Variations in Vegetation Structure , Species Dominance and Plant Communities in South of the Eastern Desert-Egypt

For two successive years, the floristic diversity and vegetation composition in the southern part of the Eastern Desert of Egypt were investigated through four transects (3 crossing the Eastern Desert and one along the Red Sea). The data collected from 142 stands covering the study area included the species composition, functional groups, chorology and occurrences (Qvalues). A total of 94 plant species belonging to 33 different families were recorded, with Asteracea, Zygophyllaceae, Fabaceae, Poaceae, Chenopodiaceae and Brassicaceae as the largest families. Shrubs represented the largest functional group (39.4%), while perennial herbs represented the smallest ones (12.8%). Species occurrence (Q-value) revealed that Zilla spinosa, Acacia tortilis subsp. raddiana, Morettia philaeana, Caroxylon imbricatum, Zygophyllum coccineum and Citrullus colocynthis had wide ecological range of distribution (dominant species, Q-values ≥ 0.2). Saharo-Arabian chorotype was highly represented (72.6 %) in the flora of this area, eventually as mono, bi or pluriregional. Classification of the data set yielded 7 vegetation groups included: (A) Zilla spinosa-Morettia philaeana, (B1) Zilla spinosa-Citrullus colocynthis-Morettia philaeana, (B2) Zilla spinosa, (C1) Zygophyllum album-Tamarix nilotica, (C2) Zygophyllum coccineum-Tamarix nilotica, (D1) Zilla spinosa-Zygophyllum coccineum and (D2) Zilla spinosa-Acacia tortilis subsp. raddiana-Tamarix aphylla-Balanites aegyptiaca. Certain vegetation groups were assigned to one or more transects. Detrended Correspondence Analysis (DCA) revealed that electrical conductivity, sodium, potassium, calcium, magnesium, chlorides, moisture content, sulphates, pH, organic matter and gravel were the soil variables that affect the species distribution in this study.


Introduction
The Eastern Desert of Egypt occupies about 223,000 km 2 , i.e. 21% of the total area of Egypt.It is characterized by two main ecological units, the Red Sea costal land and the inland desert with its wadis.In Egypt, like in the other arid lands, the desert vegetation is characterized by openness and is composed of a permanent framework of perennials, the interspaces of which may be occupied by ephemerals and their duration depends on the irregular rainfall and soil thickness (Zahran and Willis, 2009).The vegetation study areas as in northeastern Brazil (Da Costa et al., 2007), in Mount Hymettus of central Greece (Gouvas and Theodoropoulos, 2007), in Khulais region of western Saudi Arabia (Al-Sherif et al., 2013) and in Alborz Mountains of Iran (Mahdavi et al., 2013).
One approach to addressing the complexity of desert vegetation is functional analysis.Plant species can be classified into functional groups based on a variety of characteristics.Functional groups have been defined as sets of species showing either similar responses to the environment or similar effect on major ecosystem processes (Gitay and Noble, 1997).Each functional group potentially will partition the environmental gradient differently (Austin, 1990;Dale, 1998;Lyon and Sagers, 2003;Smith and Huston, 1989).
This study aimed to (1) determine the spatial distribution patterns of the recorded species in terms of plant functional groups, (2) assess the soil factors which controlle the vegetation and to identify the regional plant communities, and (3) analyze the floristic variations between the northern and southern parts of the Eastern Desert.

Study area
The study area covered nearly the southern quarter of the Eastern Desert (about 54,500 km 2 ) between 26 o 45' and 24 o 1' N latitudes and 32 o 45' and 35 o 00' E longitudes (Fig. 1) It covered the area between Qena Governorate until Aswan Governorate on the Nile Valley and from Safaga until Berenice on the Red Sea coast.According to Zahran and Willis (2009), this area covered three desert types: (1) The limestone desert (Assiut-Qena Desert), (2) The sandstone desert (Idfu-Kom Ombo Desert), and (3) The Red Sea coastal plain.Detailed studies on the geology, geomorphology, topography and lithology have been documented by Said (1962), Abu Al-Izz (1971), and Zahran and Willis (2009).
The Egyptian desert is among the most arid parts of the world characterized by extreme aridity and high temperature.Available climatic records over the period 2003-2012 in four meteorological stations (Qena, Safaga, Aswan and Marsa Alam) demonstrated that the average monthly temperature ranged between 14.9°C in January (minimum) and 33.6°C (maximum) in July.Rainfall occurs only in winter and is due to random cloudbursts, a general feature in arid desert: rain may occur once every several years.Annual average rainfall records (over 30 years) showed notable decrease along north-south direction: 5.3 mm/year in Qene in the north along the Nile Valley to 3.4 mm/year in Quseir in the south along the Red Sea coast (Abd El-Ghani, 1998).Averages of relative humidity reached to the maximum of 51.5% and 52.7% (in December), while its minimum reached 25.6% and 32.4% (in June) for Mersa Alam andSafaga (Abd El-Ghani, 1998, Abd El-Ghani et al., 2013a).

Data collection and vegetation analysis
Between 2011 and 2012, vegetation sampling was performed in the study area using 4 transects representing the 3 desert types (Fig. 1).One hundred and forty-two georeferenced (using GPS model Garmin eTrex HC) randomly selected stands (20 × 30 m) were selected along the four transects to represent apparent variations in the physiognomy of vegetation and in the physiographic features.The sandstone desert included (T1) which comprised of Aswan-Berenice road (300 km; 24°05' -24º00' N and 32º55' -35º24' E); Wadi Kharit ( 250 Correspondingly, in each transect plant species were segregated into four a priori defined functional groups: tree layer species, shrub layer species, and herb species.The presence/absence for each species was recorded in the studied stands, and a count-floristic list was obtained. The number of species within each functional group category was expressed as a percentage of total number of species in each transect.Analysis of phytogeographical ranges was carried out according to White and Léonard (1991).Taxonomic nomenclature and functional groups categorizing was according to Täckholm (1974), Boulos (1995Boulos ( , 1999Boulos ( , 2000Boulos ( , 2002Boulos ( , 2005) ) and El Hadidi andFayed (1978, 1995).Voucher specimens of each species were collected, and identified at the Herbaria of Assiut University (ASTU) and Cairo University (CAI), where they were deposited.
The degree of occurrence of each species was determined using the Q-value (Danin et al., 1985) as follows: Q= number of entries of a species X total number of species/13,348 (total number of entries).The Q-values and occurrences were categorized as follows: D=dominant, Q-value ≥ 0.2; VC=very common, Qvalue 0.1-0.199;C= common, Q-value 0.05-0.099;O=occasional or rare species, Q-value 0.01-0.049and S= sporadic or very rare, Q-value ≤ 0.01.
A floristic presence-absence data matrix of 142 stands and 94 species was subjected to classification by cluster analysis of the program Community Analysis Package (CAP) version 1.2 (Henderson and Seaby, 1999) using squared Euclidean distance dissimilarity matrix with minimum variance (also called Ward's method) as agglomeration criterion (Orlóci, 1978).The resulted vegetation groups (plant communities) were named after the dominant species that have the highest presences percentages in the stands of this group.
Detrended Correspondence Analysis (DCA) ordination based on species presence-absence data for each species was performed to examine patterns in species composition among species of different vegetation groups.The relationship between the vegetation and soil variables were assessed by calculating the simple linear correlation coefficient (r) between the DCA axes (reflect the vegetation gradient) and the soil variables.
One Way Analysis of Variance was applied to examine the statistical differences between the functional groups.

Soil sampling and analysis
Soil samples (0-50 cm depth) were collected at 3 random points from each stand as a profile (composite samples).These samples were then air-dried, thoroughly mixed, and pass through a 2 mm sieve to get rid of gravel and boulders.The weight of gravel in each stand was determined and expressed as a percentage of the total weight of the soil sample.The soil texture was determined using the sieve method; the amount of each fraction (sand, silt and clay) was expressed as percentage of the original weight used (Ryan et al., 1996).Soil moisture content was estimated by drying at 105°C; then the percentage of soil moisture was calculated based on dry weight of the soil (Kapur and Govil, 2000).The soil portion less than 2 mm in size was kept for chemical analysis according to Jackson (1967) and Allen and Stainer (1974).Soil water extracts (1:5) were prepared for determination of electrical conductivity (EC) using conductivity meter, and pH using a glass electrode pHmeter.Organic matter (OM) was determined using the Walkely and Black rapid titration (Black, 1979).Sodium and potassium were determined by flamephotometer.Calcium and magnesium were estimated by titration against EDTA (ethylenediamine dihydrogen tetraacetic acid) using ammonium purpurate and eriochrome black T as indicators (Jackson, 1967).Chlorides were determined by direct titration against AgNO3 using potassium chromate as an indicator, and bicarbonates by direct titration against HCl using methyl orange as indicator.Sulphates were determined by a turbidemetric technique with barium chloride and acidic sodium chloride solution using spectrophotometer (Model 1200) according to Bardsley and Lancaster (1965).

Classification of the vegetation
Application of classification using cluster analysis to the floristic presence-absence data matrix of the study area yielded 7 vegetation groups (Tab.2, Fig. 3).Each of the identified vegetation group will be named after the dominant species (i.e., highest presence percentages).Notably, none of the recorded species occurred in all the identified groups.Apart from coarse sand, clay and bicarbonates, the other thirteen (out of total of 16) measured soil variables showed significant differences (p < 0.05, 0.01) between the vegetation groups (Tab.3).

Group (B1): Zilla spinosa-Citrullus colocynthis-Morettia philaeana group
The 18 stands of this group (26 species) were located along Wadi Natash, W. Kharit and El-Shekh El-Shazly-Marsa Alam road (T3).Soil contents of gravels, fine sand, OM and pH were higher than the total means.The lowest contents were recorded in Na +2 and HCO3 -.Beside the dominants, Acacia tortilis subsp.and Senna italica were the co-dominants.Some species were confined to this group such as Chenopodium album and Filago desertorum.
Group (C1): Zygophyllum album-Tamarix nilotica group Most stands of this group (41 stands, 32 species) were located along the Red Sea coast transect (T4) between Marsa Alam and Qusier, and occurred on saline soil with soluble anions and cations contents higher than the groups (A, B1, B2, D1 and D2).The dominant species of this group, together with the co-dominants Nitraria retusa and Limonium axillare exhibited the saline nature of this group.Certain species showed consistency to this group such as Aeluropus littoralis, Arthrocnemum macrostachyum, Avicennia marina.
Group (C2): Zygophyllum coccineum-Tamarix nilotica group This group (8 stands) was the least diversified (19 species) among others.The stands of this group were mainly located in T4 (Qusier-Safaga transect) along the Red Sea coast occurred on saline soil with the highest silt, clay, electric conductivity, water content and all the examined ions.However, it recorded the lowest pH and coarse sand content.The co-dominant species included Phragmites australis, Nitraria retusa, Limonium axillare and Zygophyllum album.Four weed species (Chenopodium murale, Cyperus rotundus, Leptochloa fusca and Sonchus oleraceus) were recorded among the 6 confined species to this group.
Group (D1): Zilla spinosa-Zygophyllum coccineum group This group of stands (31) was the most diversified (53 species) among other groups, and collected from three different transects (T1, T2 and T4) found on soil in rich in gravels and poor in silt content.The other soil factors had intermediate position amongst the other groups.The co-dominant species included Caroxylon imbricatum, Lotus hebranicus and Ochradenus baccatus.Twelve species showed consistency to this group such as Acacia nilotica, Moringa peregrina, Ziziphus spina-christi (trees), Atriplex leucoclada, Fagonia bruguieri (shrubs), and Dichanthum annulatum, Imperata cylindrica (herbs).
Group (D2): Zilla spinosa-Acacia tortilis subsp.-Tamarixaphylla-Balanites aegyptiaca group This group (19 stands, 20 species) was characterized by the combination of the dominant species, mostly located in Wadi Gimal and its tributaries (T3) on a soil rich in fine sand, silt, pH and K + and poor in Mg +2 and water contents.The co-dominants of this group had low presence values such as Zygophyllum coccineum, Pulicaria undulata and Calotropis procera.Two species were confined to this group, Capparis decidua and Salvadora persica.

Ordination of the vegetation
Analysis of 142 stands along axes 1 and 2 (eigenvalues 0.707 and 0. 497, respectively) by DCA confirms the classification results, where the 7 vegetation groups were also segregated (Fig. 4).Linear response models were dropped because gradients along the first two axes were longer than 4 SD units (Jongman et al., 1987).The length of gradient represented by axis 1 was > 9 SD, indicating a complete turnover in species composition along this gradient.Therefore, DCA was the appropriate ordination method or indirect gradient analysis to be used.The four DCA axes explained 5.3%, 3.7%, 2.8% and 2.6% of the total variation in the species data, respectively.This low percentage of variance explained by the axes was attributed to the many zero values in the vegetation data set.It can be observed that the eigenvalue for the first DCA axis was high, indicating that it captured the greater proportion of the variation in species composition among stands.It is clear that group C1 occupied the positive end of the first DCA axis, while groups B1 and B2 occupied the negative end.This arrangement may explain a gradient of increasing soil salinity and moisture content (Tab.4), where stands of group C1 were located along the Red Sea coast transect, while B1 and B2 in the inland desert of Wadi Gimal-Aswan-Wadi Kharit transect.The first DCA axis was positively correlated with electrical conductivity (r=0.297),sodium (r=0.342),potassium (r=0.307),calcium (r=0.296),magnesium (r=0.318),chlorides (r= 0.217), moisture contents (r=0.418) and sulfates (r=0.612), and negatively with pH (r=-0.167) and gravels (r=-0.249).The second axis was positively correlated with sulfates (r=0.172) and organic matter (r=0.218).

Comparison between northern and southern parts of Eastern Desert
Tab. 5 displayed the floristic composition between two geographically distant (253 km) parts (northern and southern) of the Eastern Desert.Whereas the southern part represented by the 4 transects included in this study, the northern part (c.28,800 km 2 ; 30 o 05' -28 o 21' N and 31 o 20'-33 o 50' E) included three transects; Cairo-Suez (T1N; 112 species), Korimat-Zaafarana (T2N; 111 species) and Sheikh Fadl-Ras Gharib (T3N; 54 species) mainly in the limestone part of this desert (Abdel-Aleem, 2013).Altogether, 60 species were in common, 103 species confined to the northern part, and 34 to the southern part.
Four trees: Acacia tortilis subsp.raddiana, Tamarix aphylla, T. nilotica and Calotropis procera exhibited a wide range of distribution as they recorded in the northern and southern parts.While 9 tree species confined to the southern part, and do not penetrate northwards (e.g., Avicennia marina, Hyphaene thebaica, Balanites aegyptiaca, Moringa peregrina), the northern part devoid of any characteristic tree species.Twentyeight shrubby species were recorded in both areas and included amongst others: Zilla spinosa, Zygophyllum coccineum, Caroxylon imbricatum, Suaeda monoica, Zygophyllum album and Pulicaria incisa.

Discussion
The classification and ordination analyses proposed that the vegetation of the study area can be divided into 7 major vegetation groups (plant communities).The members of each pair of groups are, in some cases, linked together by having one of the dominant species in common.It can be noted that certain vegetation groups characterized one or more of the studied transects; group (A) in Idfu-Mersa Alam transect (T2), groups (B1), (B2) and (D2) in Aswan-Kharit-Gimal transct (T3), groups (C1) and (C2) in Qusier-Safaga transect along Red Sea coast (T4), and group (D1) was widely distributed in the study area including T1, T2 and T4.It can be noted that the salt-tolerant plant Tamarix nilotica characterized vegetation group (C1) and (C2) form hillocks of huge sizes, and vigorously growing southwards (Springuel et al., 1991) representing the natural climax community type of the desert wadis with deep deposits and an underground water reserve (Kassas and Zahran, 1962).Tamarix has been identified as a major cause of salt accumulation on the soil surface (Springuel and Ali, 1990), and concentrating a high amount of sodium chloride in specialized glands within its leaves (Bosabalidis, 1992).In addition, there is a relationship between the amount of Tamarix litter and the electric conductivity of soil (Briggs et al., 1993).Meanwhile, the lower number of recorded species in vegetation group (C1) inhabiting the coastal plains of the Red Sea may be related to its high soil salinity.Such salinity stress on floristic diversity in the study area and related areas was reported by Moustafa and Klopatek (1995) and Shaltout et al. (1997).Most of the identified vegetation groups have very much in common with that recorded in some wadis of the Eastern Desert (Salama et al., 2012(Salama et al., , 2013)), Western Desert (Abd El-Ghani, 2000; Bornkamm and Kehl, 1990), in south Sinai region (Moustafa and Zaghloul, 1996) and in northwestern Negev, Israel (Tielbörger, 1997).
In extreme deserts, as in the study area, the plant growth is triggered mainly by rain, and thus is as scarce and unpredictable as the precipitation itself.Vegetation develops in 'contracted mode' (Monod, 1954) only in habitats receiving runoff water including wadis, depressions and channels-contracted desert (Shmida, 1985).This highly dynamic vegetation is neither permanent nor seasonal, but is accidental (Bornkamm, 2001;Bullard, 1997;Kassas, 1966).The vegetation structure in the study area is relatively simple, in which the species have to withstand the harsh environmental conditions.This it can be reflected by the presence of several highly adapted, drought-resistant species.The floristic diversity of the study area included 94 species of the vascular plants (67 perennials and 27 annuals) indicating the predominance of perennials.Asteraceae, Fabaceae, Poaceae, Zygophyllaceae and Chenopodiaceae were the species-rich families which formed the major component of the flora.The first three families represent the most common in the Mediterranean North African flora (Quézel, 1978;White, 1993).These findings were in line with those of Salama et al. (2012Salama et al. ( , 2013) ) in the Eastern Desert, and Abd El-Ghani and Fahmy (1998) in south Sinai, and Salama et al. (2005) along the western Mediterranean coast.
Chorological analysis revealed that the Saharo-Arabian element (37.2% monoregional, 28.7% biregional, 11.7% as pluriregional floras) forms the major component of the floristic structure along the four transects.That is because the study area lies within the Saharo-Arabian region of the Holarctic Kingdom (White, 1993).The results were in agreement with those of El-Demerdash et al. (1990), Fossati et al. (1998) and Salama et al. (2012) who concluded that plants of Saharo-Arabian region constituting the shrub layer as good indicators for desert environmental conditions, while species of Mediterranean origin (either mono, bi or pluriregional) flourish in more mesic conditions.
Comparing the results of floristic diversity in the study area (south of the Eastern Desert) with that in the northern part (Abdel-Aleem, 2013) indicated that 60 species were in common, 103 confined to the northern part, while 34 species were consistent to the southern part.So, the floristic diversity in the northern part is three times higher than that of the southern part of the Eastern Desert, which may be attributed to the mild climatic conditions prevailing in the north.Also, increasing the aridity southwards plays a paramount role in reducing floral diversity.On the other hand, 60% of the northern vegetation (not present in south) was represented as annual herbs.Decreased numbers of annuals in the southern part of the Eastern Desert can be attributed to the environmental aridity and thermal continentality which increases from north to south (Abd El-Ghani, 1998).
Vast areas in the Egyptian deserts (Western, Eastern, and Sinai) were subjected to land reclamation due to increased population growth (Biswas, 1993).In the study area, agricultural processes were practiced in the deltaic parts of several wadis such as Wadi Kherit, W. Natash and W. El-Shikh.As the land reclamation processes entail an almost complete change of the environmental factors, several common weeds of the agro-ecosystem were recorded (e.g., Cynodon dactylon, Malva paviflora, Dicanthiun annulatum, Cyperus rotundus, Sonchus oleraceus and Chenopodium murale).Thus, weeds find the new conditions favorable for their growth.Close to the boundaries of the desert in this study, xerophytic species naturally grow among the weeds of the cultivation.This indicated that these species are native to the natural desert vegetation and can remain after the reclamation process.Therefore, the reclaimed lands found at the desert outskirts can be considered as transitional areas of the succession process between the old cultivated lands and that of the desert (Abd El-Ghani et al., 2013b;Shaheen, 2002).
As for species abundance, Zilla spinosa, Acacia tortilis subsp.raddiana , Morettia philaeana, Zygophyllum coccineum, Caroxylon imbricatum and Citrullus colocynthis (especially in Wadi Natash) had the highest Q-Values (P= 0.61, 0.36, 0.3, 0.23, 0.22, and 0.21%, respectively).This result was in line with that obtained by Abd El-Ghani et al. (2013a) and Salama et al. (2012) in northern and central parts of the Eastern Desert, and Springuel et al. (2006) in the south-eastern part of this desert.Acacia tortilis subsp.raddiana, Morettia philaeana and Citrullus colocynthis were completely absent in Red Sea transect, while the presence of the salttolerant species such as Tamarix nilotica, Limonium axillare, Arthrocnemum macrostachyum, Juncus rigidus and Nitraria retusa with high presence values in the Red Sea transect indicated its salinized habitat.The record of Avicennia marina dominating the mangal vegetation along the Red Sea coast (T4) is notable, and was documentd by Zahran and Willis (2009).
Fig. 1.Location map of the study area, showing the distribution of the studied stands along the 4 transects

Fig. 2 .
Fig. 2. Chorotype spectrum and functional groups diagram of the study area.M= species magnitude and average group abundance
Fig. 4. First two axes of the DCA ordination of 142 stands with the 7 vegetation groups (A-D2) separated by cluster analysis superimposed.
Tab. 2. Species composition of the obtained 7 vegetation groups, together with their presence values (p%).Figures in bold are the dominant species that have the highest p%.Tab.4. Simple linear correlation coefficient (r) between the soil variables and DCA axes.*= p<0.05, **: p<0.01.For soil factors abbreviations and units, see Tab. 4.