respectively, and for the remaining
signature editor tool, by sampling each of images we used Google Earth imageries
the studied areas according to Congalton without field survey because the study area
(1991) for the number of samples. was not accessible since 2011, (it is out of
Four categories of land use for all images were the government control).
identified and mapped, named: agricultural The kappa index shows the level of similarity
areas, urban areas, water and forests. The between a set of control fields and the
Maximum Likelihood Classification (MLC) classified image. It is utilized as a standard
method, as the most widely used in the part for precision evaluation to measure
scientific literature, is fast, easy to apply the accuracy between the remote sensing
and enables a clear interpretation of the derived classification map and reference
results (Bolstad and Lillesand, 1991). This data, referred to as row and column totals
algorithm can obtain a spectral image of (Jensen, 2003).
each land use class through variance and The value of Kappa is between 0 and 1.
covariance statistics of the set of training If the value is equal to zero, then there is
sites identified in the image and calculates no agreement between the map and the
the probability of belonging to each class reference. If Kappa is equal to one, it shows
according to the spectral signature; this that there is a strong agreement between the
method has been proven in works such as map and the reference.
those of (López and Plata, 2009; Hassan et The general “overall” accuracy indicates the
al., 2016), with satisfactory results. percentage of pixels properly classified, the
The classification is not fully considered percentage of producer’s accuracy sets the
until its accuracy is assessed (Lillesand et percentage of a kind of a particular land use
al., 2008). Therefore, to verify the validity change correctly classified in the image; and
of the results obtained, the accuracy was the percentage of user’s accuracy provides
assessed using the Accuracy Assessment the percentage of a land use class in the image
tool in Erdas Imagine by distributing random that matches with the class that corresponds
points representing all classes of land uses in in the land (Congalton et al., 1983).
the image, and then compare these points in The simple linear regression is a technique
terms of their classification with what it truly used to analyze the relationships between
38
Scientific Journal of King Faisal University (Basic and Applied Sciences) In Press

two quantitative variables. In linear done by SPSS V25 software.

regression, we assume that there is a linear
relation between dependent and independent RESULT AND DISCUSSION
variables (Salman-Mahini and Kamyab, 1. Land use/ Land cover maps
2012). It is used in this study to determine LULC maps were prepared and presented
the general trend of the land use change for all the images in Figure (3). The maps
and its significance during the period of show that there is an increase and a decrease
1972-2018, where the dependent variable in different land use categories. We noted
was “time” and independent variable was a decline in the extent of forest area during
“LULC class area” and the calculation was this period of (1972-2018).

Figure 3. LULC maps for the study area (Baer and al- Bassit) between 1972 and 2018.

Table (2) clarify the area and percentage land use/ land cover. In 1972, the highest
of each land cover category derived from percentage was the forest category, which
LULC maps. accounted for 55.23% of the total area of
The total area of Baer and al- Bassit region the study area, followed by agricultural land
is 72300 ha, divided into four categories of (39.43%).
39
Land Use/ Land Cover Change Detection in Baer and al- Bassit Region,... O. A. Merhej et al.

Table (2) Area (ha) and percentage of the land use classes.
Land cover class Water Forest Urban Agriculture Total
area 820.0 39959.6 3045.6 28529.01 72354.15
1972
% 1.13 55.23 4.21 39.43 100
area 818.46 35794.62 3406.77 32334.3 72354.15
1984
% 1.1 49.5 4.7 44.7 100
area 1066.5 32325.39 7545.33 31365.81 72303.03
1987
% 1.5 44.7 10.4 43.4 100
area 1115.73 29648.52 7079.04 34396.20 72239.49
1989
% 1.5 41.0 9.8 47.6 100
area 1209.15 32372.28 6196.86 32575.86 72354.15
1992
% 1.671155 44.741428 8.564623 45.022794 100
area 1174.59 34729.56 5191.74 31268.25 72364.14
1998
% 1.623166 47.992777 7.174465 43.209592 100
area 1164.42 48763.8 4084.29 18218.34 72230.85
2000
% 1.612081 67.511043 5.654495 25.222381 100
area 1276.38 37067.67 3397.59 30612.51 72354.15
2002
% 1.764073 51.230883 4.695778 42.309266 100
area 1154.97 34180.74 4111.2 32867.37 72314.28
2005
% 1.597153 47.26693 5.685184 45.450733 100
area 1151.82 31157.73 6302.7 33690.78 72303.03
2007
% 1.593045 43.093256 8.717062 46.596636 100
area 1029.06 36508.95 2980.08 31802.76 72320.85
2008
% 1.422909 50.481915 4.120637 43.974538 100
area 1266.39 35630.1 1989.9 33416.64 72303.03
2011
% 1.751503 49.278848 2.752167 46.217482 100
area 1235.34 26915.58 5450.58 38730.87 72332.37
2013
% 1.707866 37.210975 7.535464 53.545695 100
area 926.37 27661.32 5213.43 38519.73 72320.85
2015
% 1.280917 38.248057 7.208751 53.262275 100
area 1136.79 27809.82 4952.52 38455.02 72354.15
2017
% 1.571147 38.435694 6.844832 53.148327 100
area 1247.94 24205.14 6154.11 40713.66 72320.85
2018
% 1.72556 33.469103 8.509455 56.295881 100

Water bodies accounted for 1.1 % of the for the period of 1984-1987, where the
study area in 1972, which is the lowest water area increased by 0.4% (from 818.46
among land use categories, but this does not to 1066 hectares). This increase was due
make the study area water-poor. Water area to the construction of the 16 October Dam
changed slightly from year to year, except on Kabeer Shemali River. The dam lake
40
Scientific Journal of King Faisal University (Basic and Applied Sciences) In Press

was at the expense of agricultural land and lost between 1972 and 2000 has been turned
surrounding forests. In contrast, the decline into low forms of vegetation (Maquis),
in water area (from 1235.34 to 926.37 while some parts have also been lost for
hectares) between 2013 and 2015 was due to agricultural use.
the drought recorded during these two years. In Tartous governorate, which is somehow
Forests rotate with agriculture class by climatically similar to al-Baer and al-
occupying the largest proportion of land use Bassit region and varies with the nature of
categories during the study period (Table vegetation and human activities, forest area
2). The forest area decreased in favor of decreased by 262.4 km2 (651.84 km2 in
agriculture land until 1992 (from 39959.6 1987, 389.44 km2 in 2017) during the period
hectares to 32372.28 hectares) in exchange of 1987-2017. The majority of the decline
for an increase of the other categories. In the (72.5%) occurred during the period of 2002-
1990s, the proposed afforestation projects 2017 due to excessive logging heating by
began to show results, with an increase in the local communities, especially during the
forest area in 1992, 1998, 2000, by about war and was accompanied by a significant
7%, then it fell in 2005 and 2007 after Bassit increase in the area of agricultural land
Fire (from 51.23% of the study area in 2002 (Hammad et al., 2018).
to 43.1% in 2007). Urban area accounted for the lowest
Maps of 2005 and 2007 show the effects of percentage among the land use categories
Bassit fire in 2004, where the area of the fire after water class with a percentage of 4.21
has been turned into agricultural land class, % in 1972. The values of the urban category
due to the decline of vegetation biomass fluctuated widely, but we can say that it
in the region; to return back to forest class increased from 3045.6 hectares in 1972
in 2008 after the increase of the biomass to 6154.11 hectares in 2018. The largest
because of regeneration of pine forests and increase was between 1984 and 1987 (from
reforestation with several forest species 3406.77 to 7545.33 hectares). In contrast,
(Kassas, 2008). there was a decline in the area of urban areas
The sharp decline in forest area in the years in the period of 2007-2011 (from 6302.7
after 2011 is due to war and large forest hectares to 1989.9 hectares) by about 6%,
fires in Northern Latakia. In the crisis years and then there was an increase to 8.5% of the
(since 2011), the forest area has declined in study area (6154.11 hectares) in 2018. This
different parts of the study area, especially in is again due to the spectral confusion (Yang,
Hafah region, where armed operations took 2002), and this finding is consistent with the
places, in addition to fires and over-cutting results of the Aragrande and Argenti (2001)
for heating and cooking (Fig 3). The forest study, where urban expansion causes loss of
area decreased from 35630.1 hectares in 2011 agriculture land due to high competition for
to 26915.58 hectares in 2013 and continued other uses (housing, industry, etc.).
to decline to the lowest value in 2018 with The difference in the proportion of the urban
an area of 24205 hectares (equivalent to class may be due to the lower accuracy of
33.47% of the study area). This corresponds the classification of this category, which
to the study of the borders and areas of fire represents a lower producer’s accuracy for
in northern Latakia between 2002 and 2017, this category in most classified images. This
where the values of Normalized Burn Ratio can be attributed to the similarity between the
(NBR) were low in Bassit in 2005 and 2006, spectral reflection of urban and agricultural
and in 2014 and 2015 in Hafah (Merhej et land, leading to low accuracy of the maps
al., 2018). (Seto et al., 2002), especially since most of the
The largest proportion of forest area (22%1) urban areas in the study area are agricultural
villages mixed houses with cultivated land.
(1) (24200- 39960/72354)* 100
Natural and agricultural eco-systems are
41
Land Use/ Land Cover Change Detection in Baer and al- Bassit Region,... O. A. Merhej et al.

rapidly being converted into urban uses/ that the Kappa index gave very good values
built-up areas to meet the residential and for the applicable classification, ranging
economic needs of the ever-increasing urban from 0.8 in 2015 to 0.91 in 2005. The overall
populace (Yang and Liu, 2005). accuracy ranged from 87.36% in 1972 to
It is evident from the Table (2) that 92.94% in 2005, a very good proportion of
agricultural land occupied 39.43% of the the accuracy of our classification.
study area in 1972, increasing by about 17% Deriving signatures from multiple images
between 1972 and 2018 (from 28529.01 can make the signatures more robust in the
hectares in 1972 to 40713.66 hectares in sense that they result in reasonably good
2018). The increase in agricultural land can classifications across years, but on the other
be linked to the rapid poverty and the war in hand, it doesn’t necessarily produce the best
the last several years, which caused increases classification in any single year (Laborte et
in forest cutting to convert into agriculture al., 2010). The different values of the Kappa
uses and to sell fuel woods. index and overall accuracy may be due to the
fact that each image is classified according
2. Accuracy Assessment of Supervised to its spectral profile. The absence of field
Classification survey and dependence on Google Earth’s
The accuracy of the classification images images to evaluate accuracy may be the
was evaluated, Table (3) shows the accuracy reasons for the evenness accuracy of the OLI
assessment details for all the images used in images with the images of the 1970s and
the study in terms of overall accuracy and 1980s that used an aerial image of Corona
Kappa index, respectively. The results show and a set of Topographic maps.

Table 3. Overall accuracy and Kappa values for the images used in the study period.
Year 1972 1984 1987 1989 1992 1998 2000 2002 2005 2007 2008 2011 2013 2015 2017 2018
Overall
87.36 88.46 89.23 90.85 92.31 88.59 92.5 92.07 92.94 89.38 91.08 91.95 88.28 88.16 89.78 88.13
accuracy %
KAPPA (K^) 0.84 0.84 0.85 0.88 0.89 0.84 0.86 0.89 0.91 0.86 0.85 0.89 0.84 0.8 0.85 0.84

The producer’s accuracy values gave the linear regression analysis during 1972- 2018.
lowest values for the urban and agricultural The regression analysis showed a significant
land classes compared to higher values for trend towards the increase in water surface
the water category, since the producer’s area during the study period (1972-2018)
accuracy of the urban category was 61.54 (ANOVA, F = 5.85, P <0.05) (Figure 4).
and 66.67 in 2015, 2017, respectively. This
is due to the spectral confusion, where farms
at the beginning of the agricultural seasons
show a spectral reflection similar to that of
urban areas (Yang, 2002).

3. LULC change Analyses between 1972-

2018
The coefficient of determination R2 describes
the relationship between the LULC area and
the time, and determines how much variable Figure 4. General trend for water surface area
variation (LULC area) is dependent on the during the period of 1972- 2018.
independent variable (Year) (Zar, 1984;
Bihamta and Zare, 2011). Figures (4, 5 and Figure (5) presents forest area changes
6) illustrate the LULC change trends and the between 1972 and 2018, where the overall
forest area decreased during the study period.
42
Scientific Journal of King Faisal University (Basic and Applied Sciences) In Press

The regression analysis also showed a agriculture have been made by time.
significant general trend towards decreasing The regression analysis shows a significant
forest area (ANOVA, F = 8.068, P <0.05). general trend towards the increase in
According to the value of determination agricultural land area during the study period
coefficient which equals 0.38, we can say that (ANOVA, F = 17.378, P <0.05).
38 % of decreases in forest land cover have These changes are not systematic for most
been made by independent variable (time). types of land use in the study area, where the
Whereas the regression analysis for urban greater the area of the type and it importance,
area showed that there was no significant the greater the magnitude of the calculated
general trend in urban area during the study changes. Currently, the main factor in
period (1972-2018) (ANOVA, F = 0.001, P> reducing forest cover in the region is human
0.05). activities. Residents of residential districts in
the area are busy removing and cutting off
the forests to cover their fuel consumption
and the expansion of their agricultural lands.

Figure 5. General trend for forest surface area

during the period of 1972- 2018.
Figure 6. General trend for agriculture area
Figure (6) presents the changes in agriculture during the period of 1972- 2018.
land area that occurred between 1972 and
2018, and the value of R2 was the highest, Table (4) shows the analysis of changes
indicating that the trend of increase in area of between 1972 and 2018. Figure (7) provides
agricultural areas was continuous throughout a chart for a better understanding of this
the study period, and 57 % of increases in change.

Table 4. LULC changes in Baer and Al-Bassit region between 1972- 2018.
area in ha % of total area area in ha % of total area
land cover class Change in ha
(1972) 1972 (2018) 2018
water 820.0 1.13 1247.94 1.7 427.95
forest 39959.6 55.23 24205.14 33.5 -15754.4
urban 3045.6 4.21 6154.11 8.5 3108.51
agriculture 28529.01 39.43 40713.66 56.3 12184.65
Total 72354.15 100 72320.85 100  

The results show that the forest category other hand, increased by 16.87 % (about
is the only category that decreased during 12185 hectares). Therefore, we concluded
the study period, while agriculture was that the rapid increase in agriculture is due to
the category that increased significantly. the conversion of forest cover to agricultural
Statistical analysis shows a 21.8 % decline land mainly, i.e., the large loss in the forest
in the forest area between 1972 and 2018 category occurred due to increased demand
(about 15755 hectares). Agriculture, on the for agricultural land and fuel woods.

43
Land Use/ Land Cover Change Detection in Baer and al- Bassit Region,... O. A. Merhej et al.

great potential in this technique, but it depends

on the availability of quality assistive data. 

REFERENCES
Aragrande, M., and Argenti, O. 2001. Studying
food supply and distribution systems to
cities in developing countries and countries
in transition. Retrieved on 28- 1-2018 from:
https://cutt.us/xCn3N.
Al-Fares, W. 2013. Historical land use/land
cover classification and its change detection
Figure 7. Percentage change in LULC
mapping using different remotely sensed
categories between 1972- 2018.
data from landsat (MSS, TM and ETM+) and
Terra (ASTER) sensors: A case study of the
CONCLUSIONS
Euphrates River Basin in Syria with focus
In situations of rapid land use changes, on agricultural irrigation projects. Springer
satellite remote sensing provides detailed International Publishing, eBook ISBN: 978-
information to understand the land use and 3-319-00624-6.
land cover in any region. All images of the
Ali, M., Thabeet, A., Idriss, Y., and Merhej, O.
study area were selected in the same season
2018. Preprocessing of landsat imageries used
through the years to produce a series of land
to mapping NDVI in North Latakia forests.
use classification maps. Tishreen University Journal for Research
The Accuracy coefficients obtained in this and Scientific Studies - Biological Sciences
study indicate that the classification method Series. 4(5): 92- 108.
used reflected very well the real land use and
Batar, A. K., Watanabe, T., and Kumar, A. 2017.
land cover changes that took place during
Assessment of land-use/land-cover change
the study period, but this does not mean
and forest fragmentation in the Garhwal
the classification procedure should not be Himalayan Region of India. Environments.
enriched with field verification when it is 4(2): 34- 37.
possible.
The LULC changes that have occurred in Bihamta, M., and Zare-Chahouki, M. 2011.
Principles of statistics for the natural resources
Baer and al-Bassit between 1972 and 2018
science. University of Tehran Press, Iran.
have been quantified from Landsat multi-
temporal data. It was realized that the change Bolstad, P., and Lillesand, T.M. 1991. Rapid
is threating the forests, where more than maximum likelihood classification.
150 km2 of forest areas were affected by Photogrammetry Engineering Remote
Sensing. 571: 67- 74.
deforestation between 1972 and 2018; this
value impacts on future environment and Congalton, R.G. 1991. A Review of assessing the
urban management. The agricultural lands accuracy of classifications of remotely sensed
were found to have increased by 16.8% data. Remote Sensing Environment. 37: 35-
between 1972 and 2018. The majority of this 46.
change, however, occurred after 2011 with Congalton, R.G., Oderwald, R.G., and Mead,
the beginning of the war, where the increase R.A. 1983. Assessing landsat classification
of agriculture area happened at the expense accuracy using discrete multivariate statistical
of forest. techniques. Photogrammetry Engineering and
Based on our results, large scale agriculture Remote Sensing. 49(12): 1671- 1678.
activities should be restricted in Baer and al- FAO. 2010. Evaluación de los recursos forestales
Bassit region. Programs of forest restoration mundiales: Informe Nacional México; FAO:
and conservation measures should be initiated Roma, Italy. Retrieved on 31- 5- 2018 from:
and implemented in the study area. There is https://cutt.us/AFtOZ.

44
Scientific Journal of King Faisal University (Basic and Applied Sciences) In Press

Jahanifar, K., Amirnejad, H., Mojaverian, M., López, V. V. H., and Plata, R. W. 2009. Análisis
and Azadi, H. 2018. Land change detection de los cambios de cobertura de suelo
and effective factors on forest land use derivados de la expansión urbana de la Zona
changes: application of land change modeler Metropolitana de la Ciudad de México 1990-
and multiple linear regression. Journal 2000. Investigating Geography. 68: 85- 101.
of Applied Science and Environment
Merhej, O. A., Ali, M., Thabeet, A., and Idriss,
Management. 22(8): 1269 –1275.
Y. 2018. Evaluation of forest fire damage and
Kassas, H. 2008. Studying Post fire regeneration risk in northern Latakia during the crisis years
of Pinus brutia Ten. after the 2004 fire in using the normalized burn ratio. Abstract
Ras – al- Bassit and its socio- economic No 23. Proceedings of the 10th General
dimensions. Master Thesis. Tishreen Organization of Remote Sensing conference,
University, Latakia, Syria. Damascus, Syria, 8- 10 Oct. 2018.
Hammad, M., László M., and Boudewijn L. Monjardín, A. S. A., Pacheco, A. C. E., Plata, R.W.,
2018. Land cover change investigation in and Corrales, B. G. 2017. La deforestación y
the Southern Syrian coastal basins during the sus factores causales en el estado de Sinaloa,
past 30-years using landsat remote sensing México. Madera y Bosques. 231: 7- 22.
data. Journal of Environmental Geography.
Ramachandran, R. M., and Reddy, C. S. 2017.
11 (1-2), 45–51.
Monitoring of deforestation and land use
Hassan, Z., Shabbir, R., Ahmad, S.S., Malik, changes (1925–2012) in Idukki district,
A.H., Aziz, N., Butt, A., and Erum, S. 2016. Kerala, India using remote sensing and GIS.
Dynamics of land use and land cover change Journal of the Indian Society of Remote
(LULCC) using geospatial techniques: A Sensing. 45: 163-170.
case study of Islamabad Pakistan. Springer
Ramachandra, T. V., and Kumar, U. 2004.
Plus. 5: 1- 11.
GRDSS for land use land cover dynamics
Hu, Y., Batunacun, Zhen, L., and Zhuang, D. analysis. Proceedings of the FOSS/GRASS
2019. Assessment of land-use and land- Users Conference- Bangkok, Thailand, 12-
cover change in Guangxi, China. Scientific 14 Sept. 2004.
Reports, 9, Article number: 2189 (2019).
Rawat, J. S., and Kumar, M. 2015. Monitoring land
(doi.org/10.1038/s41598-019-38487-w).
use/cover change using remote sensing and
26-11-2018.
GIS techniques: A case study of Hawalbagh
Ibrahim, W.Y., Sam, B., and Paul, M. 2014. block, district Almora, Uttarakhand, India.
Agricultural policy effects on land cover and The Egyptian Journal of Remote Sensing and
land use over 30 years in Tartous, Syria, as Space Sciences. 18: 77- 84.
seen in landsat imagery. Journal of Applied
Salman-Mahini, A., and Kamyab, H. 2012.
Remote Sensing. 8(1), 083506. https://cutt.
Applied Remote Sensing and GIS with Idrisi.
us/DyQwq. 26-11-2018.
2nd Edition. Publication of Mehrmahdis.
Jensen, J. R. 2003. Introductory to Digital Image Tehran, Iran. 596 p.
Processing. A Remote Sensing Perspective.
Seto, K. C., Woodcock, C. E., Song, C., Huang,
3rd Edition. Prentice Hall, US.
X., Lu, J., and Kaufmann, R. K. 2002.
Laborte, A. G., Maunahan, A. A., and Hijmans, Monitoring land use change in the Pearl river
R. J. 2010. Spectral Signature Generalization delta using landsat TM. International Journal
and expansion can improve the accuracy of of Remote Sensing. 23: 1985- 2004.
Satellite Image Classification. PLoS One.
Teka, H., Madakadze, C. I., Botai, J. O., Hassen,
2010; 5(5): e10516. (doi:10.1371/journal.
A., Angassa, A., and Mesfin, Y. 2018.
pone.0010516). 26-11-2018.
Evaluation of land use land cover changes
Lillesand, T. M., Kiefer, R. W., and Chipman, using remote sensing landsat images and
J.W. 2008. Remote Sensing and Image pastoralists’ perceptions on range cover
Interpretation. 6th Edition. John Wiley and changes in Borana rangelands, Southern
Sons. Inc., India. Ethiopia. International Journal of Biodiversity
and Conservation. 10(1): 1-11.
45
Land Use/ Land Cover Change Detection in Baer and al- Bassit Region,... O. A. Merhej et al.

Torahi, A., and Rai, S.C. 2011. Land cover Yang, X., and Liu, Z. 2005. Using satellite
classification and forest change analysis, imagery and GIS for land use and land cover
using satellite imagery-a case study in Dehdez change mapping in an estuarine watershed.
area of Zagros Mountain in Iran. Journal of Int. J. Remote Sensing. 26: 5275- 5296.
Geographic Information System. 3: 1- 8.
Zar, J. H. 1984. Biostatistical analysis. 2nd
USGS. 2018. U.S. Geological Survey. Retrieved Edition, Prentice Hall, US.
on 15 -2- 2018 from: https://earthexplorer.
usgs.gov/.
Yang, X. 2002. Satellite monitoring of urban
spatial growth in the Atlanta metropolitan
area. Photogrammetric Engineering and
Remote Sensing. 68: 725- 734.

46
‫)‪Scientific Journal of King Faisal University (Basic and Applied Sciences‬‬ ‫‪In Press‬‬

‫ك�شف تغري ا�ستعماالت الأرا�ضي‪ /‬الغطاء الأر�ضي يف منطقة الباير والب�سيط‪ ،‬الالذقية‪� ،‬سوريا خالل الفرتة‬
‫‪2018 -1972‬‬
‫(‪)3‬‬
‫عال مرهج(‪ ،)1‬حممود عيل(‪ ،)1‬عيل ثابت(‪ ،)2‬يونس إدريس‬
‫(‪ )1‬قسم احلراج والبيئة‪ ،‬كلية الزراعة‪ ،‬جامعة ترشين‪ ،‬سورية‬
‫(‪ )2‬قسم املوارد الطبيعية‪ ،‬كلية الزراعة‪ ،‬جامعة حلب‪ ،‬سورية‬
‫(‪ )3‬اهليئة العامة لالستشعار عن بعد‪ ،‬دمشق‪ ،‬سورية‬
‫استالم ‪ 24‬مارس ‪2019‬م ‪ -‬قبول ‪ 10‬ديسمرب ‪2019‬م‬
‫‪https://doi.org/10.37575/b/sci/2070‬‬

‫امللخص‬
‫مراقبة تغريات الغطاء األريض‪ /‬استخدام األرض مهمة لتقييم الديناميات بني أنواع الغطاء األريض وفهم التأثري برشي املنشأ عىل‬
‫هذه التغيريات‪ ،‬ومتثل تقنيات االستشعار عن بعد أدوات مهمة لتحقيق هذا اهلدف‪ .‬هتدف هذه الورقة إىل رسم خرائط وحتليل تغري‬
‫استعامالت األرايض‪ /‬الغطاء األريض يف منطقة الباير والبسيط‪ ،‬حمافظة الالذقية‪ ،‬سورية‪ .‬استخدمت ‪ 15‬صورة الندسات متعددة‬
‫األزمنة للفرتة ‪ 2018 -1972‬وتم تصنيف كل صورة باستخدام خوارزمية االحتاملية القصوى للتصنيف املراقب إىل أربع فئات من‬
‫استعامالت األرايض هي‪ :‬الغابات‪ ،‬األرايض الزراعية‪ ،‬املياه واملناطق احلرضية‪ .‬تم تقييم دقة التصنيف جلميع الصور وبلغ متوسط قيم‬
‫الدقة اإلمجالية للتصنيف (‪ )% 89‬ومتوسط قيمة مؤرش كابا (‪.)0.85‬‬
‫ُحسبت مساحة كل فئة من فئات استعامالت األرايض يف كل صورة‪ ،‬وتم حتليل اجتاه التغيري هلذه الفئات خالل فرتة الدراسة باستخدام‬
‫عالقات االنحدار اخلطي‪ .‬كانت الغابات هي الفئة الوحيدة التي تعرضت لالنخفاض بني عامي ‪ 1972‬و‪ 2018‬بنسبة (‪ )% 21.8‬يف‬
‫مقابل زيادة جلميع الفئات األخرى (‪ % 16.8 ،4.3 ،0.6‬للمياه واملناطق احلرضية وللزراعة‪ ،‬عىل الرتتيب)‪ ،‬مما يشري إىل كرس أرايض‬
‫الغابة لصالح األرايض الزراعية واملناطق احلرضية‪ .‬يمكن استخدام نتائج هذه الدراسة كأداة فعالة إلدارة وحتسني غابات الباير والبسيط‬
‫فيام يتعلق باخلصائص الفيزيائية والبرشية ويمكن أن تسهل إنشاء قاعدة بيانات لتغريات ‪ LULC‬يف هذه املنطقة‪.‬‬
‫الكلامت املفتاحية‪ :‬االستشعار عن بعد‪ ،‬االنحدار اخلطي‪ ،‬التصنيف املراقب‪ ،‬تغري استعامالت األرايض‪ ،‬صور األقامر االصطناعية‪،‬‬
‫سورية‪.‬‬

‫‪47‬‬