Analysis of the vegetation in the King Abdulaziz Royal Nature Reserve, Kingdom of Saudi Arabia

1.1 Study area

KARNR, spanning an area of 2,843,192 ha (lat 26.895338° - 25.705221°; long 45.543225°- 46.424436°), is roughly divided into a northern and southern section by the Ad Dahna dune field, which runs in a northwest-to-southeast direction more or less through the center of the reserve (Fig. 1). The Ad Dahna dune field, significantly affects local and regional ecological gradients. As a result of the size and complex topography of this dune field, it acts as a major physical obstacle to regional airflow, and this alters wind direction and speed. Variations in dune shape and height across the field, furthermore, create localized turbulence.

Close

Fig. 1.

Location of the King Abdulaziz Royal Nature Reserve.

Export to PPT

Geologically, Saudi Arabia is divided into two structural provinces, viz. the Arabian Shield and the Arabian Shelf or Platform. The study area lies on the Arabian Shelf, which lies east of the Arabian Shield. The area consists of a sequence of Cambrian to Pliocene continental and shallow-water marine sedimentary rocks (Powers et al., 1966). Sand accumulations are the most recent geomorphological feature of Saudi Arabia and are represented by the Ad Dahna dune field.

According to the Köppen climate classification, KARNR lies in a hot desert (BWh) climate. The mean annual rainfall and temperature for Riyadh, approximately 70 km south of the study area, are 66 mm and 26.2°C, respectively (https://en.climate-data.org). Rainfall occurs primarily in winter and spring and is spread more or less evenly across the months from November to April. The mean annual rainfall and temperature for Hafar Al-Batin, approximately 73 km north of the study area, are 110 mm and 25.1°C, respectively.

Most plant geographers agree that there are two major phytogeographic regions in Saudi Arabia; however, there is controversy about the position of the boundary between these regions. The conventional viewpoint is based largely on the work of Eig (1931-1932 in Al-Nafie, 2008) and the earlier accounts of Zohary (1973). According to this viewpoint, almost the entire inland of Saudi Arabia, and thus the study area, is part of the Saharo-Arabian Region (Zohary, 1973). The Sudanian Region occurs along the coastline of the Arabian Peninsula.

2.1 Field survey

Vegetation surveys were conducted following the Zürich-Montpellier (Braun-Blanquet) School of Phytosociology (Werger, 1974), as has often been the approach for vegetation studies in this arid region of the KSA (Al-Anazi & Al-Qahtani, 2018; Al-Fredan, 2008; Al-Otaibi, 2017; Al-Sodany et al., 2011; Alsalem et al., 2020; El-Sheikh et al., 2013, 2021; Shaltout and Mady, 1996).

Prior to selecting sampling sites, the area was stratified into relatively homogeneous habitat types. The sampling strategy relied on a landscape approach using a grid-based system to subdivide the landscape into manageable sampling units, whereby each grid cell was considered a potential sampling grid. Due to the large size of the study area and the relative homogeneity of the landscape, 16 km² grid cells (4 km by 4 km) were considered appropriate. The 16 km² grid cell represents an upper limit for fine-scale grids, recommended for sub-regional to regional scale work where a balance is required between level of detail and potential management options (Seo et al. 2009; Ardron et al. 2010). To ensure representativeness, the following combined recommendations were used: at least 5 sampling replicates should be collected per Association (Hayek and Buzas, 2010) with the aim of reaching 15 to 20 sampling replicates per Association (Memon et al. 2020) while at least 10% of grid cells defined should be surveyed (Bartlett et al., 2001; Memon et al., 2020).

In total, 282 sample plots (1,000 m²) were surveyed from 17 November to 15 December 2022. The timing of the surveys coincided with the end of the dry season and the beginning of the wet season. Although the ephemeral component was poorly developed at the time of the surveys, the perennial species were present. Using the results of a dry season survey has the advantage of using predominantly the perennial component that can be found reliably year-round. Annual species that were present at the time of the survey were, however, included in the analysis. The ephemeral component is not present in all seasons and is also notoriously variable across years. A classification based predominantly on the ephemeral component would therefore not be recognizable year-round.

At each sample plot, the following assessments were made:

2.2 Mapping

As a backdrop for the mapping of plant communities, a 2022 Sentinel 2 (Level 2A; including atmospheric correction) natural colour satellite image with a 10 m resolution was used. The image was a 24-bit composite of bands 4,3,2 as RGB overlaid on a Hill-shading of an interpolated Copernicus Digital Surface Model with a 10 m resolution. Rivers, streams, wadis, rawdats, and depressions were mapped using a Digital Elevation Model (DEM).

2.3 Data analysis

Classification of the floristic data was informed by Ward’s cluster analysis and Principal Coordinates Analysis (PCoA) as an ordination technique, which were run in PC-Ord 6 (McCune and Mefford, 2011). The cover/abundance values were converted to percentages, and the percentage values log-transformed (log_e) to down-weight the disproportionate influence of dominant species and give greater emphasis to species with lower cover values. The Bray-Curtis distance measure was applied for the PCoA, and for the cluster analysis, the Euclidean distance measure. For the cluster analysis, the cut level was set at 50% information remaining. The table of sample plots against species was refined using Braun-Blanquet tabulation procedures (Werger, 1974) to produce a hierarchical classification.

Vegetation structure was analyzed per community by calculating the mean canopy cover and height for each stratum. Species richness (S), evenness (E), Shannon-Wiener (H’), and Simpson indices of diversity (D’) were computed per plot in PC-Ord 6, and a mean per community was calculated. Diversity indices were converted to the effective number of species, i.e., the exponent of Shannon-Wiener (Exp H’) and inverse Simpson (Inv D’), to compare communities.

Each species was assigned to a life form category according to Raunkiaer’s classification system (Mueller-Dombois & Ellenberg, 1974) (Supplementary material T1), and a phytogeographic region(s) (chorotype) following data in Al-Sodany et al. (2011), Alsalem et al. (2020), Danin & Fragman-Sapir (2016+), and El-Sheikh et al. (2013).

3.1 Classification of floristic data

In total, 172 plant species were recorded in 282 sample plots, and on average ten species/subspecies were recorded per sample plot (see Supplementary material T2 for the full species list). The mean frequency of occurrence of a plant species was 5.7%, while 22% of all species were encountered in only one sample plot. Only 15 species occurred in 20% or more of all sample plots (Table 1). The 172 species were spread across 37 families and 126 genera, with the most species-rich families being the Asteraceae (27 species), Poaceae (26 species), Fabaceae (13 species), and Amaranthaceae (13 species), which collectively represented 46% of all species recorded.

The classification of the floristic data revealed 19 plant communities (C) (Fig. 2). The first level of division in the cluster analysis separated the dunes and interdunes (C17–C19) from the other habitats. The remaining group was split into two large groups, with the first cluster containing the wetlands, the escarpments, ridges, and some plateaux and plains (C1–C10). The second cluster comprised predominantly sandy, gravelly, or rocky plains and some rawdat vegetation (C11–C16). As a result of the large number of plots surveyed, these clusters were subsequently analyzed separately and are presented as separate synoptic tables (Supplementary Material T2.1, Supplementary Material T2.2, and Supplementary T2.3). Topographically, communities were grouped into four broad habitat groups (Fig. 3).

Close

Fig. 2.

Vegetation map of the King Abdulaziz Royal Nature Reserve (A2-format of map available in (Suppl-F1)).

Export to PPT

Close

Fig. 3.

Schematic representation of the conceptual classification of habitats into four major groups.

Export to PPT

Group 1, Wetlands including pans, wadis, rawdats, and floodplains (C2, C3, C4): These communities occur in seasonally wet habitats that are focal points for herbivores. The vegetation is generally distinctly different from the surrounding plains, often being the only vegetated areas in a barren landscape. The wetland habitat is highly sensitive and should be conserved to avoid disrupting the hydrological regime and groundwater recharge. Total vegetation cover (%) was generally highest in this habitat group, and plant diversity was second highest among the habitat groups.

Group 2, Escarpments, rocky ridges with their associated wadis (C5, C6, C7): This habitat group is scenically the most striking and diverse and has a high sensitivity. Communities in this group were closely related floristically. Although these communities generally had the highest plant diversity, the total vegetation cover (%) was low.

Group 3, Gravel, sandy or rocky plains or plateaux (C1, C8–C16): These communities occur in habitats that are not scenically diverse, but occupy large expanses of rather monotonous plains or plateaux. There is generally a grading from rocky plains and plateaux that are related to Group 2 habitats, through gravelly plains and plateaux to sandy plains and plateaux that are in turn more closely related to dune habitats. These communities generally had the lowest plant diversity and plant cover.

Group 4, Dune communities (C17, C18, C19): These floristically closely related communities occupy the fairly stable, sandy interdune valleys and the mobile dunes. Dune crests are typically drier and more exposed to wind, leading to active sand movement and unstable surfaces, and consequently, a very sparse vegetation cover. In contrast, interdune depressions and lower dune slopes tend to accumulate finer sediments and retain more moisture, allowing for increased plant cover and biomass. Plant diversity in this group was generally low, although graminoid cover was often high.

A brief description of each community is provided in the supplementary material, indicating the location and environmental features and the most prominent species encountered. For the full floristic composition, the reader is referred to the synoptic tables (Supplementary Material T2). These descriptions should be read together with the structural and diversity data (Figs. 4 and 5) and the life form and phytogeographic spectra (Fig. 6).

Close

Fig. 4.

Mean values for (a) tall tree, (b) small tree, (c) shrub, (d) dwarf shrub, (e) graminoid and (f) forb cover per community.

Export to PPT

Close

Fig. 5.

Mean values for (a) species richness, (b) exponent of Shannon-Wiener, and (c) inverse Simpson indices of diversity per community.

Export to PPT

Close

Fig. 6.

Life form spectra of the communities in KARNR.

Export to PPT

3.2 Vegetation structure

Canopy cover of the strata showed large differences across communities (Fig. 4). In the dry season most of the communities were dominated by the dwarf shrub layer. However, in C2–C5 trees, most notably Acacia gerrardi and Ziziphus nummularia, made a noteworthy contribution towards the cover. Shrub cover was highest in C2–C5 and C9–C11. Dwarf shrub cover was highly variable, but highest in the northern pans (C2), followed by rawdats and plains (C11) and interdune valleys (C17). Graminoid cover was notably highest in the Ad Dahna dune communities (C17, C19). Forb cover is expected to be highly variable and dependent on seasonal rainfall. The dry season survey revealed the highest forb cover in the pan (C2, C3) and interdune communities (C17).

3.3 Plant diversity

Species richness, exponent of Shannon-Wiener, and inverse Simpson indices showed an approximately ten-fold range across the communities (Fig. 5). The lowest diversity was generally found on the sandy, gravel, or rocky plains, whereas the highest diversity was associated with the escarpments, rocky ridges, and associated wadis. Plant diversity in the pans, dunes, and some plain communities was intermediate.

3.4 Life forms

Overall, the vegetation was dominated by therophytes and chamaephytes (Fig. 6). The therophytes included annual/ephemeral dicotyledonous plants as well as annual grasses and sedges. Hemicryptophytes included perennial grasses and perennial herbs and contributed from 5% to 50% of the spectra. Geophytes, liana, and parasites were uncommon and constituted small percentages of the total number of species in a community. Considering all species, there were 37% therophytes: chamaephytes (33%), hemicryptophytes (21%), and geophytes (1%).

3.5 Phytogeographic affinities

Species from the Saharo-Arabian phytogeographical region dominated the phytogeographic spectra of all communities (31–65% of species/community) (Table 2) followed by the bi-regional Saharo-Arabian–Sudanian (10–50%) and Irano-Turanian–Saharo-Arabian (0–21%), and the monotypic Sudanian (0–16%) chorotypes. Other chorotypes made small contributions to the community spectra, except for the Mediterranean species in C10.

4.1 Flora

Surveys were undertaken at the end of the dry season, and the study was preceded by several drought years. Thus, the species listed per community (Supplementary material) do not represent a full checklist of species, since many ephemeral species were absent at the time of the surveys. Complete species lists cannot always be recorded in arid areas, because geophytes and annuals display event-driven dynamics by fluctuating in reaction to the timing and quantity of rainfall and might not be visible year-round (Van Rooyen et al., 2015). Complete species lists would thus require multiple visits to each plot site, and the time available for the KARNR study (limited by contractual agreement) did not allow for this. In such cases, it is recommended to use permanently recognizable species. Furthermore, several authors concluded that in arid regions, perennials are generally better indicators of specific habitat factors than ephemerals (Batanouny and Abu El-Souod, 1972; Leistner and Werger, 1973), and annuals are relatively unimportant in vegetation classification in arid regions (Ayyad and Ammar, 1973; Zohary, 1973).

Several other studies in the central and northern arid regions of the KSA concurred with the current study that the Asteraceae, Poaceae, and Fabaceae were the most species-rich families (e.g., Al-Anazi and Al-Qahtani, 2018; Al-Bakre, 2023; Al-Sodany et al., 2011; Alsalem et al., 2020; El-Sheikh et al., 2013, 2021). In a wet season survey of KARNR, the Brassicaceae will likely rank higher among the families.

4.2 Plant communities

The analysis of the dry season data identified 19 communities across KARNR. Overall, few communities showed strong diagnostic species groups, and communities often graded progressively from one to the next. The lack of clear diagnostic species groups could possibly be ascribed to the occurrence of an extended drought and the degraded state of the vegetation due to past overgrazing, which is reported to reduce diversity, leading to functional, taxonomic, and phylogenetic homogenization (Saifi et al., 2021; Salgado-Luarte et al., 2019).

Several phytosociological studies have been undertaken in the central, eastern, and northern arid parts of Saudi Arabia, and many of the communities described in the current study share similarities with communities described in previous studies. In particular, the wadi (Abd-ElGawad et al., 2021; Al-Rowaily et al., 2012; Alatar et al., 2012; Alsalem et al., 2020), rawdat (Al-Anazi & Al-Qahtani, 2018; Al Dosari et al., 2019; Al-Farraj et al., 1997; Shaltout and Mady, 1996); and nafud (dune) vegetation (Al-Fredan, 2008; Al-Otaibi, 2017; El-Sheikh et al., 2021) have received much attention in the literature.

Rhazya stricta communities (C11 & C12) covered large areas in the southern section of KARNR. Rhazya stricta is regarded as an indicator of rangeland degradation and is known to replace valuable range plant species in silty soils or shallow sand over silt (Assaeed and Al-Doss, 2002). The plant is poisonous but not deemed as a serious threat to livestock (Mandaville, 1990) and is reported to have allelopathic properties.

The main environmental variables determining plant community assembly in arid Saudi Arabia are moisture availability and the redistribution of water by runoff/run-on and drainage. Other environmental factors playing an important role are soil salinity (Al-Fredan, 2008; Al-Mutairi, 2017; Al-Rowaily et al., 2012), electrical conductivity and mineral content of the soil (Alatar et al., 2012, 2015), as well as grazing pressure (Al-Rowaily et al., 2015, Shaltout and Mady, 1996). In dune communities, the vegetation is related to dune stability (Bradley et al., 2019).

4.3 Vegetation structure

Most of the communities were dominated by the dwarf shrub layer; however, the vegetation in the wadis could, in some instances, be described as a savanna, while the vegetation in the Ad Dahna dune field could be described as open grassland.

4.4 Diversity

Overall, the lowest diversity was found on the gravel and rocky plains, and the highest on the escarpments and rocky ridges with their associated wadis. It is believed that species richness would be appreciably higher, especially in the rawdats, during the wet season when many ephemeral species should be present.

4.5 Life form spectrum

In a study in the Riyadh region, therophytes constituted 52%, chamaephytes 30%, phanerophytes 9%, hemicryptophytes 8% and geophytes 1% of all species (Al Shaye et al., 2020). Hemicryptophyte percentages were noticeably higher in the current study, particularly in the dune communities, and therophyte percentages were mostly substantially lower than reported by Al Shaye et al. (2020). However, surveys done by the same team in a part of the southern section of the reserve in late summer indicated that therophytes constituted 52% of all species (own data), supporting the data reported by Al Shaye et al. (2020).

Therophytes or ephemeral species generally constitute a large proportion of the total species in arid regions (Danin & Orshan, 1990; Van Rooyen et al., 1990). A high percentage of life forms that shed living shoots during times of water shortage is an effective method to control water loss. Drought-evading plant species are represented by therophytes and geophytes. These species exhibit significant plasticity in their growth rate, size, and phenology, and in years of extreme drought, evading species may be completely absent. Although chamaephytes and phanerophytes comprise only 38% of the spectrum in the current study, they represent the permanent vegetation component. These species often display morphological and physiological adaptations to drought (i.e., drought-tolerance), such as a seasonal leaf loss or shoot reduction, whereas others display xeromorphic features.

4.6 Phytogeography

The phytogeographic spectra of the communities were complex but dominated by the monoregional Saharo-Arabian chorotype. Comparing communities occurring predominantly in the southern section of KARNR with the predominantly northern communities revealed a southern to northern shift from Sudanian to Saharo-Arabian plant species. Thus, southern communities had a larger proportion of Sudanian species than the northern communities.

4.7 Grazing and off-road driving

Currently, the vegetation across large parts of the reserve is degraded due to past heavy grazing and browsing. Livestock grazing is one of the main causes of rangeland degradation in KSA. Al-Rowaily et al. (2015) found that livestock exclusion significantly increased plant cover, density, and species richness of all growth forms. Furthermore, the abundance of palatable species increased in exclosure plots, whereas unpalatable species increased in the open grazing area. Similar conclusions were drawn by (Shaltout and Mady, 1996). El-Sheikh et al. (2013) concluded that it may take 30 to 50 years for the desert perennial vegetation to recover fully from the effects of prolonged grazing and other large-scale human disturbances. Species often associated with overgrazing or disturbance included Rhazya stricta, Calotropis procera, and Citrullus colocynthis, whereas Acacia species, various grass species, and Rhanterium epapposum were more common in areas that had not been overgrazed.

Although livestock (camels, goats, and sheep) numbers had been controlled across most of the study site for a few years prior to the study, there was no noticeable improvement in range condition across most of the area at the time of the survey. In such an arid region, a recovery in range condition in the absence of grazing and browsing pressure is expected to be slow (El-Sheikh et al., 2013; Van Rooyen et al., 2015, 2018).

Additionally, off-road driving is a main contributor to land degradation in KSA (Assaeed et al., 2019). Forb and grass cover is negatively associated with distance to roads, while woody cover is positively associated. Off-road driving increases soil bulk density and soil electrical conductivity while decreasing porosity. Organic matter and nitrogen were not affected significantly by off-road driving disturbance. It is recommended that off-road vehicle restrictions be introduced via fencing, signage, or remote sensing detection systems.