About the Author(s)

Lyle E. Ground Email symbol
Environmental Planning and Climate Protection Department, eThekwini Municipality, Durban, South Africa

College of Health Sciences, University of KwaZulu-Natal, Durban, South Africa

Rob Slotow symbol
School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa

Department of Genetics, Evolution and Environment, University College, London, United Kingdom

Jayanti Ray-Mukherjee symbol
School of Liberal Studies, Azim Premji University, Bengaluru, India


Ground, L.E., Slotow, R. & Ray-Mukherjee, J., 2016, ‘The value of urban and peri-urban conservation efforts within a global biodiversity hotspot’, Bothalia 46(2), a2106. http://dx.doi.org/10.4102/abc.v46i2.2106

Original Research

The value of urban and peri-urban conservation efforts within a global biodiversity hotspot

Lyle E. Ground, Rob Slotow, Jayanti Ray-Mukherjee

Received: 01 June 2016; Accepted: 14 Oct. 2016; Published: 02 Dec. 2016

Copyright: © 2016. The Author(s). Licensee: AOSIS.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


Background: Systematic conservation planning (SCP) is a key tool in conservation prioritisation. It has recently been applied within the eThekwini Municipal Area (EMA), South Africa, a rapidly developing metro located within a biodiversity hotspot. Most vegetation types within this region are threatened, yet have received limited scientific attention.

Objectives: To assess forb biodiversity of coastal and near-coastal grasslands and contextualise potential drivers of species variation to the EMA context.

Method: We quantified forb species richness, frequency and species composition, determined the variation of these amongst plots and sites, and assessed which variables were responsible for this variation.

Results: Site forb species richness ranged from 33 to 84 species per site, with the most frequent species differing across the seven sites. Beta diversity was higher across than within sites. Distance to the closest traditional medicine market and site perimeter-to-area ratio were the greatest drivers of species richness, whilst altitude, aspect, mean temperature and rainfall seasonality were the main contributors to species composition patterns.

Conclusion: SCP can use key environmental and climatic categories to improve selection of grasslands to maximise both species richness and variation in species composition. A more fine-scale systematic conservation plan will, in turn, provide a more robust basis for development decisions. The outcomes illustrate the importance of urban and peri-urban conservation efforts, including fine-scale conservation planning, in contributing to regional, national conservation targets.


Conservation planning and sustainable environmental management have limited resources for implementation owing to competition with other socio-economic imperatives (Heywood & Iriondo 2003; Shankaran 2009; Schneider et al. 2011). Prioritisation of areas for conservation, through systematic conservation planning (SCP), has become a key tool in conservation resource allocation (Heywood & Iriondo 2003; Margules & Pressey, 2000; Pressey, Cowling & Rouget 2003). These algorithm-based planning tools use available data on habitats, species and ecosystem services to select priority sites based on important criteria (Margules & Pressey 2000). However, SCP outputs are only as good as the input data (Heywood & Iriondo 2003; Margules & Pressey 2000; Smith, Goodman & Matthews 2006), and conservation plans are often based on limited data, which requires conservation planners to define different environmental features by using environmental breaks or changes (termed mesofilters; Crous, Samways & Pryke 2013). These ‘surrogate’ habitats are then assumed to be representative of a set of species known to occur within them (Crous et al. 2013; Smith et al. 2006). Incorrect selection of surrogates may result in inaccurate SCP that does not sufficiently represent reality, leading to potentially wasted resource expenditure on marginal areas (Crous et al. 2013) and inappropriate protection for the species that need it most (Coppolillo et al., 2004).

The eThekwini Municipal Area (EMA), which includes the metropolitan city of Durban, is located on the central-southern coastal region of KwaZulu-Natal (KZN) Province and includes 99 728 ha of the original extent of the KwaZulu-Natal Coastal Belt Biome (KZNCB; Mclean, Ground & Boon 2014; Rutherford et al. 2006). This municipality is a hub of development, and the remaining grassland vegetation within the municipality currently represents less than 25% of its historical extent (eThekwini Municipality SDF 2013; Mclean et al. 2014; O’Conner 2005; Scott-Shaw & Styles 2012; Uys, Bond & Everson 2004). The EMA has distinct variations in climatic conditions, with subtropical to temperate conditions at higher elevations (high rainfall seasonality, high temperature seasonality, lower mean temperatures) and subtropical to tropical conditions at lower elevations approaching sea level (lower rainfall seasonality, lower temperature seasonality, higher mean temperatures) owing to climatic regulation from the warm Mozambique ocean current (Rutherford et al. 2006). This variation has created a heterogeneous region that supports tropical, sub-tropical and temperate mist-belt species (Eeley, Lawes & Piper 1999; Rutherford et al. 2006).

The grassland ecosystems of KZN have, until recently, had fairly limited scientific attention (O’Conner 2005; Scott-Shaw & Styles 2012; Uys et al. 2004; Zaloumis & Bond 2011), which imposes the risk of poor selection and representation of critical biodiversity areas (Crous et al., 2013). Graminoides make up the majority of the biomass of a grassland system and have historically been the main focus of scientific studies (Zaloumis & Bond 2011). Conversely, the non-graminoid herbaceous component (commonly termed forbs, which include geophytes) contributes most to overall grassland species richness (Uys et al. 2004; Zaloumis & Bond 2011).

The scale of transformation in the EMA is likely to have affected ecosystem integrity and associated functioning. Fragmentation will have led to the disruption of natural spatial disturbance patterns, particularly fire and grazing, critical for the maintenance of the successional state of grasslands (Boakye et al. 2013; Snyman 2006; Uys et al. 2004). Management (including block burning, grazing regulation, alien plant clearing) is currently the primary method for ensuring simulation of natural disturbance. It is in these managed grasslands where one would expect to find higher species richness and a more representative species composition (comparative to unmanaged grassland; Uys et al. 2004; Van Wilgen et al. 2014). The discrepancy between species assemblages within and outside managed areas adds an additional anthropogenic-derived level of heterogeneity to an already complex ecosystem (Uys et al. 2004).

Understanding the primary drivers of species heterogeneity within a grassland ecosystem is critical in ensuring the development of both adequate and accurate conservation targets, and associated conservation site selection (Crous et al. 2013; Pressey et al. 2003; Uys, Hamer & Slotow 2009). Scale is an important factor when investigating the mechanisms explaining species or ecosystem heterogeneity, as various factors have differing levels of influence at different scales (Gonzalez-Megias, Gomez & Sanchez-Pinero 2007; Huston, 1999). What may affect a particular site or area may not be as significant at a regional scale; similarly, a locally dominant species may be rare regionally, or a species recorded throughout a region may be sporadic in its local distribution (Gonzalez-Megias et al. 2007; Honnay & Hermy 2004; Hanski, 1994). These distributional differences may be caused by site-specific anthropogenic disturbances, or might be explained by environmental heterogeneity. Knowing the drivers of plant species abundance (including rarity) will significantly improve SCP output accuracy (Crous et al. 2013).

The coastal and near-coastal grasslands of the EMA are predominantly included in the Durban Metropole Grassland vegetation type (KZN2), Northern Coastal Grassland (KZN16), Ngongoni Veld (SVs4), or KZN Sandstone Sourveld (SVs5; Rutherford et al. 2006). These grasslands are nationally recognised as either vulnerable (SVs4), endangered (SVs5, KZN16) or critically endangered (KZN2), respectively (Rutherford et al. 2006). The percentage of these grassland types that are formally conserved, ranges between 1% and 4% (Rutherford et al. 2006), and the total remaining area of coastal and coastal-hinterland grassland within the EMA is 13 534 ha, which is only 12% of the pre-1850 distribution (Mclean et al. 2014). These grassland ecosystems are, therefore, under significant threat and require urgent scientific attention to optimise their conservation planning and management within the EMA spatial development plan (Rutherford et al. 2006; Uys et al. 2009).

The present study therefore attempts to improve the understanding of coastal and near-coastal grasslands within the central region of KZN, with a particular focus on key anthropogenic and environmental factors influencing species richness, species composition, species frequency, and species turnover (or beta diversity). For this purpose, we sampled sites situated in or near to the urban/peri-urban matrix of the EMA region of KZN, which contained primary coastal and near-coastal grassland. The main objectives of the study were: (1) to quantify the species richness and identify the level of variation in forb species composition at plot, site and regional scales of central KZN coastal grasslands; (2) to determine which disturbance, or biogeographic or environmental factors, may be responsible for any observed variation in forb species richness and composition; and (3) to understand the spatial distribution and frequency of occurrence of species at different spatial scales.


Study sites and sampling

After initial site inspections, seven sites (four formal municipal reserves, two formal provincial nature reserves, and one municipal managed site) were sampled within the coastal region of KZN (Table 1 and Figure 1). Six of the seven sites were located within the eThekwini Municipality (EM) while Vernon Crookes, considered a characteristic site for primary central KZN grassland in good condition, is located approximately 60 km south of the EM’s southern boundary, within a peri-urban setting, similar to that of outlying areas of the EM (Figure 1; Rutherford et al. 2006).

TABLE 1: Description of the managing agent, altitude, surrounding land uses, geology, location, condition, management period, size, number of plots, species richness, plot richness range and mean plot richness, of the sampled sites.
FIGURE 1: Map illustrating the location of the different sites sampled in relation to major roads and the eThekwini Municipality boundary.

A total of 36 plots were sampled across the seven sites (Table 1). At each site, a number of 10 m x 10 m plots were set out using tape measures and marker pegs. Species accumulation curves were used to determine required sampling effort (Willott 2001). Sampling was undertaken between October and March to ensure that seasonal forbs would be conspicuous. The positioning of plots within each site was done in a manner which covered as many grassland patches as possible. Ten 1 m x 1 m subplots (quadrats) were then randomly sampled within the plot. Forb species presence-absence was recorded for each subplot. Species identifications were confirmed using local field guides (Jeppe 1975; Pooley 2005), local experts, and the KZN Herbarium. Photographic vouchers were taken for each unique species found at each site for offsite identification and record keeping.

Using ArcGIS v.10.1 (ESRI 2011. ArcGIS Desktop: Release 10. Redlands, CA: Environmental Systems Research Institute), sampling sites were digitised. Environmental, climatic and disturbance variables for each of the 36 plots were extracted. Altitude was extracted from EM’s 2 m contour shapefile (Survey and Land Information Department, eThekwini Municipality). Aspect was extracted from the EM’s 10 m digital elevation model raster dataset (Corporate GIS Department, eThekwini Municipality). Mean rainfall, mean temperature, rainfall seasonality, and temperature seasonality were extracted from the WorldClim (Hijmans et al. 2005) raster dataset. The size of the contiguous grassland patch in which a plot was located and the size of the total natural open spaces (total natural area, regardless of the relative composition of vegetation types) were digitised using EM’s 2012 high-resolution aerial imagery (Survey and Land Information Department, eThekwini Municipality 2012) as reference. The areas of these shape files were then calculated in ArcGIS v. 10.1 using the ‘calculate field’ option in the shape file attribute table section. The reserve perimeter/area ratio was calculated initially by utilising the ‘calculate field’ option to determine the boundary perimeter of each reserve, and then by completing a ratio calculation with the area of each reserve. Distance to the nearest anthropogenic structure (e.g. house, factory, road) and distance to the closest major traditional medicinal market, in this case the Durban City Centre Victoria Market (-29.856689°S, 31.015287°E), was measured using the ‘distance tool’ in the ArcGIS v.10.1 and EM’s 2012 high-resolution aerial imagery as a reference (Hartebeesthoek94 co-ordinate system).

Data analysis
Species frequency

The frequency of each species was calculated across all seven sites, based on their presence in the subplots. Species were ranked according to their frequencies within each site to understand the general trends of species distribution patterns within and across sites. In order to determine the most frequent species at a site scale, a one-way similarity percentages species contributions (SIMPER) analysis was performed in PRIMER v.6 (Clarke & Gorley 2006) to describe the most frequent species, as a percentage of the frequency of occurrence within each plot of a site. Species frequencies of occurrence were examined at local and regional scales to determine distribution patterns. ‘Regional distribution’ assesses the distribution of species across the different sites in the study (widespread = present in ≥6 sites; restricted = present in >1 and ≤4 sites; highly restricted = present in only 1 site). ‘Local distribution’ assesses the distribution of species within the sites by measuring the subplots in which they were present (widespread = present in ≥40 subplots; restricted = present in ≤10 subplots).

Beta diversity

Shared and unique species between combinations of plots and subplots were calculated using EstimateS (Colwell 2013). Beta diversity was assessed based on the framework proposed by Carvalho, Cardoso & Gomes (2012); where Βcc (value range 0–1) reflects overall beta diversity and comprises components of species turnover (β-3, replacement of species between sites) and species richness differences (βrich, difference in species richness between sites; Carvalho et al. 2012). A Kruskal-Wallis H-test was performed in SPSS (SPSS Inc., 2011) to determine if the mean of beta diversity measures within a site was significantly different from beta diversity across sites. The βrich and β-3 values were standardised and compared in a Mann-Whitney test to determine which of these made a greater contribution to Bcc (Gutt & Piepenburg 2003).

Nestedness of the plots was assessed using a nestedness metric based on overlap and decreasing fill (NODF) run in ANINHADO (Guimarães & Guimarães 2006). In comparison with the nestedness temperature metric (which until recently had been the most popular nestedness metric used), the former metric is not sensitive to matrix shape and size, and therefore has a low probability for type I and type II error (Almeida-Neto et al. 2008). The maximum level of nestedness is obtained when a matrix has a complete overlap of presence across columns and rows, and decreasing totals between paired rows and columns. The value range of NODF is 0–100, where 100 is a completely nested system. Most observed matrices are known to have a degree of nestedness between 40 and 70 (Almeida-Neto et al. 2008). Observed nestedness values were then compared with expected nestedness values obtained from Er (the occurrence of a connection is attributed to a random cell in the matrix) and Ce (the varied placement of a connection in a cell based on the relationship between the number of ones, denoting the presence of a species, in a particular row and column in the observed matrix) models (Jędrzejewska-Szmek & Zych 2013).

Environmental variables

Univariate and multivariate analyses of environmental factor effects were performed in Primer v.6 (Clarke & Gorley 2006). Based on the presence and absence of species, a Bray-Curtis resemblance matrix (dataset standardised) was generated, and cluster and multi-dimensional scaling (MDS) analyses were performed in PRIMER. A principle component analysis (PCA) was performed on the continuous environmental variables, to identify the key variables contributing most to the variability in the species composition dataset (Ringner 2008). PC1 contributed 64.1% and PC2 24.0% to the observed variation. Variables that displayed eigenvalues (an indicator of significance of the corresponding component) >0.4 (the value at which a break was observed in the scree plot) were selected from PC1 and PC2 for further analysis (Johnstone 2001). The relationship between the distances between sites and site species composition similarity was investigated using the RELATE function within PRIMER (Clarke 1993). RELATE examines the correlation between the distance dataset and the species composition Bray-Curtis similarity index, and provides a coefficient and significance value as an output (Clarke & Gorley 2006).

The determination of significant difference between variability in species composition of each of the levels within each factor was tested using one-way PERMANOVA tests for each factor (Clarke 1993). Unfortunately, owing to limited degrees of freedom, multi-factor PERMANOVA could not be performed to determine the contribution of multiple variables and their interactions. A principal coordinate analysis (PCO) was employed to investigate patterns of correlations between environmental distances and species composition. BIOENV was used in PRIMER to determine the best combination of environmental variables that correlated with the Bray-Curtis species composition similarity matrix. BIOENV examines patterns within an environmental factor matrix and a species composition resemblance matrix, and explores all the possible combinations of factors to determine what combination of which factors best correlates with the species resemblance pattern, and gives a coefficient value related to this (Clarke 1993).

The Bayesian information criterion (BIC) was used to identify the combinations of environmental variables that best explain species richness variation, identifiable by comparatively low BIC values (Schwarz 1978). The variables within this model were then included in a regression (Schwarz 1978). A normality test (Shapiro-Wilk’s test, Shapiro & Wilk 1965) was performed, after which a regression commonality analysis (RCA) was used to identify each variable’s unique (the amount of variance accounted for by a single variable) and common variance (the amount of variance attributed to a set of variables). One key aspect of RCA is its ability to also identify specific suppressor variables (Ray-Mukherjee et al. 2014). Suppressor variables are variables that suppress irrelevant variance (variance not shared with the dependent variable) and improve the predictive capabilities of other key variables (Ray-Mukherjee et al. 2014).


Species richness and frequency

A total of 192 native forb plant species was recorded across the seven sites. At a plot scale, forb species richness ranged from 14–52. Mean plot richness across all sites was 23.5 (± 2.4, 95% confidence interval (CI)). At a site scale, Vernon Crookes had the highest total forb species richness with 84 species, and Sherwood the lowest with 33 species (Table 1).

For the most part, plots within the same site clustered together in terms of forb community assemblages (Figure 2). The exceptions to this were in Giba East, where two plots clustered separately from the other four, in New Germany where two sets of three plots clustered separately, and in Stainbank where two plots clustered separately from the other three. Interestingly, four Giba East and three New Germany plots clustered together as a separate group, and the other two Giba East and three New Germany plots clustered together as a group with Vernon Crookes (Figure 2a and 2b). The Stainbank plots clustered either with Silverglen or Sherwood. The similarity amongst plots from different sites was relatively low (<35%), given the close proximity of sites. The three groupings which split plots within the same site were, in the case of Giba East, two sets of plots located above and below a scarp at two different elevations and, in the case of New Germany and Stainbank, two sets of plots located on different slope aspects. This observation reinforces the findings of the analysis of similarities (ANOSIM) and best subset of environmental variables with maximum correlation with community dissimilarities (BIOENV) analyses, which indicated aspect and elevation as important drivers in observed forb species composition variation.

FIGURE 2: (a) Cluster dendogram and (b) MDS plot of the similarity of plots within sites illustrating that the majority of plots clustered with those within the same site.

Frequency data for plots within each site showed a consistent inverse proportional or hollow curve profile (Figure 3) which indicates that the majority of forb species present within a site have average to low frequencies of occurrence, with a limited number of very common species (McGill et al. 2007). The species recorded as common varied between sites. Frequency of occurrence of forb species within the local landscape varied amongst sites (Figure 3), with those which have been under protection for longer (Table 1) displaying a gentler slope with longer tails, suggesting a greater number of rare forb species (lower in frequency) and fewer high frequency species. The slopes for Giba East and Sherwood were comparatively steeper with shorter tails, indicating a comparative lack of rare forb species and dominance of a few, common species (although with only three shared species in the top ten most common species in each site).

FIGURE 3: Proportional frequency rank curves illustrating the different profiles of ranked species proportional frequency for each of the sites sampled. The Vernon Crookes profile is considered the closest to a pristine state. Sherwood and Giba East have a higher proportion of common species with high frequencies. Axis ranges are consistent to allow ease of comparison.

The majority of forb species were categorised as either ‘restricted in regional distribution’ or ‘restricted in local distribution’ (39 species) or as ‘very restricted in regional distribution and restricted in local distribution’ (63 species). Only two species, Tephrosia macropoda (E. Mey.) Harv. var. macropoda (Fabaceae) and Senecio glaberrimus DC. (Asteraceae), were widespread at both the regional and local scale. Eight species were considered locally common but restricted in their regional distribution.

Beta diversity

The means of all three forb beta diversity index values were significantly higher amongst sites than within sites (Table 2). B-3 had a significantly larger mean rank and sum of ranks than Brich (654 and 41 1821, compared with 607 and 382 609 respectively; Mann-Whitney U-test: p < 0.05), i.e., B-3 was a larger contributor to Bcc. This suggests that forb species turnover, or changes in forb species compositions across sites, is a more important factor in driving beta diversity amongst these grasslands than species richness differences across sites.

TABLE 2: Assessment of beta diversity across and within sites. Kruskal-Wallis H-test significance values for within site versus between site Bcc (overall beta diversity), B-3 (replacement of species between sites), and Brich (difference in species richness between sites) value comparisons.

The NODF value of the nestedness of plots was 27.90, which indicates a medium-to-low level of plot nestedness. Observed nestedness was, however, not randomly distributed and the null model nestedness value of 19.36 was significantly lower than the expected nestedness (p < 0.05 for both the Er and Ce models). The observed nestedness of the pooled site species compositions can therefore be attributed to ecological processes. The NODF value of the nestedness of plots was 40.34, indicating a higher level of nestedness amongst the sites. This value was statistically significant for the Er model (p < 0.05), but not significant for the Ce model (p = 0.34), suggesting that the nestedness may be randomly derived (Jędrzejewska-Szmek & Zych 2013).

Drivers of species composition and species richness

Three principle components contributed 92.6% of the forb species’ composition variability (PC1 = 57.0%, PC2 = 24.3%, PC3 = 11.3%). All environmental variables were included within these three components, and all had eigenvalues higher than 0.4 (threshold for consideration as an important contributor) in at least one principle component (Ringner 2008). Hence, all the continuous environmental variables were used for further analysis.

Forb species composition differed significantly with altitude, mean temperature, mean rainfall, temperature seasonality, rainfall seasonality, reserve size, contiguous grassland patch size, distance to settlement, distance to the traditional medicine market, and reserve perimeter to area ratio (Table 3a). Combinations of altitude, aspect, mean temperature and rainfall seasonality had the highest correlations with the species Bray-Curtis similarity matrix (Table 3b). No significant correlation was found between distances between sites and their similarities (rho 0.027, p > 0.05).

TABLE 3A: The effect of environmental and anthropogenic variables on species composition variation between plots. One-way PERMANOVA output F scores, and P values for environmental and anthropogenic factors.
TABLE 3B: The effect of environmental and anthropogenic variables on species composition variation between plots .The top 10 best combination of variables (numbers as per Table 3a) which correlate with the species composition resemblance matrix and which indicate the best relationship between environmental change and species composition variaition.

Grassland size, distance to traditional medicine market, perimeter to area ratio, temperature seasonality, and rainfall seasonality together explained 70.67% of variation in species richness (regression commonality analysis: p < 0.0001). In terms of the variables’ unique contribution (UC), grassland size, distance to traditional medicine market, perimeter to area ratio, and rainfall seasonality contributed significantly to the variation in the model, whilst temperature seasonality contributed marginally (Table 4) to the regression model. Grassland size was, however, reflected as a suppressor variable, mainly because it had a high negative commonality coefficient (-0.1626), cancelling out its unique contribution (0.1730), and resulting in a negligent total contribution (0.0104) to the model. Removal of grassland size from the regression lowered the R2 value, a further indication of a suppressor variable, as grassland size improved the predictive capabilities of other variables. Distance to traditional medicine market (total correlation (TC) 0.1967) and perimeter to area ratio (TC 0.0671) were the two variables with greatest total contribution to the variation in species richness model (Table 4).

TABLE 4: An indication of the different contributions of key variables, selected by the Bayesian information criterion, to the variation in species richness across plots.


The EMA’s coastal and near-coastal grasslands had a high number of forb species when compared with similar South African grassland studies. Between 30 and 44 forb taxa per plot were recorded within 1000 m2 Whittaker plots in natural dolerite and sandstone grasslands of the Drakensberg, KZN (O’Conner 2005), and 8 – 30 graminoid and forb taxa per 16 m2 plots in urban Bloemfontein, Karoo-Supergroup underlain grasslands (Dingaan & Dupreez 2013). Closer to the central KZN region, 118 taxa were recorded from 27 (100 m2) mesic grassland plots within the rural midlands of KZN (840 metres above sea level (m.a.s.l.); Uys et al. 2004). Whilst the level of species richness we recorded is expected at a coastal and near-coastal region, it is high, given the location of our sites within an urban and peri-urban landscape (Dingaan & Dupreez 2013; Knight, Crowe & Siegfried 1982; et al. 2004). This finding, combined with the absence of disturbance as a factor influencing species composition, highlights the ongoing conservation value of these threatened grasslands within this global biodiversity hotspot (Maurer, Durka & Stocklin 2003; Steenkamp et al. 2004).

The high beta diversity and nestedness observed, both within and across sites, is in accord with other KZN grasslands studies, which showed significant correlation between environmental and climatic mesofilter types and species composition and richness variation (Crous et al. 2013; Uys et al., 2004; Pausas & Austin, 2001). That plots within a site were more similar to plots in other sites, and that there was a biological basis to this pattern, is an important result which indicates that grassland patches within close proximity are not necessarily the same, and that stratified conservation planning, which incorporates quantified determinants of species variation, and uses the boundaries or thresholds of these determinants for defining subtypes, is needed to ensure adequate species variation selection (mesofilter determination; Crous et al. 2013; Laliberte et al., 2009 Gabriel et al., 2006; Coppolillo et al., 2004).

The species frequency curves, however, suggest that these grassland systems are not entirely unaffected by urbanisation-associated impacts. The observed higher number of dominant species, and the relatively higher frequencies of those dominant species in Giba East and Sherwood (both subjected to sporadic management or periods of no management), may be an outcome of modified disturbance and patch size patterns resulting in the diminishing competitiveness of certain species, as well as the entire exclusion of other species from the grassland (as shown by the significant influence of disturbance and size factors on total species richness; O’Conner & Hurt, 2009). These patterns are consistent with observations made in other grassland disturbance response studies (Menge & Sutherland 1987; Pooley et al. 2005; Zaloumis & Bond 2011).

Reserve managers should therefore strive to create an environment where natural disturbance is maintained at historical levels, and anthropogenic disturbance is excluded as far as possible, to ensure maintenance and improvement of species diversity within these grasslands (Connell 1978; Roberts & Gilliam 1995; Tilman & Lehman 2001). The size of a contiguous grassland patch, its distance to the traditional medicine market, and the shape of the grassland patch, were the biggest drivers of species richness variation. Special management attention should therefore be given to (1) the maintenance and expansion of grassland patch size, and the reduction of perimeter-to-area ratios by controlling bush encroachment and restoring appropriate adjacent areas to grassland; and (2) managing public access to and activities within grassland sites (O’Conner 2005; O’Conner & Hurt 2009; Scott-Shaw & Morris 2014; Zaloumis & Bond 2011). Considering the limitations of plant seed dispersal, reintroduction of key species, based on local ecological requirements, would assist in species richness improvement (Soons & Heil 2002; Soons et al. 2005).

The outcomes of the present study highlight the often underappreciated importance of conservation efforts in urban and peri-urban landscapes. Considering that global biodiversity hotspots often have above average population densities, urban and peri-urban conservation, through protection of relatively small (when compared with traditional ‘game reserves’ such as Kruger National Park and Hluhluwe Imfolozi Park) high biodiversity areas cannot be ignored in the pursuit of global conservation goals (Miller & Hobbs 2002). Broad-scale systematic conservation plans cannot be solely relied on, and urban and peri-urban systematic conservation plans will need to be developed with a focus on finer-scale, local-level variances in ecosystem types (Knight et al. 2006). Urban and peri-urban systematic conservation plans should consider variables such as aspect, altitude, and climate as potential indicators of fine-scale variation in ecosystem types (Crous et al. 2013). These fine-scale variations will, in turn, provide a more robust basis for development decisions (Crous et al. 2013). Without these efforts, urban biodiversity hotspots are likely to suffer continued degradation and associated species losses (Malcolm et al. 2005).


This research is supported by eThekwini Municipality through the Durban Research Action Partnership: KwaZulu-Natal Sandstone Sourveld Programme, together with the South African Research Chairs Initiative of the Department of Science and Technology and the National Research Foundation of South Africa (grant no. 84157).

Competing interest

The authors declare that they have no financial or personal relationships which may have inappropriately influenced them in writing this article.

Authors’ contributions

L.G. was the principal author, conceptualised key aspects of the paper, and undertook field work, data analysis, discussion formulation, and the write-up. R.S. undertook the discussion formulation, conceptual thinking around the paper focus, and participated in the transcript review. J.M. undertook data analysis and participated in the transcript review.


Almeida-Neto, M., Guimaraes, P., Guimaraes, P., Loyola, R. & Ulrich, W., 2008, ‘A consistent metric for nestedness analysis in ecological systems: Reconciling concept and measurement’, Oikos 117(8), 1227–1239. http://dx.doi.org/10.1111/j.0030-1299.2008.16644.x

Boakye, M., Little, I., Panagos, M. & Jansen, R., 2013, ‘Effects of burning and grazing on plant species percentage cover and habitat condition in the Highland Grassland of Mpumalanga Province, South Africa’, The Journal of Animal & Plant Sciences 23(2), 603–610.

Carvalho, J., Cardoso, P. & Gomes, P., 2012, ‘Determining the relative roles of species replacement and species richness differences in generating beta-diversity patterns’, Global Ecology and Biogeography 21(7), 760–771. http://dx.doi.org/10.1111/j.1466-8238.2011.00694.x

Clarke, K., 1993, ‘Non-parametric multivariate analyses of changes in community structure’, Australian Journal of Ecology 18, 117–143. http://dx.doi.org/10.1111/j.1442-9993.1993.tb00438.x

Clarke, K. & Gorley, R., 2006, PRIMER v6: User Manual/Tutorial. PRIMER-E, Plymouth.

Colwell, R., 2013, EstimateS: Statistical estimation of species richness and shared species from samples, Version 9, viewed 5 June 2014, from Purl.oclc.org/estimates.

Connell, J.H., 1978, ‘Diversity in tropical rain forests and coral reefs’, Science 199(4335), 1302–1310. http://dx.doi.org/10.1126/science.199.4335.1302

Coppolillo, P., Gomez, H., Maisels, F. & Wallace, R., 2004, ‘Selection criteria for suites of landscape species as a basis for site-based conservation’, Biological Conservation 115, 419–430. http://dx.doi.org/10.1016/S0006-3207(03)00159-9

Crous, C., Samways, M. & Pryke, J., 2013, ‘Exploring the mesofilter as a novel operational scale in conservation planning’, Journal of Applied Ecology 50, 205–214. http://dx.doi.org/10.1111/1365-2664.12012

Dingaan, M. & Du Preez, P., 2013, ‘Grassland communities of urban open spaces in Bloemfontein, Free State, South Africa’, Koedoe 55(1), 1–8. http://dx.doi.org/10.4102/koedoe.v55i1.1075

Eeley, H., Lawes, M. & Piper, S., 1999, ‘The influence of climate change on the distribution of indigenous forest in KwaZulu-Natal, South Africa’, Journal of Biogeography 26, 595–617. http://dx.doi.org/10.1046/j.1365-2699.1999.00307.x

Framework Planning, 2013, Spatial development framework report 2013/14, eThekwini Municipality, Durban.

Gabriel, D., Roschewitz, I., Tscharntke, T. & Thies, C., 2006, ‘Beta diversity at different spatial scales: Plant communities in organic and conventional agriculture’, Ecological Applications 16(5), 2011–2021. http://dx.doi.org/10.1890/1051-0761(2006)016[2011:BDADSS]2.0.CO;2

Gonzalez-Megias, A., Gomez, J. & Sanchez-Pinero, F., 2007, ‘Diversity-habitat heterogeneity relationship at different spatial and temporal scales’, Ecography 30, 31–41. http://dx.doi.org/10.1111/j.0906-7590.2007.04867.x

Guimarães, P.R. & Guimarães, P., 2006, ‘Improving the analyses of nestedness for large sets of matrices’, Environmental Modelling and Software 21, 1512–1513. http://dx.doi.org/10.1016/j.envsoft.2006.04.002

Gutt, J. & Piepenburg, D., 2003, ‘Scale-dependent impact on diversity of Antarctic benthos caused by grounding of icebergs’, Marine Ecology Progress Series 253, 77–83. http://dx.doi.org/10.3354/meps253077

Hanski, I., 1994, ‘Patch-occupancy dynamics in fragmented landscapes’, Trends in Ecology and Evolution 9(4), 131–135. http://dx.doi.org/10.1016/0169-5347(94)90177-5

Heywood, V. & Iriondo, J., 2003, ‘Plant conservation: Old problems, new perspectives’, Biological Conservation 113, 321–335. http://dx.doi.org/10.1016/S0006-3207(03)00121-6

Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones P.G. & Jarvis A., 2005, ‘Very high resolution interpolated climate surfaces for global land areas’, International Journal of Climatology 25, 1965–1978. http://dx.doi.org/10.1002/joc.1276

Honnay, B. & Hermy, M., 2004, ‘Scale-dependent frequency distributions of plant species in dune slacks: Dispersal and niche limitation’, Journal of Vegetation Science 15, 321–328.

Huston, M.A., 1999, ‘Local processes and regional patterns: Appropriate scales for understanding variation in the diversity of plants and animals’, Oikos 86, 393–401. http://dx.doi.org/10.2307/3546645

IBM Corp. Released 2011. IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp.

Jędrzejewska-Szmek, K. & Zych, M., 2013, ‘Flower-visitor and pollen transport networks in a large city: Structure and properties’, Arthropod-Plant Interactions 7(5), 503–516. http://dx.doi.org/10.1007/s11829-013-9274-z

Jeppe, B., 1975, Natal wild flowers, Prunell (S.A.) (PTY) Ltd., Cape Town, South Africa

Johnstone, I., 2001, ‘On the distribution of the largest eigenvalue in principle components analysis’, The Annals of Statistics 29(2), 295–327. http://dx.doi.org/10.1214/aos/1009210544

Knight, R., Crowe, T. & Siegfried, W., 1982, ‘Distribution and species richness of trees in southern Africa’, Journal of Southern African Botany 48, 455–480.

Knight, A., Driver, A., Cowling, R., Maze, K., Desmet, P., Lombard, A. et al., 2006, ‘Designing systematic conservation assessments that promote effective implementation: Best practice from South Africa’, Conservation Biology 20(3), 739–750. http://dx.doi.org/10.1111/j.1523-1739.2006.00452.x

Laliberté, E., Paquette, A., Legendre, P. & Bouchard, A., 2009, ‘Assessing the scale-specific importance of niches and other spatial processes on beta diversity: a case study from a temperate forest’, Oecologia 159, 377–388. http://dx.doi.org/10.1007/s00442-008-1214-8

Malcolm, J., Liu, C., Neilson, R., Hansen, L. & Hannah, L., 2005, ‘Global warming and extinctions of endemic species from biodiversity hotspots’, Conservation Biology 20(2), 538–548. http://dx.doi.org/10.1111/j.1523-1739.2006.00364.x

Margules, C.R. & Pressey, R.L., 2000, ‘Systematic conservation planning’, Nature 405, 243–253. http://dx.doi.org/10.1038/35012251

Maurer, K., Durka, W. & Stocklin, J., 2003, ‘Frequency of plant species in remnants of calcareous grassland and their dispersal and persistence characterisitics’, Basic Applied Ecology 4, 307–316. http://dx.doi.org/10.1078/1439-1791-00162

McGill, B.J., Etienne, R.S., Gray, J.S., Alonso, D., Anderson, M.J., Benecha, H.K. et al., 2007, ‘Species abundance distributions: Moving beyond single prediction theories to integration within an ecological framework’, Ecology letters 10(10), 995–1015. http://dx.doi.org/10.1111/j.1461-0248.2007.01094.x

Mclean, C., Ground, L. & Boon, R., 2014, eThekwini Municipality systematic conservation plan, Unpublished raw data.

Menge, B., & Sutherland, J., 1987, Community regulation: variation in disturbance, competition, and predation in relation to environmental stress and recruitment. American Naturalist 1, 730–757.

Miller, J. & Hobbs, R., 2002, ‘Conservation where people live and work’, Conservation Biology 16(2), 3300–3337. http://dx.doi.org/10.1046/j.1523-1739.2002.00420.x

O’Conner, T., 2005, ‘Influence of land use on plant community composition and diversity in Highland Sourveld grassland in the southern Drakensberg, South Africa’, Journal of Applied Ecology 42, 975–988. http://dx.doi.org/10.1111/j.1365-2664.2005.01065.x

O’Conner, T. & Hurt, C., 2009, ‘Medium-term changes in grass composition and diversity of Highland Sourveld grassland in the southern Drakensberg in response to fire and grazing management’, African Journal of Range & Forage Science 20(1), 1–10.

Pausas, J. & Austin, M., 2001, ‘Patterns of plant species richness in relation to different environments: An appraisal’, Journal of Vegetation Science 12, 153–166. http://dx.doi.org/10.2307/3236601

Pressey, R.L., Cowling, R.M. & Rouget, M., 2003, ‘Formulating conservation targets for biodiversity pattern and process in the Cape Floristic Region, South Africa’, Biological Conservation 112(1–2), 99–127. http://dx.doi.org/10.1016/S0006-3207(02)00424-X

Pooley, E., 2005, ‘A field guide to wildflowers: KwaZulu-Natal and the eastern region’, Natal Flora Publications Trust c/o Natal Herbarium, Durban.

Ray-Mukherjee, J., Nimon, K., Mukherjee, S., Morris, D., Slotow, R. & Hamer, M., 2014, ‘Using commonality analysis in multiple regressions: A tool to decompose regression effects in the face of multicollinearity’, Methods in Ecology and Evolution 5(4), 320–328. http://dx.doi.org/10.1111/2041-210X.12166

Ringner, M., 2008, ‘What is principle component analysis?’ Nature Biotechnology 26, 303–304. http://dx.doi.org/10.1038/nbt0308-303

Roberts, M., & Gilliam, F., 1995, Patterns and mechanisms of plant diversity in forested ecosystems: Implications for forest management. Ecological Applications 5, 969-977.

Rutherford, M.C., Mucina, L., Lotter, M.C., Bredenkamp, G.J., Smit, J.H.L., Scott-Shaw, R. et al., 2006, ‘Savanna Biome’, in L. Mucina & M.C. Rutherford (eds.), The Vegetation of South Africa, Lesotho and Swaziland, Strelitzia 19, pp. 510–511, South African National Biodiversity Institute, Pretoria.

Schneider, R., Hauer, G., Farr, D., Adamowicz, W. & Boutin, S., 2011, ‘Achieving conservation when opportunity costs are high: Optimizing reserve design in Alberta’s oil sands region’, PLoS One 6(8), e23254. http://dx.doi.org/10.1371/journal.pone.0023254

Schwarz, G., 1978, ‘Estimating the dimension of a model’, Annals of Statistics 6(2), 461–464. http://dx.doi.org/10.1214/aos/1176344136

Scott-Shaw, R. & Styles, D., 2012, ‘The florisitics and conservation status of Sandy Coastal Belt Grassland’, Plantlife 41 & 42, 28–39.

Seibold, D. & Mcphee, R., 1979, ‘Commonality analysis: A method for decomposing explained variance in multiple regression analyses’, Human Communication Research 5(4), 355–365. http://dx.doi.org/10.1111/j.1468-2958.1979.tb00649.x

Shankaran, M., 2009, ‘Diversity patterns in savanna grassland communities: Implications for conservation strategies in a biodiversity hotspot’, Biodiversity Conservation 18, 1099–1115. http://dx.doi.org/10.1007/s10531-008-9519-9

Shapiro, S. & Wilk, M., 1965, ‘An analysis of variance test for normality (complete samples)’, Biometrika 52(3–4), 591–611. http://dx.doi.org/10.1093/biomet/52.3-4.591

Smith, R., Goodman, P. & Matthews, W.S., 2006, ‘Systematic conservation planning: A review of perceived limitations and an illustration of the benefits, using a case study from Maputaland, South Africa’, Oryx 40(4), 400–410. http://dx.doi.org/10.1017/S0030605306001232

Snyman, H., 2006, ‘Estimating grassland production loss due to fire for a semi-arid climate’, South African Journal of Animal Science 36(5), 38–41.

Soons, M. & Heil, G., 2002, ‘Reduced colonisation capacity in fragmented populations of wind-dispersed grassland forbs’, Journal of Ecology 90, 1033–1043. http://dx.doi.org/10.1046/j.1365-2745.2002.00729.x

Soons, M., Messelink, J., Jongejans, E. & Heil, G., 2005, ‘Habitat fragmentation reduces grassland connecitivity for both short-distance and long-distance wind-dispersed forbs’, Journal of Ecology 93, 1214–1225. http://dx.doi.org/10.1111/j.1365-2745.2005.01064.x

Steenkamp Y, Van Wyk B, Victor J, Hoare D, Smith G, Dold T. et al., 2004, ‘Maputaland-Pondoland-Albany’, in R.A. Mittermeier, P. Robles-Gil, M. Hoffman, J.D. Pilgrim, T. Brooks et al., (eds.), Hotspots revisited: Earth’s biologically richest and most endangered ecoregions, pp. 219–228, Cemex, Mexico City.

Tilman, D. & Lehman, C., 2001, ‘Human-caused environmental change: Impacts on plant diversity and evolution’, Proceedings of National Academy of Sciences of United States of America 98(1), 5433–5440. http://dx.doi.org/10.1073/pnas.091093198

Uys, C., Hamer, M. & Slotow, R., 2009, ‘Turnover in flightless invertebrate species composition over different spatial scales in Afrotemperate forest in the Drakensberg, South Africa’, African Journal of Ecology 47, 341–351. http://dx.doi.org/10.1111/j.1365-2028.2008.00968.x

Uys, R., Bond, W. & Everson, T., 2004, ‘The effect of different fire regimes on plant diversity in southern African grasslands’, Biological Conservation 118, 489–499. http://dx.doi.org/10.1016/j.biocon.2003.09.024

Van Wilgen, B., Govender, N., Smit, I. & MacFadyen, S., 2014, ‘The ongoing development of a pragmatic and adaptive fire management policy in a large African savanna protected area’, Journal of Environmental Management 132, 358–368. http://dx.doi.org/10.1016/j.jenvman.2013.11.003

Willott, S., 2001, ‘Species accumulation curves and the measure of sampling effort’, Journal of Applied Ecology 38, 484–486. http://dx.doi.org/10.1046/j.1365-2664.2001.00589.x

Zaloumis, N. & Bond, W., 2011, ‘Grassland restoration after afforestation: No direction home?’ Austral Ecology 36, 357–366. http://dx.doi.org/10.1111/j.1442-9993.2010.02158.x


Crossref Citations

1. Improving the management of threatened ecosystems in an urban biodiversity hotspot through the Durban Research Action Partnership
Mathieu Rouget, Sean O’Donoghue, Chantal Taylor, Debra Roberts, Rob Slotow
Bothalia, African Biodiversity & Conservation  vol: 46  issue: 2  year: 2016  
doi: 10.4102/abc.v46i2.2199