Advertisement

Regional Environmental Change

, Volume 18, Issue 6, pp 1707–1719 | Cite as

Impacts of war in Syria on vegetation dynamics and erosion risks in Safita area, Tartous, Syria

  • Hazem Ghassan Abdo
Original Article

Abstract

Vegetation change and soil erosion are among the most serious environmental issues associated with the current the war in Syria. About 13 million people have been displaced, of which 8 million are inside Syria. The Syrian coastal region has received 1.4 million people of refugees since the onset of the war in Syria, resulting in an increased demand of services and goods from ecosystems, thus increasing the overall pressure from human activities on natural resources, especially vegetation and soil. The Syrian coastal region constitutes an important economic, touristic, and agricultural center in Syria. More than 90% of the vegetation of Syria is concentrated in the western coastal region. The study aims to assess the changes of vegetation cover and soil erosion in Safita area during the period 2011–2017 using satellite observations (NDVI) and a soil loss model (RUSLE). The results indicate a massive variation of vegetation cover, where degradation has occurred mainly in areas with high and very high densities of vegetation cover. As a result, soil erosion rates and their risk grades have increased remarkably. The estimated soil erosion amounted to 20.14 and 23.19 t ha−1 year−1 in 2011 and 2017, respectively. However, the results of this study make it a must for local planners to intervene quickly by using reliable and effective conservation techniques. A comprehensive analysis revealed that this unexpected pressure created by refugees can also significantly and swiftly alter the environmental characteristics of the area with potential serious consequences. So, impact of war in Syria on natural resources in the safe areas is a complex process which requires further detailed studies.

Keywords

Syrian coastal region Vegetation dynamics Soil erosion risk NDVI RUSLE model 

Introduction

War may cause environmental harms directly or indirectly, reducing human welfare (Reuveny et al. 2010; Mitri et al. 2014). Wars often destroy natural resources in conflict areas completely, along with putting the heavy pressure on them as a result of the changes caused by refugees. These changes are best exemplified in inhabitants’ lifestyle and their abuse of natural resources in safe areas inside country of war or even in the neighboring countries. Thus, the degradation of quality and productivity of environmental resources occurs simultaneously in safe (refugees pressure) and unsafe areas (war actions) in the same war country and with neighboring countries (Biswas 2002; Machlis and Hanson 2008; Mubareka and Ehrlich 2010; Reuveny et al. 2010; Mitri et al. 2014).

The growing global interest in the environment–migration nexus related to sociopolitical transitions is the product of growing concern over the potential impacts of future climate change and what these might mean for human well-being (Morrissey 2014). Environmental degradation, however, is one of the most important causes for human mobility (Swift 1973; Lahav 2016).

Since the outbreak of Syrian war in March 2011, it has led to the displacement of more than half of the Syrian population (11 million). The internally displaced people, who are estimated to be at around 8 million (Faour and Fayad 2014; Baldacchino and Sammut 2016), have been forced to flee their homes to safe areas, especially to Syrian coastal region (SCR) (OCHA 2017). The war in Syria has had devastating effects not only on humans but also on several vital sectors over the Syrian geography in general, such as archeological sites (Casana and Panahipour 2014), the complicated connections between water and war in Syria (Gleick 2014), water environment (Faour and Fayad 2014), irrigated agriculture (Jaafar et al. 2015), urban areas (Marx 2016), and land use and transboundary freshwater resources (Müller et al. 2016).

Reliable and updated information on vegetation degradation and soil erosion during the war in Syria is an essential prerequisite for prioritization of land as well as formulation of appropriate management strategies, which are key components for environmental re-strengthening in Syria when the war ends. Vegetation is one of the most important natural resources in Syria, and more 90% of it is located in the humid coastal area (Barakat et al. 2014a, b). SCR forest land is a notably important ecosystem in terms of biodiversity, timber, carbon storage, and recreational aspects. Recently, massive areas of vegetation in Syria have been shockingly destroyed, especially in a relatively safe Syrian area such as coastal area, as a result of several factors: (1) the war has taken place in the areas of fossil resources, which have created a severe shortage of fuel for heating and electricity supplies. So, locals have started cutting timber; (2) unexpected increase of residents caused by war refugees, at their various economic-social levels, has contributed to increasing the demand for the various goods and services especially housing. Consequently, urban areas have widely expanded, and finally, (3) there have been repeated and premeditated fires to produce charcoals. In the case of the ongoing war in Syria, however, the assessment of vegetation change in Syrian safe regions is more logical and objective than it is in zones where military war is still active.

Soil is functionally a non-renewable resource (Pal 2016). Soil erosion is a dynamic natural process which involves detachment, transportation, and accumulation of productive surface soil caused by accelerated interaction among humans, nature, and land, which negatively impacts natural resources and ecosystems (Fagbohun et al. 2016). Although changes in both vegetation and soil erosion are complex processes (Zhou et al. 2006), SCR represents a favorable geomorphologic, lithologic, hydro-climatic, and socioeconomic pattern for recourse degradation. However, the soil of SCR is extremely fertile soil (Saleh et al. 2016). In details, soils in SCR are highly sensitive to erosion due to two main factors: the system of the area’s rainfall which happens after long periods of drought in the form of violent, focused, and short-duration form, over only several days of the annual rainfall total. Second, the area has very steep terrain. These two factors drift tons of fragile soil of slopes (Mohammed et al. 2016; Abdo and Salloum 2017a, 2017b; Abdo 2017). Soil erosion is among the agro-environmental indicators developed by the European Commission services for monitoring agricultural and environmental policies (Panagos et al. 2015).

During war in Syria, many complicated accelerating factors have led to the loss of a great deal of vegetation cover, which has a significant role in reducing soil erosion (Elwell and Stocking 1976; Wu and Wang 2011). Thus, the spatial and quantitative assessment of the impact of vegetation change on soil erosion rates during war contains several different values for environmental enhancement phase which is logically to be started as soon as the war ends. The outcomes could help local planners develop strategies to assist preservation of the ecosystem in Safita area and SCR generally. To this end, this paper was organized to explore the vegetation change using two ETM+ Landsat sensors images, acquired on March 17, 2011, and March 16, 2017, on soil erosion by utilizing universal equation of soil loss (RUSLE) model in-depth multivariate analysis in Safita area.

Materials and methods

Study area

Safita area as shown in Fig. 1 on digital elevation model (DEM) with 335-km2 area is geographically situated in Tartous Governorate in the west of Syria. Safita area is boarded by Messyaf district from the east, the Akkar plain from the west, Tal-kalakh district from the south, and Drekish district from the north. In Safita area, 73.4% of the area is mountainous area, leaving only 26.6% as hilly land. The area under study belongs to the humid region in west of Syria, which have a typical Mediterranean climate with warm and rainy winters and hot and dry summers. The annual rainfall ranges from 750 to 1225 mm. More than 90% of the annual rainfall occurs in the winter from October to April, with no rain falls during the summer. The mean monthly temperature is 13.5 to 21.2 °C. June, July, and August are the hottest months of the summertime (Fig. 2).
Fig. 1

Map of Safita area showing location of study area on digital elevation model (DEM)

Fig. 2

Mean monthly temperature and rainfall in Safita area (1970–2016)

Satellite remote sensing and monitoring of vegetation change

The flowchart presented in Fig. 3 summarizes all the required data in this study. Currently, remotely sensed image has become a reliable and effective data source in vegetation dynamic studies (Zeng and Yang 2008; Liu et al. 2009; Wei et al. 2009; Hu et al. 2010), mainly on the basis of spectral reflection information and the contrast between strong absorption in the visible spectrum and high reflection in the near-infrared spectrum (Jiang et al. 2008). In addition, spectral information that has been obtained from remote sensing is the most widely used method for estimating vegetation cover (Ferreira and Panagopolous 2014).
Fig. 3

Schematic representation of work flow showing the methodology adopted for generating soil erosion susceptible areas during the war in Syria in Safita area

The NDVI (Normal Deference Vegetation Index) is the most widely relied vegetation index used because it effectively responds to vegetation fluctuations (Chen et al. 2006; Diodato and Bellocchi 2008; Zeng and Yang 2008; Rastmanesh et al. 2010; Barati et al., 2011). The NDVI is not only very sensitive to the physical characteristics of vegetation, but it can also effectively lower the impact from the sensor observation angle, solar radiation, and different soil backgrounds. Using the NDVI has become an important means for monitoring the changes in global vegetation cover and its conditions both at inter-annual and intra-annual time-scale (Ma et al. 2006; Propastin et al. 2008).

Further, many researchers have reported on the use of NDVI for vegetation monitoring (Lan 2009; Yang et al. 2011). Therefore, rigorous analysis and interpretation of vegetation changes in surface reflectance during war in the area under study adopted NDVI derived from the Landsat Enhanced Thematic Mapper Plus (ETM+), which were both downloaded from US Geological Survey (USGS) (http://earth explorer.usgs.gov). In this paper, the data has been projected to the Universal Transverse Mercator (UTM) projection system with WGS-84 datum, zone – 37. In details, the spectral specifications of the image data are described in Table 1. March was chosen as a reference for the distribution of vegetation cover due to several objective reasons. March 2011 is the beginning of the war in Syria, and it also represents the end of annual cutting timber period. However, Demirci and Karaburun (2012) have recommended determining the C factor values in RUSLE model during March, April, or May specifically, relying on Landsat imageries. However, Figs. 4 and 5 present the spatial distribution of NDVI values in 2011 and 2017, respectively.
Table 1

Landsat ETM+ sensor system characteristics

Date of acquisition

Sensors

Spatial resolution (m)

Spectral bands

Spectral resolution (μm)

17 March 2011

Landsat ETM+

30 × 30

3 (red)

0.63–0.69

30 × 30

4 (near IR)

0.78–0.90

16 March 2017

Landsat ETM+

30 × 30

3 (red)

0.63–0.69

30 × 30

4 (near IR)

0.78–0.90

NIR near infrared

Fig. 4

Spatial variation of NDVI values between 2011 and 2017

Fig. 5

Spatial distribution of R factor values

In detailed calculation as shown in Eq. 1, NDVI is a popular spectral index which indicates the amount, vigor, and distribution of green vegetation in the area by calculating spectral reflectance difference between red and near-infrared band of the image (Tucker 1979). Generally, NDVI ranges from − 1 to 1. The negative value indicates land cover without chlorophyll, such as cloud, rock, water, and storm. The positive value represents vegetation coverage, with the higher value for the denser coverage and healthy of green vegetation. The NDVI is a simple numerical indicator that can be calculated as Eq. 1:
$$ NDVI=\left(\mathrm{NIR}-\mathrm{RED}\right)/\left(\mathrm{NIR}+\mathrm{RED}\right) $$
(1)
where NIR is the near-infrared band (band 4, 0.76–0.90 μm), and RED is the red band (band 3, 0.63–0.69 μm).

RUSLE model description

The universal equation of soil loss RUSLE (Wischmeier and Smith 1978; Renard et al. 1997) has been widely used to predict the spatial annual average soil loss, along with for conservation goals, with effective and acceptable level of accuracy (Balasubraman et al. 2015). This spatial-mathematical model is based on thematic layers of kinetic energy of raindrops, soil portability to erosion due to its physical and chemical properties, topography, vegetation, and maintenance, respectively. The average annual soil loss per unit area and per year was quantified as per the following equation (Eq. 2) of RUSLE (Wischmeier and Smith 1978):
$$ A=R\times K\times LS\times C\times P $$
(2)

where A is the average annual soil loss in tons per ha per year (t ha−1 year−1), R is the rainfall erosivity, K is the soil erodibility, LS is the hill slope length and steepness, C is the vegetation factor, and P is the support practice. In order to standardize pixel resolution DEM and Landsat imageries resolution, all inputs and outputs for the calculation of erosion risks is in 30-pixel resolution for each sub-factors of RUSLE model.

Rainfall erosivity factor (R)

The rainfall erosivity (R) factor represents the effect of rainfall intensity on soil erosion and requires continuous, detailed data (Wischmeier and Smith 1978). R value is greatly affected by the volume, intensity, duration, and pattern of rainfall whether for single storms or a series of storms, and by the amount and rate of the resulting runoff (Farhan et al. 2013). R factor index within the RUSLE model should be obtained by multiplying the total storm energy (E) and the maximum 30-min rainfall intensity (I30) (Renard et al. 1997). Moreover, the calculation of the R values by the previous method can apply only in areas which are equipped by recording gauging stations that record the rains instantaneously. To overcome this problem, global rainfall erosivity map (spatial resolution 30 arc-seconds) was used in this study to generate the rainfall erosivity map of the study area (Panagos et al. 2017). In study area, the R values were calculated between 812.002 and 1672.98 MJ mm ha−1 year−1.

Soil erodibility factor (K)

The soils of the Safita district are divided into four different categories based on the soil type, and it reflects the physical and chemical properties of the soil. Soil erodibility factor (K) represents the susceptibility of soil or surface material to erosion, transportability of the sediment. The amount of runoff given a particular rainfall input as measured under a standard condition is also taken into account. The standard condition is the unit plot, 22.6 m long with a 9% gradient, maintained in continuous fallow, tilled up and down the hill slope (Kim 2006; Chadli 2016). The K factor rated on a scale from 0 to 1, with 0 indicating soil with least susceptibility to erosion, and 1 refers to soils which are highly susceptible to erosion by water. Air-dried soils passed through sieve were used for determining physical and chemical characteristics. The results of analyzed soil samples in Beit-kammouneh Center for Agricultural Research (BCAR) in Tartous are the basis for mapping K factor by using the following equation introduced by (Wischmeier et al. 1971; Renard et al. 1997; Panagos et al. 2014):
$$ K=\frac{2.1\times {10}^{-4}\left(12- OM\right)\times {M}^{1.14}+3.25\left(s-2\right)+2.5\left(p-3\right)}{100}\times 0.1317 $$
(3)
where M is the textural factor with M = (msilt+mvfs) × (100 − mc); mc [%] is clay fraction content (> 0.002 mm); msilt [%] is silt fraction content (0.002–0.05 mm); mvfs [%] is very fine sand fraction content (0.05–0.1 mm); OM [%] is the organic matter content; s is the average row of weighted granular diameters which are calculated by using the sieves with graduated diameters (0.25–0.5–1–2–3–5–8–10) and the following equation Eq. 4:
$$ MWD=\sum \limits_{i=1}^n\left({W}_i\times {X}_i\right) $$
(4)
where n is the order number of samples sizes, Wi is the weight of the secondary granules on the sieve as a percentage in the case of total drought, and Xi is average of diameter of the granules on the sieve. Then, the weighted granules were classified into four rows, as in Table 2. s is the soil structure class (s = 1, very fine granular; s = 2, fine granular; s = 3, medium or coarse granular; s = 4, blocky, platy, or massive); p is the permeability class (p = 1, very rapid; p = 6, very slow). By using the previous equation, the K values (t MJ−1 hmm−1) were computed and subsequently, the K factor map was designed (Fig. 6); p is the saturated hydraulic conductivity of soil which calculated using soil water characteristics calculator program depending on the percentage of individual granules involved in soil composition and soil content of organic matter. The saturated hydraulic conductivity was classified as six rows as the Table 3 indicates.
Table 2

Average row of weighted granular diameters (Wischmeier and Smith 1978)

Row

Average weighted diameter

1

< 1

2

1–2

3

2–10

4

> 10

Fig. 6

Soil erodability factor (K) map prepared from soil data

Table 3

Rows of hydraulic conductivity (Wischmeier and Smith 1978)

Row

Hydraulic conductivity (mm/day)

1

< 10

2

10–100

3

100–400

4

400–1000

5

1000–3000

6

> 3000

Topographic factor (LS)

The effect of terrain factor on soil erosion rates is expressed by the combined effect of slope length (L), slope steepness (S), and slope morphology on rill, inter-rill erosion and sediment production (Farhan and Nawaiseh 2015). Slope length (L) is the effect of slope length on erosion. The slope length is defined as the distance from the point of origin of overland flow to the point where either the slope decreases to the extent that deposition begins, or runoff water enters a well-defined channel. Thus, the soil loss per unit area increases as the slope length increases. Slope steepness (S) represents the effect of slope steepness on erosion. The effects of slope steepness have a greater impact on soil loss than slope length (Ganasri and Ramesh 2016). In this context, in the Mediterranean region, the slopes are related to lithology and soil type (Roose et al. 1993; Elaloui et al. 2017). By utilizing digital elevation model (DEM) with 30 m resolution (ASTER GDEM Validation Team 2009), LS factor in the study area was calculated depending on the following equation (Eq. 5):
$$ LS={\left( FlowAccumulation\times \frac{CellSize}{22.13}\right)}^{0.5}\times {\left(\frac{sinslope}{0.0896}\right)}^{1.3} $$
(5)
where FlowAccumulation is the grid layer of flow accumulation expressed as the number of grid cells, and CellSize is the length of a cell side. LS values of the study area are in the range of 0–21.98 as illustrated Fig. 7.
Fig. 7

Slope length and steepness factor (LS) prepared from DEM

Vegetation factor (C)

Vegetation cover plays an important role in soil erosion control spatially, owing to dissipating the rainfall energy before it reaches the soil surface, slowing down the movement of surface runoff, and promoting surface water infiltration (Hu et al. 2015; Salloum and Abdo 2016). Vegetation factor (C) is defined as the ratio of soil loss from land with specific vegetation cover to the corresponding soil loss from continuous fallow. In this context, C factor is an important factor in RUSLE since it represents the conditions that can be easily changed to reduce erosion than others factors involved (Wischmeier and Smith 1978; Beskow et al. 2009). However, C factor is one of the most complex factors because it associated with the status of vegetation that is changing rapidly. These changes are related to morphological characteristics, growth phase, deteriorating, and removing (Abdo and Salloum 2017a; Salloum and Abdo 2016b). As there was a significant changes in vegetation cover in the study area over temporal dimension of this study, NDVI distribution map of 2011 and 2017 respectively are used to map the spatial distribution of C factor utilizing the following equation (Eq. 6) (Zhou et al. 2008; Kouli et al. 2009):
$$ {C}_{factor}=\mathit{\exp}\left[-\alpha \frac{NDVI}{\left(\beta - NDVI\right)}\right] $$
(6)
where α and β parameters determine the shape of the NDVI curve. Reasonable results are produced using values of α = 2 and β = 1. Indeed, Fig. 8 shows the variation of C factor values between 2011 and 2017 and this indicates more vulnerability to soil erosion, as the area is considered as an unprotected land. However, C factor values are comparable with similar studies carried out Ozsoy and Aksoy (2015).
Fig. 8

Spatial variation of C factor derived from NDVI values between 2011 and 2017

Maintenance factor (P)

Conservation practice factor (P) is the ratio of soil loss after a specific support practice to the corresponding soil loss after up and down cultivation (Samanta et al. 2016). These practices principally affect erosion by modifying the flow pattern, grade or direction of surface runoff, and by reducing the amount and rate of runoff. For cropland, the support practices considered included contouring, terracing, strip cropping, and subsurface drainage (Renard et al. 1997). The value of the P factor ranges from 0 to 1, the value approaching to 0 indicates good conservation practice and the value approaching to 1 indicates poor conservation practice (Wischmeier and Smith 1978). Field observations supported by Fig. 9 indicate to the absence of active maintenance procedures of lands in study area. In order to neglect the P factor from soil erosion estimation, P equals to 1 as suggested by Wischmeier and Smith (1978).
Fig. 9

The sample of field checking of maintenance criteria (taken on 30 March 2017)

Result and discussion

Analyzing of vegetation patterns changes

As NDVI values represent different vegetation types and fraction of vegetation cover (Carlson and Ripley 1997; Gutman and Ignatov 1998), and in order to compare the resultant maps of vegetation cover, each NDVI values map is differentiated into four grades as > 0.2 (low), 0.2–0.4 (moderate), 0.4–0.6 (high), and < 0.6 (very high), respectively. However, these four types of NDVI densities were used to describe the vegetation changes. Overall, vegetation coverage in terms of NDVI is the highest for forests, intermediate for farmlands, and the lowest for grasslands and crops.

The spatial distribution of the four NDVI density classes was extracted to detect vegetation cover change information. Obviously, the vegetation coverage and landscape features in studied area have changed greatly and more fragmented than before. Figures 4 and 5 depict the spatial-temporal changing of NDVI indicator values in Safita area during 2011–2017. The results show that the changes are both positive by gaining and negative by losing vegetation. It can be noted that the areas in northeastern part of study area have higher NDVI change than those in the west or south.

Figures 4 and 5 also indicate that the major loss of vegetation was in the high and very high densities. Based on the deep fieldwork observations, high densities consist of olive and citrus trees, and the very high densities consist of well-grown broadleaved forests, coniferous forests, and shrubs. Moreover, these two densities are forming the first target for the loggers. Quantitatively, Table 4 summarizes the individual NDVI class change and percentage change of each one for the two years 2011 and 2017. In details, forest areas have diminished from 19.5% (65.48 km2) in 2011 to 17.2% (54.53 km2) in 2017, and olive and citrus trees areas have dwindled from 60.7% (203.3 km2) in 2011 to 58.4% (185.3 km2) in 2017. Meanwhile, the annual rate of vegetation change in 6 years reached to 0.39% and 0.37% in very high and high densities of vegetation classification, respectively. Furthermore, spatial characteristics of the vegetation restoration region are also noticeable. However, the moderate and low densities areas of sparse vegetation cover, such as senescing crops, have increased 0.4 and 4.1% continuously through monitored period.
Table 4

Vegetation coverage changes in 2011 and 2017 of Safita area

NDVI class

2011

2017

Changing (%)

NDVI dense

Area (km2)

Percentage

NDVI dense

Area (km2)

Percentage

Low

< 0.2

15.19

4.5

< 0.2

15.72

5.0

+ 0.4

Moderate

0.2–0.4

51.20

15.3

0.2–0.4

61.61

19.4

+ 4.1

High

0.4–0.6

203.30

60.7

0.4–0.6

185.31

58.4

− 2.2

Very high

> 0.6

65.48

19.5

> 0.6

54.53

17.2

− 2.3

Strictly speaking, spatially explicit patterns of vegetation cover changes derived from analysis show more considerations for the elevated impacts of human activities. The effects are realized by high deforestation rate and fires caused by enormous refugee pressure. According to the UN OCHA, there are an estimated 500,000 refugees in Tartous (Faour and Fayad 2014), of which about 189,114 refugees in the study area. Tartous Governorate, however, with a displaced population accounting for 47% of the pre-conflict population, also has an exceptionally high burden of displacement despite the smaller absolute size of the displaced population (452,000) (Doocy et al. 2015). Out of the total newly displaced people, an estimated 9000 people moved to Safita by 2017 (DOSA 2017).

Moreover, such increase of population requires additional land for horizontal expansion of agricultural and constructional development needs, combined with gradual absence of forest rangers during the war years in particular. Similar findings were found in a study by Zeleke and Hurni (2001) and Ahmed (2016). According to field observations, the positive changing of vegetation cover in moderate densities of vegetation classification can be attributed to the intensification of agricultural activities especially field crops and granular whose price has remarkably risen recently due to the accelerated demand to keep up with the nutritional requirements for refugees, thereby increasing vegetation coverage with limited biomass accumulation. However, farmlands in some degradation vegetation sites tended to substitute forests during war (Fig. 10).
Fig. 10

Spatial causes as drivers of increasing of soil erosion rates (a, b taken on 25 March 2017; c taken on 6 September 2017)

Spatial-temporal assessment of soil erosion

After identifying the changes of vegetation through NDVI, the Map Algebra tool in GIS environment was used to determine at cell-by-cell level the raster layers of RUSLE model criteria (R, K, LS, and C). The two spatial multiplication processes of previous criteria were done according to Eq. 2. This process yielded to two of the final output maps of monitoring of sites for soil erosion risks over the studied period. Figure 11 represents the final output of the spatial modeling process of soil erosion over the studied period. The soil erosion areas have shown a dramatic upward trend for the last 6 years dramatically. However, the estimated values of soil erosion range from 0 to 20.14 t ha−1 year−1 with a mean value of 10.75 t ha−1 year−1 in 2011 and from 0 to 23.18 t ha−1 year−1 with a mean value of 10.61 t ha−1 year−1 in 2017. The visual interpretation of spatial patterns changing of soil erosion for two monitored years is summarized in Table 5 and Fig. 11, where each final output map was divided into four grades according to the erosion risks. These grades are low, moderate, high, and very high areas. However, the estimated average annual soil erosion rate in Safita area before and during the war exceeds the acceptable soil erosion tolerance limits from 2 to 12 t ha−1 year−1 for the Mediterranean environments (Nearing et al. 1990; Rojo 1990; İrvem et al. 2007; Trabucchi et al. 2012; Farhan and Nawaiseh 2015).
Fig. 11

Maps of soil erosion setting showing spatial correlation between vegetation loss and acceleration of soil erosion risks between 2011 and 2017

Table 5

Soil erosion changes through 2011 and 2017 of Safita area

Soil erosion class (t ha−1 year−1)

2011

2017

Changing (%)

Rate of soil erosion class in t ha−1 year−1

Area (km2)

Percentage

Rate of soil erosion class in t ha−1 year−1

Area (km2)

Percentage

Low

< 3

189.18

59.65

< 3

87.24

27.51

− 32.14

Moderate

3–7

98.65

31.10

3–7

157.18

49.56

+ 18.45

High

7–14

25.65

8.09

7–14

66.52

20.97

+ 12.89

Very high

14–20.14

3.69

1.16

14–23.19

6.23

1.96

+ 0.80

As revealed in Table 5, not only the spatial distributions of soil erosion areas have increased, but also the annual rate of soil erosion during the time dimension of this study that has increased by 3.05 t ha−1 year−1. The estimated increase in the annual rate of soil erosion between 2011 and 2017 represents a real disaster on the level of land degradation if this outcome can be circulated to SCR which constitutes an important agricultural pillar in Syria especially during war. In details, the results have shown that high and very high grade areas of potential soil erosion have changed by + 12.89% (8.09 to 20.79%) and + 0.8% (1.16 to 1.96%), respectively. Along with low grade areas which decreased by − 32.14% (59.65 to 27.51%). The rest of the study area, moderate grade areas, have changed positively by + 18.45% (31.1 to 49.56%).

The variations of soil erosion risks areas can be attributed to the negative vegetation changes. Interestingly, the enormous monitored fluctuations in soil erosion estimations denoted to the fact that the high and very high areas of soil erosion have focused in the same areas of vegetation degradation. These results have demonstrated that the vegetation parameter has changed in study area more rapidly compared with other potential changes of the other involved spatial factors into RUSLE model. In this context, Fig. 10 illustrates the main reasonable spatial causes as drivers of increasing of soil erosion rates such as deforestation, increasing residential area, absence of maintenance factors, and frequent fires and so on.

Meanwhile, high and very high areas of soil erosion have concentrated in the north-eastern and upper-central parts of the study area, which are essentially hilly terrain with steep slope areas. Therefore, the degradation of vegetation has increased soil erosion process. However, the decrease of vegetation coverage at high slope area often results in severe soil erosion and serious ecological disturbance in the long term (Liu et al. 2009). A progressive decrease in the forest’s ability to sustain soil can be observed in areas where forest logging and sparsely vegetated areas on steep slopes have caused intense soil mobilization (Fig. 12). In this context, management criteria must be adopted to control soil erosion in areas of high to very high erosion. Special prioritization with effective adjustment and optimization of vegetation must be given to the protection of these critical places as quick as possible. Such procedure must stay a priority in these areas to mitigate impending land degradation. Appropriate vegetation types should be chosen according to site-specific and vegetation patterns conditions. In addition to reconverting vegetation to its natural condition, spatial modification must be done to steep slopes by creating the terraces with vegetation barriers. This should take place, side by side, with firmly reactivating the role of forest rangers. Determination of directions of fire line is recommended, taking into account the criteria of plant moisture, and eastern dry wind times that enhance the mass of fire each year.
Fig. 12

Soil mobilization under NDVI classification of forest

Conclusion

The indirect impacts of war in Syria on the Safita’s vegetation-soil resources have revealed the potential interactions between internally displaced people, vegetation changes, and soil erosion for a Syrian safe area during war years. On the whole, the study is motivated by the fast changes occurring in vegetation coverage during the war in Syria in the safe coastal area. It must be emphasized that negative vegetation cover changes have occurred under impact of the increasing of heating needs and settlements, along with the recurrent and premeditated fires. However, the effects of NDVI values dynamics on spatial distributions of soil erosion grades based on RUSLE model in GIS technology were evaluated. It was clear that the significant changes in vegetation cover attributed to increasing pressures created by refugees. The high and very high densities of vegetation were the most serious change. Consequently, these alarming shifts of vegetation have affected soil erosion rates dramatically. Quantitatively, the high and very high densities of vegetation coverage exposure to degradation case recorded 2.4% of area between 2011 to 2017, followed by an increase in the rates of annual soil erosion by 3.05 t ha−1 year−1 over monitored period. Besides, there was shocking expanding of the spatial grades of soil erosion risks. In this accelerated case of environmental change, the dramatic reduction of vegetation cover and soil erosion in Safita area will have inevitable consequences on the level of land degradation within a short period. Therefore, the limit ability of vegetation reconstruction must be considered with soil erosion, especially for the degraded region with high slope.

The analysis showed that unexpected pressure of refugees can also significantly and swiftly alter the area-scale environmental quality with potentially serious land implications. This study also illustrated how spatial techniques combined with geographic information system and remote sensing data can be used to approach vegetation and soil erosion dynamics in a safe zone in Syria during war years, and more generally for the acquisition of critical spatial measurements for land management in unstable country. In addition, the spatial representation of the vegetation cover change has presented useful information about the most sensitive locations to soil erosion over the period of war. Hence, the consequences of this study are valuable and necessary for local decision makers to taking necessary procedures to mitigate the erosion problem in environmental rehabilitation phase as soon as Syrian war ends.

Notes

Acknowledgments

The author gratefully thanks rector of Tartous University Prof. Dr. Issam Al-dali for unlimited support. Sincerely thanks also go to Prof. Dr. Fatina Yaseen Alshaal and Prof. Dr. Juliet Salloum for academic support, and Ms. Homam Knaj and Mrs. Rania Hassan for administrative support. The author expresses his deep and sincere thanks to the anonymous reviewer and the editor for their valuable comments, suggestions, assistance, and constructive criticism in the improvement of earlier version of the manuscript, and Ms. Wafeeq Asaad for editing the English of this manuscript.

References

  1. Abdo H, Salloum J (2017a) Mapping the soil loss in Marqya basin: Syria using RUSLE model in GIS and RS techniques. Environ Earth Sci 76(3):114.  https://doi.org/10.1007/s12665-017-6424-0 CrossRefGoogle Scholar
  2. Abdo H, Salloum J (2017b) Spatial assessment of soil erosion in Alqerdaha basin (Syria). Model Earth Syst Environ 3(1):26.  https://doi.org/10.1007/s40808-017-0294-z CrossRefGoogle Scholar
  3. Abdo HG (2017) Geo-modeling approach to predicting of erosion risks utilizing RS and GIS data: a case study of Al-Hussain Basin, Tartous, Syria. J Environ Geol 1(1):1–4Google Scholar
  4. Ahmed N (2016) Application of NDVI in vegetation monitoring using GIS and remote sensing in northern Ethiopian highlands. Abyss J Sci Technol 1(1):12–17Google Scholar
  5. Aster GDEM, Validation Team (2009) ASTER global DEM validation summary report. METI/ERSDAC, NASA/LPDAAC, USGS/EROSGoogle Scholar
  6. Balasubramani K, Veena M, Kumaraswamy K, Saravanabavan V (2015) Estimation of soil erosion in a semi-arid watershed of Tamil Nadu (India) using revised universal soil loss equation (rusle) model through GIS. Model Earth Syst Environ 1:10.  https://doi.org/10.1007/s40808-015-0015-4 CrossRefGoogle Scholar
  7. Baldacchino G, Sammut C (2016) The migration crisis: No human is illegal. The Round Table 105(2):231–233.  https://doi.org/10.1080/00358533.2015.1112092 CrossRefGoogle Scholar
  8. Barakat M, Mahfoud I, Kwyes A (2014a) Estimation of vegetation factor (C) using normalized difference vegetation index (NDVI) in lower and middle basin of Al-Kabeer Alshamali River in Lattakia/Syria. Tishreen Univ J Res Sci Stud 36(5):99–107Google Scholar
  9. Barakat M, Mahfoud I, Kwyes A (2014b) Study of soil erosion risk in the basin of Northern Al-Kabeer river at Lattakia-Syria using remote sensing and GIS techniques. Mesopot J Mar Sci 29(1):29–44Google Scholar
  10. Barati S, Rayegani B, Saati M, Sharifi A, Nasri M (2011) Comparison the accuracies of different spectral indices for estimation of vegetation cover fraction in sparse vegetated areas. Egypt J Remote Sens Space Sci 14(1):49–56.  https://doi.org/10.1016/j.ejrs.2011.06.001 CrossRefGoogle Scholar
  11. Beskow S, Mello CR, Norton LD, Curi N, Viola MR, Avanzi JC (2009) Soil erosion prediction in the Grande River Basin, Brazil using distributed modeling. Catena 79(1):49–59.  https://doi.org/10.1016/j.catena.2009.05.010 CrossRefGoogle Scholar
  12. Biswas AK (2002) Scientific assessment of the long-term environmental consequences of war. In: Austin JE, Bruch CE (eds) The environmental consequences of war: Legal, economic, and scientific perspectives. Cambridge University Press.  https://doi.org/10.1080/07900627.2015.1023892
  13. Carlson TN, Ripley DA (1997) On the relation between NDVI, fractional vegetation cover, and leaf area index. Remote Sens Environ 62(3):241–252.  https://doi.org/10.1016/S0034-4257(97)00104-1 CrossRefGoogle Scholar
  14. Casana J, Panahipour M (2014) Satellite-based monitoring of looting and damage to archaeological sites in Syria. J Eastern Mediterr Archaeol Heritage Stud 2(2):128–151.  https://doi.org/10.5325/jeasmedarcherstu.2.2.0128 CrossRefGoogle Scholar
  15. Chadli K (2016) Estimation of soil loss using RUSLE model for Sebou watershed (Morocco). Model Earth Syst Environ 2:1–10Google Scholar
  16. Chen PY, Fedosejevs G, Tiscareno-Lopez M, Arnold JG (2006) Assessment of MODIS-EVI, MODIS-NDVI and VEGETATION-NDVI composite data using agricultural measurements: an example at corn fields in western Mexico. Environ Monit Assess 119:69–82.  https://doi.org/10.1007/s10661-005-9006-7 CrossRefGoogle Scholar
  17. Demirci A, Karaburun A (2012) Estimation of soil erosion using RUSLE in a GIS framework: a case study in the Buyukcekmece Lake watershed, northwest Turkey. Environ Earth Sci 66(3):903–913CrossRefGoogle Scholar
  18. Diodato N, Bellocchi G (2008) Modelling vegetation greenness responses to climate variability in a Mediterranean terrestrial ecosystem. Environ Monit Assess 143:147–159.  https://doi.org/10.1007/s10661-007-9964-z CrossRefGoogle Scholar
  19. Doocy S, Lyles E, Delbiso TD, Robinson CW (2015) Internal displacement and the Syrian crisis: an analysis of trends from 2011–2014. Confl Heal 9(1):33.  https://doi.org/10.1186/s13031-015-0060-7 CrossRefGoogle Scholar
  20. DOSA (2017) Official records of the Directorate of Social Affairs in Tartous. Ministry of social affairs, SyriaGoogle Scholar
  21. Elaloui A, Marrakchi C, Fekri A, Maimouni S, Aradi M (2017) USLE-based assessment of soil erosion by water in the watershed upstream Tessaoute (Central High Atlas, Morocco). Model Earth Syst Environ 1–13Google Scholar
  22. Elwell HA, Stocking MA (1976) Vegetal cover to estimate soil erosion hazard in Rhodesia. Geoderma 15(1):61–70.  https://doi.org/10.1016/0016-7061(76)90071-9 CrossRefGoogle Scholar
  23. Fagbohun BJ, Anifowose AY, Odeyemi C, Aladejana OO, Aladeboyeje AI (2016) GIS-based estimation of soil erosion rates and identification of critical areas in Anambra sub-basin, Nigeria. Model Earth Syst Environ 2(3):159.  https://doi.org/10.1007/s40808-016-0218-3 CrossRefGoogle Scholar
  24. Faour G, Fayad A (2014) Water environment in the coastal basins of Syria—assessing the impacts of the war. Environl Processes 1(4):533–552.  https://doi.org/10.1007/s40710-014-0043-5 CrossRefGoogle Scholar
  25. Farhan Y, Zregat D, Farhan I (2013) Spatial estimation of soil erosion risk using RUSLE approach, RS, and GIS techniques: a case study of Kufranja Watershed, Northern Jordan. J Water Resour Prot 5(12):1247.  https://doi.org/10.4236/jwarp.2013.512134 CrossRefGoogle Scholar
  26. Farhan Y, Nawaiseh S (2015) Spatial assessment of soil erosion risk using RUSLE and GIS techniques. Environ Earth Sci 74(6):4649–4669.  https://doi.org/10.1007/s12665-015-4430-7 CrossRefGoogle Scholar
  27. Ferreira V, Panagopoulos T (2014) Seasonality of soil erosion under Mediterranean conditions at the Alqueva Dam watershed. Enviro Manag 54:67–83.  https://doi.org/10.1007/s00267-014-0281-3 CrossRefGoogle Scholar
  28. Ganasri BP, Ramesh H (2016) Assessment of soil erosion by RUSLE model using remote sensing and GIS - A case study of Nethravathi Basin. Geosci Front 7(6):953–961CrossRefGoogle Scholar
  29. Gleick P (2014) Water, drought, climate change, and conflict in Syria. Am Meteorol Soc 6:331–340.  https://doi.org/10.1175/WCAS-D-13-00059.1 CrossRefGoogle Scholar
  30. Hu CJ, Fu BJ, Liu GH, Jin TT, Guo L (2010) Vegetation patterns influence on soil microbial biomass and functional diversity in a hilly area of the Loess Plateau, China. J Soils Sediments 10:1082–1091.  https://doi.org/10.1007/s11368-010-0209-3 CrossRefGoogle Scholar
  31. Hu Y, Tian G, Mayer AL, He R (2015) Risk assessment of soil erosion by application of remote sensing and GIS in Yanshan Reservoir catchment, China. Nat Hazards 79(1):277–289.  https://doi.org/10.1007/s11069-015-1841-4 CrossRefGoogle Scholar
  32. İrvem A, Topaloğlu F, Uygur V (2007) Estimating spatial distribution of soil loss over Seyhan River Basin in Turkey. J Hydrol 336(1–2):30–37CrossRefGoogle Scholar
  33. Jaafar HH, Zurayk R, King C, Ahmad F, Al-Outa R (2015) Impact of the Syrian conflict on irrigated agriculture in the Orontes Basin. Int J Water Resour Dev 31(3):436–449.  https://doi.org/10.1080/07900627.2015.1023892 CrossRefGoogle Scholar
  34. Jiang Z, Huete AR, Didan K, Miura T (2008) Development of a two-band enhanced vegetation index without a blue band. Remote Sens Environ 112(10):3833–3845.  https://doi.org/10.1016/j.rse.2008.06.006 CrossRefGoogle Scholar
  35. Kim HS (2006) Soil erosion modelling using RUSLE and GIS on the IMHA watershed, South Korea. Doctoral dissertation. Colorado State University USAGoogle Scholar
  36. Kouli M, Soupios P, Vallianatos F (2009) Soil erosion prediction using the revised universal soil loss equation (RUSLE) in a GIS framework, Chania, Northwestern Crete, Greece. Environ Geol 57(3):483–497.  https://doi.org/10.1007/s00254-008-1318-9 CrossRefGoogle Scholar
  37. Lahav G (2016) The global challenge of the refugee exodus. Curr Hist 115(777):10Google Scholar
  38. Lan Y (2009) Development of an integration sensor and instrumentation system for measuring crop conditions. CIGR Journal, Agricultural Engineering InternationalGoogle Scholar
  39. Liu LY, Jing X, Wang JH, Zhao CJ (2009) Analysis of the changes of vegetation coverage of western Beijing mountainous areas using remote sensing and GIS. Environ Monit Assess 153:339–349.  https://doi.org/10.1007/s10661-008-0360-0 CrossRefGoogle Scholar
  40. Ma M, Wang J, Wang X (2006) Advance in the inter-annual variability of vegetation and its relation to climate based on remote sensing. J Remote Sens 10(3):421–431Google Scholar
  41. Machlis G, Hanson T (2008) Warfare ecology. Bioscience:729–736.  https://doi.org/10.1007/978-94-007-1214-0_5
  42. Marx A (2016) Detecting urban destruction in Syria: a Landsat-based approach. Remote Sens Appl: Soc Environ 4:30–36.  https://doi.org/10.1016/j.rsase.2016.04.005 CrossRefGoogle Scholar
  43. Mitri G, Nader M, Van der Molen I, Lovett J (2014) Evaluating exposure to land degradation in association with repetitive armed conflicts in North Lebanon using multi-temporal satellite data. Environ Monit Assess 186(11):7655–7672.  https://doi.org/10.1007/s10661-014-3957-5 CrossRefGoogle Scholar
  44. Mohammed S, Kbibo I, Alshihabi O, Mahfoud E (2016) Studying rainfall changes and water erosion of soil by using the WEPP model in Lattakia, Syria. J Agric Sci 61(4):375–386.  https://doi.org/10.2298/JAS1604375M CrossRefGoogle Scholar
  45. Morrissey J (2014) Environmental change and human migration in sub-Saharan Africa. In People on the move in a changing climate (pp. 81-109). Springer Netherlands.  https://doi.org/10.1007/978-94-007-6985-4_4
  46. Mubareka S, Ehrlich D (2010) Identifying and modelling environmental indicators for assessing population vulnerability to conflict using ground and satellite data. Ecol Indic 10(2):493–503.  https://doi.org/10.1016/j.ecolind.2009.09.002 CrossRefGoogle Scholar
  47. Müller MF, Yoon J, Gorelick SM, Avisse N, Tilmant A (2016) Impact of the Syrian refugee crisis on land use and transboundary freshwater resources. Proc Natl Acad Sci 113(52):14932–14937.  https://doi.org/10.1073/pnas.1604566113 CrossRefGoogle Scholar
  48. Nearing MA, Deer-Ascough L, Laflen JM (1990) Sensitivity analysis of the WEPP hillslope profile erosion model. Trans ASAE 33(3):839–849.  https://doi.org/10.13031/2013.31409 CrossRefGoogle Scholar
  49. Ozsoy G, Aksoy E (2015) Estimation of soil erosion risk within an important agricultural sub-watershed in Bursa, Turkey, in relation to rapid urbanization. Environ Monit Assess 187(7):419.  https://doi.org/10.1007/s10661-015-4653-9 CrossRefGoogle Scholar
  50. Pal S (2016) Identification of soil erosion vulnerable areas in Chandrabhaga river basin: a multi-criteria decision approach. Model Earth Syst Environ 2(1):5.  https://doi.org/10.1007/s40808-015-0052-z CrossRefGoogle Scholar
  51. Panagos P, Meusburger K, Ballabio C, Borrelli P, Alewell C (2014) Soil erodibility in Europe: a high-resolution dataset based on LUCAS. Sci Total Environ 479:189–200.  https://doi.org/10.1016/j.scitotenv.2014.02.010 CrossRefGoogle Scholar
  52. Panagos P, Borrelli P, Poesen J, Ballabio C, Lugato E, Meusburger K et al (2015) The new assessment of soil loss by water erosion in Europe. Environ Sci Pol 54:438–447.  https://doi.org/10.1016/j.envsci.2015.12.013 CrossRefGoogle Scholar
  53. Panagos P, Borrelli P, Meusburger K, Yu B, Klik A, Lim KJ et al (2017) Global rainfall erosivity assessment based on high-temporal resolution rainfall records. Sci Rep 7(1):4175.  https://doi.org/10.1038/s41598-017-04282-8 CrossRefGoogle Scholar
  54. Propastin PA, Kappas M, Muratova NR (2008) A remote sensing based monitoring system for discrimination between climate and human-induced vegetation change in Central Asia. Manag Environ Qual 19(5):579–596.  https://doi.org/10.1108/14777830810894256 CrossRefGoogle Scholar
  55. Rastmanesh F, Moore F, Kharrati-Kopaei M, Behrouz M (2010) Monitoring deterioration of vegetation cover in the vicinity of smelting industry, using statistical methods and TM and ETM+ imageries, Sarcheshmeh copper complex, Central Iran. Environ Monit Assess 163:397–410.  https://doi.org/10.1007/s10661-009-0843-7 CrossRefGoogle Scholar
  56. Renard KG, Foster GR, Weesies GA, McCool DK, Yoder DC (1997) Predicting soil erosion by water: a guide to conservation planning with the Revised Universal Soil Loss Equation (RUSLE). U.S. Department of Agriculture Handbook No 703 p 404Google Scholar
  57. Reuveny R, Mihalache-O’Keef A, Li Q (2010) The effect of warfare on the environment. J Peace Res 47(6):749–761CrossRefGoogle Scholar
  58. Rojo L (1990) Plan Nacional de Restauracio’n Hidrolo’gico-Forestal y Control de la Erosio’n, Tomo I. Memoria; Tomo II. Mapas. ICONA, MadridGoogle Scholar
  59. Roose E, Kabore V, Guenat C (1993) Le zaï: fonctionnement, limites et amélioration d’une pratique traditionnelle africaine de rehabilitation de la végétation et de la productivité des terres degrades en région soudano-sahélienne (Burkina Faso). Cahiers-ORSTOM Pédol 28(2):159–173Google Scholar
  60. Saleh HA, Allaert G, Sutter RD (2016) Towards efficient use of water resources management: a case study of the Syrian coastal region. Int J Water 10(1):28–54.  https://doi.org/10.1504/IJW.2016.073740 CrossRefGoogle Scholar
  61. Salloum J, Abdo H (2016) Statistical modeling of conservation the vegetation of the land in Alqadmous area from rainfall erosion. Tishreen Univ J Res Sci Stud-Arts Human Ser 38(3):667–683Google Scholar
  62. Samanta S, Koloa C, Pal DK, Palsamanta B (2016) Estimation of potential soil erosion rate using RUSLE and E30 model. Model Earth Syst Environ 2(3):149.  https://doi.org/10.1007/s40808-016-0206-7 CrossRefGoogle Scholar
  63. Swift J (1973) Disaster and a Sahelian nomad economy. In: Dalby D, Church R (eds) Drought in Africa: Report of the 1973 symposium. University of London, LondonGoogle Scholar
  64. Trabucchi M, Puente C, Comin FA, Olague G, Smith SV (2012) Mapping erosion risk at the basin scale in a Mediterranean environment with opencast coal mines to target restoration actions. Reg Environ Chang 12(4):675–687CrossRefGoogle Scholar
  65. United Nations Office for the Coordination of Humanitarian Affairs (2017) About the Crisis, OCHA. 〈http://www.unocha.org/syrian-arab-republic/syria-country-profile/about-crisis〉, (accessed 29. 1.17).
  66. Wei W, Chen LD, Fu BJ, Lv YH, Gong J (2009) Responses of water erosion to rainfall extremes and vegetation types in a loess semiarid hilly area, NW China. Hydrol Process 23:1780–1791.  https://doi.org/10.1002/hyp.7294 CrossRefGoogle Scholar
  67. Wischmeier WH, Smith DD (1978) Predicting rainfall erosion losses, a guide to conservation planning. USDA Handb. 537. U.S. Gov. Print. Off., Washington, DCGoogle Scholar
  68. Wischmeier WH, Johnson CB, Cross BV (1971) A soil erodibility nomograph for farmland and construction sites. J Soil Water Conserv 26:189–192Google Scholar
  69. Wu X, Wang X (2011) Spatial influence of geographical factors on soil erosion in Fuyang County, China. Procedia Environ Sci 10:2128–2133CrossRefGoogle Scholar
  70. Yang Y, Zhu J, Zhao C, Liu S, Tong X (2011) The spatial continuity study of NDVI based on kriging and BPNN algorithm. Math Comput Model 54(3):1138–1144.  https://doi.org/10.1016/j.mcm.2010.11.046 CrossRefGoogle Scholar
  71. Zeleke G, Hurni H (2001) Implications of land use and land cover dynamics for mountain resource degradation in the northwestern Ethiopian Highlands. Mt Res Dev 21(2):184–191. https://doi.org./10.1659/0276-4741(2001)021[0184:IOLUAL]2.0.CO;2Google Scholar
  72. Zeng B, Yang TB (2008) Impacts of climate warming on vegetation in Qaidam area from 1990 to 2003. Environ Monit Assess 144:403–417.  https://doi.org/10.1007/s10661-007-0003-x CrossRefGoogle Scholar
  73. Zhou ZC, Shangguan ZP, Zhao D (2006) Modeling vegetation coverage and soil erosion in the Loess Plateau Area of China. Ecol Model 198(1):263–268.  https://doi.org/10.1016/j.ecolmodel.2006.04.019 CrossRefGoogle Scholar
  74. Zhou P, Luukkanen O, Tokola T, Nieminen J (2008) Effect of vegetation cover on soil erosion in a mountainous watershed. Catena 75(3):319–325.  https://doi.org/10.1016/j.catena.2008.07.010 CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Geography Department, Faculty of Arts and HumanitiesTartous UniversityTartousSyria

Personalised recommendations