Study area: The study was conducted in the Tankwa Karoo National Park, one of the driest areas in South Africa. Historic overgrazing has resulted in the poor and often degraded state of vegetation in large parts of the Tanqua Karoo region. Aim: This study assessed the spatial variation of vegetation structure and composition in the three main vegetation types namely: Tanqua Karoo (TK), Tanqua Wash Riviere (TWR) and Tanqua Escarpment Shrubland (TES), along an elevational gradient. Method: Using the point intercept survey method, vegetation cover, plant height, species diversity, life forms, proportions of perennials and annuals were examined in 43 sites, widely distributed in these vegetation types. Results: There were 150 vascular plant species belonging to 83 genera and 29 families recorded across all sites. The most diverse vegetation type was TES with 96 species belonging to 61 genera and 26 families. The vegetation structure was comprised mainly of shrubs and dwarf shrubs with a high proportion of leaf and stem succulent species. The mean perennial vegetation cover throughout the study area was 28 and annuals covered 22%, but this cover varied significantly between the vegetation types. The most dominant life forms were chamaephytes, which comprised 64% of all species, with cryptophytes (18%), therophytes (16%) and nanophanerophytes (2%) less abundant. Surveyed sites in the TES showed a clear association with each other but there was an overlap in the species composition and environmental conditions between some TWR and TK sites. This study highlighted the important role of elevation and topography as drivers of vegetation characteristics. Conclusion: The findings from this study can be used as a vegetation baseline to identify and prioritise degraded areas for active restoration in order to limit further degradation. Considering climate change, elevational studies may provide additional insight into species dynamics across landscapes.

Taxonomic reference: Plants of Southern Africa Checklist (South African National Biodiversity Institute 2016).

Abbreviations: ANOVA = analysis of variance; GPS = Global Positioning System; LFA = Landscape Function Assessment; NDVI = Normalized difference vegetation index; PCA = principal component analysis; PCo-A = principal co-ordinate analysis; TES = Tanqua Escarpment Shrubland; TK = Tanqua Karoo; TKNP = Tankwa Karoo National Park; TWR = Tanqua Wash Riviere.

arid environment, land degradation, life form, spatial variation, Succulent Karoo, vegetation composition, vegetation types

The Succulent Karoo Biome is a semi-desert region that extends along the western parts of South Africa and Namibia. It covers an area of approximately 111,000 km² on the arid and semi-arid fringes of South Africa’s Cape Floristic Region (Mucina and Rutherford 2006) and includes the Tanqua Karoo area, one of the driest landscapes of this biome. The Succulent Karoo Biome recorded the highest diversity value for any arid region in the world (Cowling et al. 1989). It is recognised as one of only two biodiversity hotspots within arid and semi-arid areas (Mittermeier et al. 2004) and comprises about 30% of the world’s succulent species (Smith et al. 1993) and has long been identified as a major conservation priority (Cowling and Hilton-Taylor 1994).

It is well documented that arid and semi-arid lands of the world have been subjected to vast and often radical vegetation changes (Li 2018). Overgrazing is considered one of the greatest threats to the indigenous vegetation in the Succulent Karoo (Hoffman and Ashwell 2001; Desmet 2007). As a consequence, vegetation degradation and the declining ability of arid and semi-arid rangelands to support livestock nutritional demands have been major concerns for land managers in southern Africa (Milton and Dean 1995; Hoffman et al. 2009).

The common features of degradation in the Succulent Karoo region are a decline in vegetation cover, and changes in plant species and life form composition (Anderson and Hoffman 2007). This leads to the replacement of palatable perennial grasses and non-succulent shrubs by less palatable or unpalatable, shorter-lived dwarf shrubs and annual species, and as a consequence, an increase in denuded areas and loss of fertile topsoil (Beukes and Ellis 2003; Van Rooyen et al. 2015). Although the vegetation degradation in the area has mainly been ascribed to over-utilisation by livestock and variations in rainfall regime (Kraaij and Milton 2006), the contribution of other environmental factors such as soils, microclimate, elevation and topography exacerbating degradation has received little attention.

Vegetation indicators such as vegetation cover, structure, and diversity are usually considered important indicators of rangeland health and degradation (Schulz et al. 2009). A detailed understanding of vegetation characteristics and their drivers may clarify the current state of an ecosystem and can explain future environmental responses and vegetation dynamics. All of this could benefit conservation management and restoration initiatives in this ecologically significant arid system. Effective restoration programmes in arid areas and sustainable use of its resources cannot be achieved without a thorough understanding of the ecosystem dynamics and the main factors that influence its components and spatial distribution (Saaed et al. 2018).

In the Succulent Karoo context, there is still insufficient information about the vegetation, particularly on in the Tankwa Karoo National Park (TKNP) and adjacent areas. A previous study conducted by Rubin (1998) in the TKNP before the park expanded its boundaries, assessed the physical conditions and major plant communities inside the park, but did not specifically investigate the role of elevation and survey sites were randomly selected in subjectively demarcated plant communities. Other studies in the Tanqua region did not specifically assess the role of elevation either and mainly studied the vegetation in relation to the physical environment such as geology, soils and land types (e.g., Milton et al. 1997; Land Type Survey staff 2010, 2012) with some studies focusing on the impacts of grazing that usually result in the dominance of short-lived plants where disturbance is the highest (Beukes and Ellis 2003; Rutherford and Powrie 2010). Roux and Vorster (1983) and Van der Merwe et al. (2008) studied broad scale vegetation change in the Karoo whereas Kraaij and Milton (2006) assessed these changes on a fine scale in relation to land use change, rainfall and herbivory. The few particularly detailed studies in the TKNP within its recent boundaries were on the floral composition (Bester et al. 2012; Steyn et al. 2013), soil seedbank (Saaed et al. 2018), rangeland resilience and impact of overgrazing (Saeed et al. 2020) and landscape units (Van der Merwe et al. 2015). However, these studies cover small sections of the park and the findings might not be applicable to its overall extent.

Saaed et al. (2018) investigated the composition of the soil seedbank and its role in ecosystem dynamics and restoration potential, while Osman et al. (2019) focused on soil and vegetation differences across a few ecological boundaries in the park. Using NDVI, Saaed et al. (2020) showed that vegetation conditions remained stable in the last 20 years for about 81% of the land despite it being under conservation. The study argued that vegetation productivity is largely a factor of rainfall and not a change in land use from livestock farming to conservation. Saaed et al. (2022a) used Point-Centred Quadrants to show that landscape functionality in the three main vegetation types remained stable due to the low vegetation cover and patchiness, and low mean annual precipitation. However, data and our understanding on vegetation structure and floristic composition and its variation across elevations are still insufficient and scarce across large areas of the park.

The main aim of this study was to assess the vegetation structure and composition in the TKNP in relation to different elevations and determine the variation in these vegetation characteristics within and among the three main vegetation types i.e. Tanqua Karoo (TK, SKv5), Tanqua Wash Riviere (TWR, AZi7) and Tanqua Escarpment Shrubland (TES, SKv4) as described by Mucina and Rutherford (2006). These three vegetation types comprised 94% of the total area of the park and are located at different elevations ranging from the lowland plains to the steep slopes of the escarpment (Saaed et al. 2020).

The survey sites spread throughout the three vegetation types (TK, TWR, and TES) in the TKNP spanned a wide range of elevations that varied between 338 m above sea level on the flatter plains to 1149 m on the escarpment. Slope inclination ranged from zero on the plains to about 40% at some sites on the escarpment. A total number of 150 vascular plant species belonging to 83 genera and 29 families were identified along the transects across all sites (Suppl. material 1). About 86% of the species were dicotyledons, and 14% were monocotyledons. The low proportion of grasses in the park apart from small sandy pockets on the flat open plains is noteworthy. In TK there were 67 species belonging to 48 genera and 18 families, while in TWR, 63 species belonged to 43 genera and 16 families, and in TES, 96 species belonged to 61 genera and 26 families with some of these species shared amongst the three vegetation types.

The vegetation structure of the study area was comprised mainly of shrubs and dwarf shrubs with a high proportion of leaf and stem succulent species. Annuals (therophytes) and cryptophytes were comprised mostly of winter growing species and were abundant after good rains. The study demonstrated that annual plant species were well represented in the open plains and their cover proportion was relatively high (33%), and were less frequent on the escarpment (12%), where perennial shrubs dominated. The most frequently occurring species on the open plains (TK) were Ursinia nana, Euryops annuus, and Stipagrostis ciliata (Table 1, with a detailed list of phytosociological information of the plant species shown in Suppl. material 2). Flood plains (TWR) were dominated by Drosanthemum framesii, Salsola aphylla, and Galenia sarcophylla. The most common species in the escarpment (TES) were Montinia caryophyllacea, Pteronia pallens and Tylecodon wallichii. Small tree species recorded in TES were M. caryophyllaceae and Searsia undulata. Long-lived trees such as Vachellia karroo were confined to rivers and watercourses.

The most common families in all the surveyed vegetation types were Aizoaceae, Iridaceae, and Asteraceae. Amaranthaceae and Zygophyllaceae were more common in TK and TWR while Fabaceae was only observed in TWR and TES. Poaceae was more frequent in TK and TES than in TWR. Oxalidaceae and Solanaceae were more frequent in TES than in TK and TWR. Asparagaceae, Crassulaceae, Euphorbiaceae, Geraniaceae, and Malvaceae were only recorded on the escarpment (Table 2).

The mean perennial vegetation cover was 28% across all the study sites and was significantly different amongst the vegetation types. Perennial vegetation was sparser in the TK (9%) and TWR (29%) and was significantly higher in the TES (50%) (Table 3). Low vegetation cover was associated with high mean values for denuded areas and a higher annual cover (Table 3). The mean shrub height value was 0.27 m and showed a significant difference amongst the vegetation types (Table 3).

The results showed a low rock cover (mean 5%) with the mean rock cover lowest for both TWR and TK and highest in TES. The mean litter cover was also low (4%), with the most litter found in the TWR vegetation and least litter cover (1%) in the TK (Table 3).

The most dominant life forms were chamaephytes, which comprised 64% of the species, cryptophytes (geophytes) 18%, therophytes 16%, and nano-phanerophytes 2%. Chamaephytes dominated the TES while, therophytes were most abundant in the TK as were the Cryptophytes (Table 3). The nano-phanerophytes were represented by only a few species (e.g. Lycium cinereum. Searsia undulata and Montinia caryophyllacea) with the highest percentage cover in the TES (Table 3).

The mean species richness (S) was 16 species, mean Shannon-Wiener index (H’) 1.86 and species evenness (E) 0.7. The highest mean species richness and Shannon-Wiener index (H’) were recorded in the TES and the lowest in the TK (Table 3).

The principal component analysis (PCA) showed that the first two PC axes explained 58% of the total variation amongst the examined variables (Figure 3). The first axis was positively related to slope, elevation, rock cover, perennial cover, shrub height, nano-phanerophytes, and negatively related to denuded areas and therophytes. The second axis was positively related to evenness, Shannon-Wiener index, litter cover, species richness, succulent species, chamaephytes, and negatively related to cryptophytes (Figure 3).

Figure 3. 

Bi-plot of the Principal Component Analysis (PCA) showing the distribution of the examined variables according to the first two Principal Components (PCs), the lines meet at the (0, 0) and represent the original variables. The angle between the arrows of a pair of variables indicated similarity in the distribution of the variables. The eigenvalue of the first PC axis was 6.252 explaining 42% of the variation while the eigenvalue for the second PC axis was 2.377 explaining 16% of the variation.

The cluster analysis of the survey sites based on the variables listed in Table 3, demonstrated a clear association amongst the TES survey sites. However, there was interference between the TWR and the TK in a few sites. The Bray-Curtis similarity at 80% categorised the survey sites into two main groups and at 85% similarity into six sub-groups (Figure 4A). The ordination using principal coordinate analysis (PCo-A) showed that the first and second axes explain 61% of the total variation amongst the survey sites. The eigenvalue along the first axis was 336.08 and explained 47% of the total variation. Along the second axis, the eigenvalue was 98.284 and explained 14% of the total variation (Figure 4B).

Figure 4. 

(A) Clustering diagram of Bray-Curtis similarity amongst the survey sites in the different vegetation types; Tanqua Karoo (TK), Tanqua Wash Riviere (TWR), and Tanqua Escarpment Shrubland (TES) based on the life-form and environmental variables examined during this study), (B) Principal Co-ordinate Bi-plot for the distribution of the survey sites based on the different species, life-form and environmental variables, the first two axes explain 61% of the variation. The eigenvalue along the first axis was 336.08 and captured 47% of the total variation. Along the second axis, the eigenvalue was 98.284 and captured 14% of the total variation. The solid black lines denote an 80% similarity between sites wheres the purple broken lines denote a 85% similarity.

Elevation has been shown to be a key factor in determining the distribution of plant species and vegetation communities in different landscapes (Hegazy et al. 2011; Piers et al. 2019). This is due to the effect of elevation on climatic variables such as temperature and precipitation (Xu et al. 2018), soil nutrient and water infiltration status (Petersen et al. 2004) but also the role elevation plays in grazing distribution (Samuels et al. 2007, 2022). Isohyets from meteorological records (2006–2014) for the study area shows aridity increases from the eastern uplands of the Roggeveld Mountains to the western low relief plains. In this context, some previous studies indicated that the temporal and spatial fluctuations in climatic factors enhance the variation of the vegetation in sensitive arid ecosystems mainly due to elevation variation (Roux and Vorster 1983; Chapin 2003).

The percentage cover values obtained for the denuded areas were significantly higher for the plains (TK and TWR) than in the escarpment (TES) (Table 3). This may be a result of the relatively higher aridity and decades of overgrazing in preferred habitats such as on plains and flatlands in general in comparison to the elevated and steeply sloped landscapes that are mostly used during drought periods as stated in previous studies (Bester et al. 2012; Steyn et al. 2013; Saaed et al. 2022b; Samuels et al. 2007, 2016). The similarity in denuded area cover between TK and TWR was also supported by the clustering of these two vegetation types on the lowlands. Larger denuded areas were associated with low soil cover (i.e., litter and rock), particularly on the open plains. Litter cover in many sites was completely absent due to the low standing vegetation cover (Saaed et al. 2022b), which is typical of arid regions. The low surface litter may lead to an increase in the evaporation rate from the soil, lower organic material and nutrient contents retained and recycled in the soil (Tongway and Hindley 2003), and higher degradation potential of these areas by decreasing soil stability and increasing soil erosion through agents such as wind and water (Saaed et al. 2022a).

This study illustrated that each vegetation type in the study area has adapted to its local environmental conditions which determines its characteristic structure and floristic composition. The vegetation cover and species diversity on the escarpment slopes were higher, as a result of a mixture of succulent and woody, long-lived non-succulent shrubs, in comparison to the plains dominated by therophytes and cryptophytes. This turnover of species across the Tanqua landscape has also been found in the Richtersveld, which is believed to be driven primarily by climatic factors and local habitat features, i.e., elevation, slope, and soil cover (Jürgens et al. 1999). In this part of the Tanqua Karoo, the elevational gradient and its associated effects on microclimates such as rock cover and slope as noted in Figure 3 is a crucial influential environmental factor affecting plant community composition.

Throughout the study area, the most common families were Aizoaceae, Iridaceae, and Asteraceae (Table 2), which support earlier work of Steyn et al. (2013). These authors generated a comprehensive species list using herbarium specimens (lodged in herbaria over decades/centuries) and active collecting expeditions to collect every species found across the park. Previous studies within the broader Succulent Karoo (e.g., Gibbs-Russell 1987; Milton et al. 1997; Schmiedel and Jürgens 1999; Ellis et al. 2006) reported similar findings. The abundance of these particular families may reflect the aridity prevailing in the area, since these families are generally more common in arid habitats (Sayed 1998).

The distributions of some plant families in the study area were restricted to particular habitats. Asparagaceae, Crassulaceae, Euphorbiaceae, Geraniaceae, and Malvaceae were only recorded on the escarpment. The geology, soils and microclimates of the escarpment may be more favourable to species in these families. Studies (e.g. Anderson et al. 2010) have shown that geology and soils play a major influence on the distribution of plant families. Zygophyllaceae and Amaranthaceae were more frequent in the open and on flood plains as some of these species seem to prefer disturbed, more saline, areas often found on the plains. This is in contrast to other perennial shrubs in the Zygophyllaceae family that are not indicators of disturbance. This study also observed that Atriplex lindleyi (Amaranthaceae), a naturalised dwarf shrub from Australia was present on the plains in the TKNP. Fabaceae species were common on the escarpment and very low in the open plains. The very low contribution of the Fabaceae in the plains contrasts with the finding of Gibbs-Russell (1987) who stated that this family is frequent in the Succulent Karoo biome. This may be a result of the long grazing history (Driver et al. 2002; Saaed et al. 2022b) since the Fabaceae species are highly palatable and nutritious forage species (Shavanov 2021).

Generally, the perennial plant cover throughout the study area was low. The lower perennial and higher annual cover, particularly on the plains, is evidence of low productivity and substantial vegetation degradation (Anderson and Hoffman 2007). Distinguished from other areas in the Succulent Karoo region (Born et al. 2007), the low proportion of grasses in the study area is also a sign of the poor health status of the rangeland. In such degraded rangelands in other parts of the Succulent Karoo, exposure to continuous overgrazing for decades resulted in a decline of the palatable shrubs and perennial grasses (Anderson and Hoffman 2007; Desmet 2007), in addition to the negative impact of inappropriate stocking rates on species richness.

As a consequence of the higher precipitation and lower temperature and evapotranspiration rate as well as reduced grazing densities, the escarpment is characterised by denser and taller shrubs, with the presence of some small trees. This contrasts with the low open plains characterised by a sparse cover of dwarf shrubs and numerous annual species which are promoted by heavy grazing and are strongly dependent on rainfall (Anderson and Hoffman 2007).

Even though there was a significant difference amongst the vegetation types in terms of cover of succulent species (leaf and stem), the succulent species contribution was notably high in all of the vegetation types. This high abundance of succulent species is common throughout the Tanqua Karoo region (Van der Merwe and Van Rooyen 2011b). In this study the common succulent species included Augea capensis, Drosanthemum lique, and Malephora crassa. Many of these succulent shrubs are shallow-rooted and mortality amongst this guild is high during severe droughts (Von Willert et al. 1992; Jürgens et al. 1999) resulting in a high turnover of plants in the area. This finding of generally high succulent abundance in the TKNP was also confirmed by Saaed et al. (2022b).

Different plant life forms are usually distributed along an elevational gradient (Whittaker 1960; He and Chen 1997). The most dominant life form in the study area was the chamaephyte group, which comprised 64% of the species. On the plains, chamaephytes were co-dominated by therophyte and cryptophyte species. On the escarpment, chamaephytes were co-dominated by cryptophyte species, which corresponds with the findings of Van der Merwe and Van Rooyen (2011a). Thus, the escarpment is less dynamic when compared to the open plains due to the higher contribution of chamaephytes and nano-phanerophytes, which have a longer life span when compared to the plains where therophytes and cryptophytes have higher abundance. In the plains, the periodically high contribution of cryptophytes and therophytes is strongly coupled with rainfall seasonality and events and their dominance is consistent with what may be expected in an arid climate (Van der Merwe and Van Rooyen 2011a; Van Rooyen et al. 2015, Van Rooyen et al. 2018). The significant variation in life form structure in the TKNP, particularly in chamaephyte, therophyte and cryptophyte cover across the vegetation types, provides additional evidence of indigenous vegetation structure driven by environmental factors. Vegetation structure is an important variable for monitoring complex changes in disturbed ecosystems where changes related to rangeland degradation can be detected and described (Hegazy et al. 2011; Hanke et al. 2014).

The variation in vegetation composition is reflected in the biodiversity patterns, which also varied spatially across the study area, with higher biodiversity values found for the escarpment. Species richness on the escarpment was about double the species richness on the open plains and about 1.6 times greater than on the flood plains. The Shannon-Wiener index showed a similar trend, while there was a relatively high species evenness index across the vegetation types. These plant diversity patterns may be attributed to the variation in the environmental factors (e.g. elevation gradient, rainfall, air temperature), which create a variety of habitats with different environmental conditions (Zhao et al. 2005; Hegazy et al. 2011). Elevation is amongst the most important environmental factors that influence biodiversity distribution across a landscape (Bhat et al. 2020). Plant species diversity is affected by the large scale and micro patterns of moisture and temperature conditions that usually change along an elevational gradient (Zhao et al. 2005). The high evenness scores may be related to a breakdown of hierarchical dominance and competitive structures with greater disturbance across the park (Rutherford and Powrie 2010). In addition to the variation in abiotic factors across the different elevations, the stresses caused by the past overgrazing and cropping disturbances particularly on the plains would have significantly influenced species richness and diversity.

There were significant variations amongst the surveyed sites within the same vegetation types in many of the examined parameters such as vegetation cover, litter cover, denuded areas, succulent species, species richness, and diversity illustrating variability within the same vegetation type. This may indicate different states of vegetation condition and different degradation potentials within the same vegetation type. The finer differences in soils, microclimates and plant-plant and plant-animal interactions have been shown to also play a role in variability within arid landscapes such as in the case of heuweltjies (termitaria) and quartz patches specifically in the Succulent Karoo (Kunz et al. 2012; Schmiedel et al. 2015; Eibes et al. 2021).

Although the vegetation differences amongst the vegetation types are likely a result of an interaction of the biotic and abiotic components of the system, the results of a principal component analysis (PCA) (Figure 4) demonstrated the particularly high influence of elevation on vegetation characteristics. The large influence of elevation on vegetation cover, productivity, distribution, and composition in the area is most likely through its influence on the edaphic and microclimatic factors. However other variables, such as slope and rock cover, also contribute to variability. This corresponds with the findings of other studies elsewhere, where small fluctuations in environmental parameters were shown to lead to relatively large shifts in vegetation states (Körner 2003; Stage and Salas 2007; Anderson et al. 2010; Hegazy et al. 2011).

This study revealed the high heterogeneity in vegetation structure and composition and their related indices that included cover, plant height, species richness and diversity and life form diversity across the Tankwa Karoo National Park. Areas in the park located at higher elevations where TES vegetation prevails, were more vegetated and species rich due to the effect of elevation and topography on climate and habitat heterogeneity. The low-lying TK was more arid and environmentally stressed and exhibited a low vegetation and litter cover. The effect of elevation on vegetation characteristics improves our understanding of the likely trajectories of change in vegetation when the climate changes in the future. Concomitantly, this information could be used to prioritise areas for active interventions such as degraded areas which have significantly lower plant cover and diversity, and are therefore more susceptible to erosion and further degradation. Personal observations by the authors noted that some restoration interventions by the state have been hugely successful in the park thus results of this study can also be used to inform those interventions.

The data would be made available by the authors on reasonable request.

M.I.S. conceived the study, analyzed the data, authored and reviewed the manuscript and approved the final draft. M.S. conceived the study, analyzed the data, authored and reviewed the manuscript and approved the final draft. S.J.J. conceived and designed the experiments, reviewed the drafts of the manuscript. M.L.M. conceived study, authored and reviewed drafts of the manuscript and approved the final draft. H.v.d.M. assisted with plant identification, authored and reviewed the manuscript and approved the final draft. L. K. conceived and designed the experiments, analyzed the data, prepared figures and tables, reviewed drafts of the manuscript and approved the final draft.

We thank SANParks for permission to work in the park, with a special mention of of Letsie Coetzee, Kennet Makondo and all of Tankwa Karoo National Park. Thank you to Adrian Simmers, Edwina Marinus and S. P. Bester for helping identify plant species. Manam Saeed was sponsored for a PhD by the Lybian Embassy in South Africa.

• Mogamat Igshaan Samuels (Corresponding author, isamuels@uwc.ac.za), ORCID: https://orcid.org/0000-0002-5594-2623
• Manam Saeed (manam.saaed@uob.edu.ly), ORCID: https://orcid.org/0000-0002-3740-4274
• Shayne Jacobs (posthumous), ORCID: https://orcid.org/0000-0003-1829-7650
• Mmoto Leonard Masubelele (mmoto.masubelele@sanparks.org), ORCID: https://orcid.org/0000-0003-3656-1347
• Helga van der Merwe (h.vandermerwe@saeon.nrf.ac.za), ORCID: https://orcid.org/0000-0002-7677-5123
• Lesego Khomo (lesegok@gmail.com), ORCID: https://orcid.org/0000-0002-8977-8008