Effect of grazing on vegetation and soil of the heuweltjieveld in the Succulent Karoo, South Africa

Acta Oecologica 77 (2016) 27e36

Contents lists available at ScienceDirect

Acta Oecologica journal homepage: www.elsevier.com/locate/actoec

Original article

Effect of grazing on vegetation and soil of the heuweltjieveld in the Succulent Karoo, South Africa € wer b, 1, Jona Luther-Mosebach b, Jürgen Dengler c, d, Ute Schmiedel a, *, Inga Ute Ro € ngro € ft b Jens Oldeland a, Alexander Gro a

Biocentre Klein Flottbek, University of Hamburg, Ohnhorststr. 18, 22609 Hamburg, Germany Institute of Soil Science, University of Hamburg, Allende-Platz 2, 20146 Hamburg, Germany €tsstr. 30, 95447 Bayreuth, Germany Plant Ecology, Bayreuth Center of Ecology and Environmental Research (BayCEER), University of Bayreuth, Universita d German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig, Deutscher Platz 5e, 04103 Leipzig, Germany b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 March 2016 Received in revised form 14 August 2016 Accepted 29 August 2016

We asked how historical and recent grazing intensity affect the patchy landscape of the heuweltjieveld in the semi-arid biodiversity hotspot Succulent Karoo. The study was carried out on a communal farmland 80 km south-west of Springbok, in Namaqualand. Heuweltjies are roughly circular earth mounds that are regularly distributed in this landscape. We sampled plant species and life-form composition, diversity measures, habitat and soil variables in 100 m2 plots, placed in three visually distinguishable heuweltjie zones (centre, fringe, and matrix) and distributed across grazing camps with different recent and historic grazing intensities. Differences between heuweltjie zones were assessed with ANOVAs and multiple linear regressions. The effect of past and recent grazing intensity on soil and plant variables was analysed by Generalized Linear Models for each heuweltjie zone separately. The three zones constituted clearly distinguishable units in terms of vegetation and soil characteristics. Soil pH and cover of annual plants increased from matrix to centres, while total vegetation cover, species richness and perennial plant cover decreased in the same direction. Historic (pre-2000) grazing patterns had the strongest effects on fringes, showing the strongest soil and vegetation-related signs of overutilization with increased stocking density. Centres showed signs of overutilization irrespective of the stocking density. The much shorter exposure to recent grazing pattern (post-2000), which was nearly inverse to the historic grazing pattern, showed increase of vegetation cover (centres) and species richness (matrix) with recent grazing intensity. We interpret these effects as still visible responses of the lower grazing intensity in these camps during the historic period. No recovery under recent grazing was observed at any of the zones. We conclude that irrespective of their conducive growing conditions, once transformed to a disturbed state, heuweltjie centres recover slowly, whereas the less impacted soil and vegetation of fringes are more responsive than centres and matrix. © 2016 Elsevier Masson SAS. All rights reserved.

Keywords: Dryland Land-use impact Rangeland Soil indicator Termitarium Indicator patch

1. Introduction

* Corresponding author. E-mail addresses: [email protected] (U. Schmiedel), inga.roewer@ €wer), [email protected], [email protected] (I.U. Ro hamburg.de (J. Luther-Mosebach), [email protected] (J. Dengler), [email protected] (J. Oldeland), [email protected]€ngro €ft). hamburg.de (A. Gro 1 Current address: Neuengammer Hausdeich 631, 21037 Hamburg, Germany. http://dx.doi.org/10.1016/j.actao.2016.08.012 1146-609X/© 2016 Elsevier Masson SAS. All rights reserved.

Rangeland farming, i.e. the farming with livestock on natural vegetation, is the dominating land-use type in drylands. Climatic variability, low productivity of the soils and limiting socioeconomic conditions in the farming communities make these lands prone to overgrazing, which can lead to hardly reversible degradation. Even though the relative role of the potential drivers for these changes is controversial (Dean et al., 1995; Kiage, 2013), the threat to drylands worldwide for degradation or “desertification” is undisputed (Reynolds et al., 2007). This threat calls for

28

U. Schmiedel et al. / Acta Oecologica 77 (2016) 27e36

monitoring systems of rangeland states and reliable early-warning indicators for overutilization (Reynolds et al., 2011; Winslow et al., 2011). A broad range of soil- and plant-related indicators for rangeland conditions have been suggested such as invertebrates (Hoffmann, 2010; Prieto-Benitez and Mendez, 2011), plant species or functional traits (Diaz et al., 2007; McIntyre and Lavorel, 2001; € ngro €ft Wesuls et al., 2013), as well as soil chemical features (Gro et al., 2010; Manley et al., 1995; Smet and Ward, 2006), that show characteristic responses to grazing. However, rangelands are spatially heterogeneous and variable over time with varying responses to grazing (Illius and O'Connor, 1999; Vetter, 2005). Habitat types that are particularly attractive to herbivores can thus provide early information about consequences of disturbances (Stokes et al., 2009). Heuweltjies in the southern African winter rainfall region (Lovegrove and Siegfried, 1986; Picker et al., 2007) have been shown to be attractive to herbivores (Esler and Cowling, 1995; Knight et al., 1989; Kunz et al., 2012; Midgley and Musil, 1990). Heuweltjies are earth mounds of 3e5 m in height and up to 30 m in diameter (Lovegrove and Siegfried, 1986). They occupy up to 25% of the land surface of the south-western part of southern Africa (Picker et al., 2007). Age and origin of heuweltjies remain a matter of contention (Cramer and Midgley, 2015; Picker et al., 2007), though most authors assume that they are fossil termitaria of the species Microhodotermes viator (Petersen, 2008; Picker et al., 2007) and dating back to 20,000 years of age (Midgley et al., 2012). They are recently sustained by rodent burrowing activity. Like the mima mounds in Argentina (Roig et al., 1988), termite mounds in East African Savannas (Okullo and Moe, 2012) and fairy circles of Africa's summer rainfall region (Jürgens, 2013), heuweltjies contribute heterogeneity to the landscape, are persistent over time, and influence distribution of biomass and diversity. Due to their special habitat conditions, heuweltjies in the Succulent Karoo support distinct vegetation units (Luther-Mosebach et al., 2012), with soil properties following concentric patterns (Petersen, 2008). Their particular species composition is commonly explained by characteristic edaphic features of the heuweltjies, such as element composition, enrichment in organic compounds, pH value, and physical properties (Esler and Cowling, 1995; Midgley and Musil, 1990). In particular, heuweltjie centres are characterised by better water availability (Midgley and Musil, 1990; Turner, 2003), significantly lower stone content (Kunz et al., 2012) and higher silt content (Midgley and Musil, 1990). Heuweltjie vegetation is either denser and higher or sparser than the surrounding vegetation, depending on the climatic conditions and disturbance regime (Turner, 2003). Compared to the surrounding matrix, heuweltjie vegetation contains more deciduous and opportunistic species, which has been ascribed to the better mineral, nutrient, and water supply (Esler and Cowling, 1995; Midgley and Musil, 1990; Rahlao et al., 2008). The latter characteristics may be responsible for the observed attractiveness of the vegetation to herbivores (Milton and Dean, 1990) due to which Stokes et al. (2009) tested heuweltjies in the southern Succulent Karoo vegetation as “indicator patches” for overutilization in camps, that were each subjected to a different grazing intensity, from low to high, for the last 50 years. Following the same understanding of heuweltjies as azonal habitat types that are subject to grazing and disturbance, we studied rangeland camps that have been subject to different grazing intensity over several decades until 1999. In the year 2000, the tenure changed and brought along changes in land-use patterns, resulting in resting of some of the formerly overutilized camps and more intense use of some of the formerly moderately used camps. We therefore also tested for the effect of this recent grazing intensity. For this study, we distinguished between three heuweltjie

zones, i.e. centre, fringe and the surrounding matrix, and compiled information on soil and vegetation variables for each of the heuweltjie zones. The following key questions were considered: o Which biotic (vegetation structure, plant diversity) and abiotic environmental parameters (soil chemistry) characterise the three different heuweltjie zones (i.e. heuweltjie centres, fringes and the surrounding matrix vegetation) in the study area? o How do vegetation and soil of the three heuweltjie zones respond to different levels of long-term historic and short-term recent grazing and which parameter is the most sensitive indicator of grazing intensity?

2. Materials and methods 2.1. Study area The study was conducted on the communal farm land of the Soebatsfontein community (Fig. 1; between 30
50 58.0200 S, 17
34012.2000 E and 30
15013.0900 S, 17
33015.9300 E), in the Namaqualand lowland, Northern Cape Province of South Africa. For further information on the study area, see description of Observatory S22 Soebatsfontein in Haarmeyer et al. (2010). The climate is subtropical semi-arid with an average of 130 mm annual precipitation, falling mainly in winter from May to September, and additionally as fog. The mean annual temperature is 19.4
C, but temperatures fall below 10
C in winter, while frost hardly occurs (Haarmeyer et al., 2010; Mucina et al., 2006). The area is of slightly undulating topography and ranges from about 200 to 392 m above sea level. Geologically, the area is lined by igneous rocks such as gneisses and granites under red and yellow colluvial soils. Dominating soil groups are Leptosols, Durisols, and Cambisols (Haarmeyer et al., 2010). The major vegetation type forms part of the Namaqualand Heuweltjieveld within the Namaqualand Hardeveld Bioregion, which is part of the Succulent Karoo Biome (Mucina et al., 2006) and has been recognized as a biodiversity hotspot (Myers et al., 2000). Approximately 25% of the flora is endemic to Namaqualand (Desmet, 2007). The vascular flora of the Succulent Karoo with approximately 6300 species on 100,251 km2 is extraordinarily species-rich, particularly for a semi-arid to arid region. It is predominantly composed of leaf-succulent dwarf shrubs (mainly Aizoaceae), geophytes, and few grasses (Haarmeyer et al., 2010; Luther-Mosebach et al., 2012; Milton et al., 1997). Until the year 1999, the study area was owned and managed commercially for small livestock production (mainly sheep). Due to the land reforms of the post-Apartheid government of South Africa, approximately 15,000 ha of the farmland were handed over to the local municipality to be used as communal farmland by the adjacent community. It is since managed by a Commonage Committee and the local municipality. The farmland is subdivided into 16 camps of 200e1400 ha in size, which are used by small groups of farmers for livestock farming (sheep and goats). 2.2. Grazing intensity data The historic (from 1986 until 1999) grazing intensities per camp were calculated based on all known stock movements (the number of sheep multiplied by the number of days they stayed in the camp) over the period of one year (sheep grazing days per year: SGD). An assumed carrying capacity of 9 ha/small-stock units (SSU, i.e. sheep or goats) was used by the former farm manager (Floors Brand, pers. comm.) to calculate “potential sheep grazing days” [SGDpot ¼

U. Schmiedel et al. / Acta Oecologica 77 (2016) 27e36

29

Fig. 1. Study area: Soebatsfontein commonage in the Northern Cape Province, South Africa, with camp borders. Historic (left) and recent (right) grazing intensity classes per camp. Grazing intensity categories: 3 ¼ high (>120%); 2 ¼ intermediate (80e120%); and 1 ¼ low (<80%). The two most northern camps were excluded from this study due to the lack of heuweltjies.

(Camp area/Carrying capacity of the camp)  365 days]. Grazing intensity was then expressed as the ratio of actual SGD to SGDpot (in %), and averaged over the recorded years. The grazing intensities of the camps were classified into the categories: 3 ¼ high (>120%); 2 ¼ intermediate (80e120%); and 1 ¼ low (<80%). Data on recent grazing intensity (determined for the years 2008e2009) per camp was derived from data on livestock numbers collected by the paraecologist Reginald Christiaan (pers. comm.). The data of 2008/2009 were processed the same way as for the historic stocking data, assuming the stock numbers to be constant for one year. The resulting spatial patterns for historic and recent grazing intensity per camp are shown in Fig. 1. 2.3. Sampling design and vegetation survey The camps of Soebatsfontein commonage and parts of the adjacent Namaqua National Park were classified into ten landscape units based on a satellite image from 25 November 2003 (Source: Google Earth), which was overlaid by 4000 regular grid points of 300 m distance. From these landscape units we considered those for this study that were heuweltjie dominated and situated inside the Soebatsfontein commonage (see Fig. 1). From those 777 grid points, we randomly chose 30 positions, stratified according to the historic land-use intensity on the camps (Table 1). At each of these s of 10 m  10 m in a nested positions, we sampled 3e5 releve  was sampled at the exact position of the grid design: one releve point (M ¼ matrix), one was placed in the centre of the nearest heuweltjie (C ¼ centre). One or two plots, respectively, depending on the heterogeneity and size of the zone, were placed orthogonally to the direction of the slope in the hilly landscape in the fringe zone next to the central plot (F ¼ fringe). If the position of the grid point  was was situated on a heuweltjie centre or fringe, a matrix releve

placed randomly in the matrix zone of the heuweltjie. This resulted s (i.e. 30 releve s on heuweltjie centres, 67 on heuin 121 releve s of weltjie fringes and 24 in the surrounding matrix). The releve this study were incorporated into a comprehensive vegetation classification for the study area by Luther-Mosebach et al. (2012). s were conducted in August and September 2008, The releve which is the growing season in the Succulent Karoo. For the s, the projected cover of all vascular plant species vegetation releve (hereafter referred to as species) was estimated in percent. Species were grouped according to life-history (annual, perennial, occasional perennial) and life-form categories (chamaephytes, geophytes, hemicryptophytes, lianas, phanerophytes, and therophytes, see Table S1). The following environmental data were recorded:  estimated cover in percent of the mineral soil surface according to grain size classes (i.e. fine material, stones 2e6 cm, stones 6e20 cm, stones 20e60 cm, blocks, rock outcrops).  estimated projected cover in percent of dead wood standing, dead wood litter, soft litter of dicotyledonous, soft litter of grass, dung, biotic soil crusts, bryophytes and lichens, area affected by burrowing mammals  aspect (
), slope (
), position of plot on the slope (i.e. summit, shoulder, middle, foot slope, toe slope, valley)

2.4. Soil sampling and analysis Soil sampling was conducted at the same time as the vegetation s (Aug./Sept. 2008). Mixed topsoil samples at 0e10 cm depth releve were taken and combined by use of a shovel from nine positions evenly distributed along two diagonal lines across the plots. Of each

30

U. Schmiedel et al. / Acta Oecologica 77 (2016) 27e36

Table 1 Camp names, historic and recent grazing intensity in percent and classes classified into the categories: 3 ¼ high (>120%); 2 ¼ intermediate (80e120%); and 1 ¼ low (<80%), and s per camp, sorted according to decreasing historic grazing intensity. Camps marked with an asterisk (*) were subdivided in 2001 to allow for better rangeland number of releve management. Camp name

Size of camp in ha Historic relative grazing intensity

Historic grazing intensity Recent relative grazing class intensity

s per Recent grazing intensity Number of releve class camp

Kruisvlei Bakenkop Huiskamp Langkamp Kateklip a* Kateklip b* Rooshoogte West* Rooshoogte East* Koeislagkoppie Witbanken Ramkamp Skurweberg Mean value per camp

1190 827 1044 924 2196

163 142 135 104 96

3 3 3 2 2

1 1 1 3

2370 After 1994: 1601 1406 2139 196 1196

82

2

68 60 43 41 93

2 1 1 1 2.0

mixed soil sample 300 g was air dried, sieved to 2 mm, and used for further soil chemical analyses. A soil subsample was ground to <0.06 mm with a vibration disc mill and dried at 105
C. To determine the total amount of carbon (TC) and total amount of nitrogen (N) a C:N-analyzer (Vario Max, Elementar) was used to combust a fine-ground sample (about 0.7 g) at 900
C with oxygen (DIN, 1995). For determining the total amount of inorganic carbon (IC), a fine-ground sample (0.1e2.0 g) was dried at 105
C overnight and thereafter treated with 5 ml 50%-phosphoric acid in a closed system. The released CO2 was measured by gas chromatography (GC 14B. Shimadzu). The amount of soil organic carbon (OC) was calculated by TC e IC. The electrical conductivity (EC) was measured in a demin water-solution with a 1:2.5 ratio with a conductivity sensor (DIN, 1997). For determination of the pH-value, a soil suspension was prepared by addition of 0.01 M CaCl2 with a 1:2.5 ratio, and measured with a pH-electrode after 1 h, while repeatedly stirring the suspension up to until 10 min before measurement (PSA, 2002).

2.5. Heuweltjie density Based on satellite images from the year 2003 (Google Earth) the density of and area covered by heuweltjies in an area of 300 m  300 m per camp was determined by digitizing polygons around visible heuweltjie structures by using the geographical information system software ArcGIS 9.1 (ESRI Inc. 2005). The sampled area was placed in the approximate centre of each camp covering an area that was representative for the camp.

2.6. Cumulative palatability of plant species We asked three local livestock farmers to rate the most dominant plant species, according to palatability (in four ordinal classes: eaten frequently (2), eaten (1), rarely eaten (0), poisonous (1)) to sheep and goats respectively, and the seasonal importance of these plants. The three farmers were selected from the total of 20 farmers farming on the commonage based on their long-standing farming experiences and sound local knowledge. The data were complemented with information from comparative interviews conducted by Linke (2009) in the same study area. Ratings per species were averaged, and species with negative means were excluded. The mean values were multiplied by the respective cover value per species and plot, and added up for each plot. The resulting value is hereafter called cumulative palatability (Table S2).

25 0 46 159 23 194 36 94 49 111 2 109 71

3 1 2 1 2 1 2 1.6

7 21 12 3 13 7 4 4 8 19 7 16

2.7. Differences between heuweltjie zones To identify habitat conditions for plants on heuweltjies compared to the surroundings, we tested for the differences between the three zones regarding soil chemical, vegetation, and structural parameters of soil surface as dependent variables (Table 2). We used Analyses of Variance (One-way ANOVAs) of the software STATISTICA 10.0 (StatSoft Inc, 2007). Variances and frequency distribution of response variables were inspected visually (Quinn and Keough, 2002). If there was a clear relation between means and variances or if frequency distributions were skewed, data were either rank-transformed (when data contained 0) or otherwise log-transformed and inspected for a better distribution. Nevertheless, means and additional p-values of non-transformed data are given, as it is assumed that a rejection of the zero hypothesis (H0) based on transformed data will justify a rejection of H0 with original data, and maintaining the original H0 enables interpretation that is more straightforward. Tukey's HSD post-hoc tests were applied to reveal which zones differed from each other. For unequal sample sizes (Zones M, F, C: NM ¼ 24, NF ¼ 67, NC ¼ 30) Type I errors may be inflated, especially if the smaller sample had a larger variance (Quinn and Keough, 2002). Therefore, in those cases a more conservative significance level was adopted (p ¼ 0.01). 2.8. Influence of abiotic environmental parameters on vegetation We analysed the relative influence of heuweltjie zones and topographical, as well as soil chemical parameters, on vegetation features, by applying multiple linear regressions. Response variables were diversity and structural parameters of vegetation (Table 2). Predictors were heuweltjie zones, topographical and soil chemical variables. To avoid multicollinearity (Gotelli and Ellison, 2004), we identified collinear predictor variables (at jrj > 0.8) with Spearman's rank correlation and used only one of the correlating variables. CaCO3 and pH were highly correlated (r ¼ 0.85), also OC and N (r ¼ 0.84). Thus, OC and pH were used. Normality of distributions was checked for the remaining parameters. For each response variable, we conducted a multiple regression with the stated predictor variables. Variables were added in automated stepwise forward selection (StatSoft Inc, 2007) in order to select predictor variables with the least collinearity, and to produce higher tolerances (1-r2) and F-ratios (Gotelli and Ellison, 2004). F-to-enter was set at 4.0 in accordance with other ecological studies (Loredo and Vuren, 1996; Slotow and Rothstein, 1995). Residuals

U. Schmiedel et al. / Acta Oecologica 77 (2016) 27e36

31

Table 2 Description of variables used in the analyses. Variable Soil parameters C:N ratio CaCO3 EC N pH OC Diversity parameters Cumulative palatability Evenness Shannon index Simpson index Species richness Vegetation parameters Annuals Chamaephytes Geophytes Hemikryptophytes Phanerophytes Vegetation cover Structural parameters for soil surface Classes rill erosion Classes sheet erosion Cover biotic crust Cover bioturbation Cover dead wood Cover dung Cover fine material Cover soft litter Grazing intensity Distance to disturbance sites Proportional resting periods Relative historic grazing intensity Relative recent grazing intensity Topography Easting Northing Slope Heuweltjie zones

Description Carbon-to-nitrogen ratio Calcium carbonate content [%] Electrical conductivity (1:2.5) [mS/cm] Total nitrogen content [%] pH in CaCl2 (1:2.5) Total organic carbon content [%] P

m; i¼1 (ni)¼(Average palatability classplant i x Coverplant i) m ¼ Number of species Shannon based evenness Cover based Shannon index Simpson index (abundance-based diversity index Number of species per plot [100 m2]

Estimated Estimated Estimated Estimated Estimated Estimated

cover of annual plants [%] cover of chamaephytes [%] cover of geophytes [%] cover of hemicryptophytes [%] cover of phanerophytes [%] total vegetation cover [%]

Classes of rill erosion in slight, moderate or severe Classes of sheet erosion in slight, moderate or severe Estimated cover of biotic soil crust [%] Estimated cover of bioturbated area [%] Estimated cover of woody litter þ dead plants [%] Estimated cover of dung [%] Estimated cover of mineral soil surface <2 mm [%] Estimated cover of soft topsoil litter [%] Shortest distance to water point and stock post, in ordinal scale Resting period/recorded time span [%] SGD/SGDpotential [%] until 1999 SGD/SGDpotential [%] in 2008 East-West orientation of aspect. Calculated as sine of aspect in radian measure or x-coordinate. North-South orientation of the aspect. Calculated as cosine of aspect in radian measure or y-coordinate Inclination of slope [
] M e matrix zone, F e fringe zone, C e centre zone

were plotted against predicted values from the regression equation to check for a potential relation.

3. Results 3.1. Heuweltjie topography and density

2.9. Impact of grazing intensities on vegetation and soil To establish the response of heuweltjie zones to historic and recent grazing intensities and to test for their suitability as indicator patches sensu Stokes et al. (2009), we conducted a Generalized Linear Model with historic and recent grazing intensity for each heuweltjie zone separately. Historic and recent grazing days were used as predictors, while a list of soil properties and vegetation parameters served as response variables (Table 2). In order to minimise the clustering effect of multiple plots per camp, the categorical variable “camp” was constantly included as a covariate. Furthermore, we checked the Pearson residuals for spatial autocorrelation using Moran's I computed with the R software (R Core Team, 2013) and its packages gstat (Pebesma, 2004) and ncf version 1.1e5 (Bjornstad, 2013). Moran's I correlograms for specific distance classes were computed with the package spdep (Bivand, 2014). For the spatial verification of residuals we projected the coordinates of each vegetation plot into the coordinate system UTM 34S/WGS 84 (EPSG: 32,734) using the rgdal package version (Bivand et al., 2014). The GLM were performed using the basic stats package (R Core Team, 2013). Pseudo R2 for model evaluation was calculated using the package pscl version 1.04.4 (Jackman, 2012).

Abundance, individual size, and area covered by heuweltjies in the study area were highly variable (Table 3, Table S3). The average density of heuweltjies per 90,000 m2 (300 m  300 m) was 14.9 heuweltjies, covering 12% of the surface. The surface area of a single heuweltjie ranged from 83 m2 to 2749 m2. 3.2. General characterisation of vegetation s a total of 267 vascular plant species were In the 121 releve recorded. All zones were dominated by chamaephytes with a mean vegetation cover of 25.2% (M), 24.2% (F), and 14.7% (C) (Table 4). Second most important life form for heuweltjie centres were the therophytes with a mean vegetation cover of 4.9%, and for the matrix and fringe vegetation the phanerophytes with a mean vegetation cover of 2.1% (M) and 1.4% (F), respectively (Table 4). 3.3. Differences between heuweltjie zones To identify habitat conditions for plants on heuweltjies compared to the surroundings, we tested for the differences between the three zones regarding soil chemical, vegetation, and structural parameters of the soil surface applying one-way ANOVA.

32

U. Schmiedel et al. / Acta Oecologica 77 (2016) 27e36

Table 3 Abundance, individual size and area covered by heuweltjies per 90,000 m2 for each camp, minimal and maximal values. Camp

Kleinkoeroe Witbanken N Witbanken S Koeislagkoppie Skurweberg Langkamp Rooshoogte W Rooshoogte E Kateklip W Kateklip E Kruisvlei Ramkamp Bakenkop Huiskamp Mean ± SD

Number of heuweltjies

Area covered by Heuweltjies

3 5 10 24 14 29 15 16 8 1 38 28 11 7 14.9 ± 10.4

Area individual heuweltjies

[m2]

[%]

Min [m2]

Max [m2]

3074 3344 4890 18,365 9597 15,169 12,442 8774 8778 919 29,517 20,878 13,198 3850 10,914 ± 7726

3.4 3.7 5.4 20.4 10.7 16.9 13.8 9.7 9.8 1.0 32.8 23.2 14.7 4.3 12.1 ± 8.6

888 83 139 312 241 149 178 166 693 919 237 218 613 131 355 ± 282

1099 1462 975 1582 1538 850 1648 999 1639 919 1413 1478 2749 1221 1398 ± 463

The ANOVA revealed significant differences between at least two of the zones at a time for most of the parameters (Table 4). There was a significant increase in soil pH from matrix to centre. This gradient was accompanied by a significant increase in the plant-relevant nutrients Ca and N in the same direction. The C:N ratio and OC, in contrast, showed lowest values for the fringes compared to both, matrix and centres. Species richness and Shannon index decreased from the matrix to the heuweltjie centres, whereas Evenness and Simpson index did not show any significant differences between the zones. The lifeform composition had clear differences between the three zones. Total vegetation cover, chamaephyte cover and hemicryptophyte

cover decreased from matrix to centres whereas the therophyte cover increased in the same direction. Among the structural parameters of the soil surface only bioturbation showed a significant increase from matrix to heuweltjie centres.

3.4. Influence of abiotic parameters and grazing on vegetation The resulting significant models are shown in Table 5. Soil pH turned out to be the most frequent predictor for vegetation variables (e.g., cover of most life form groups, richness- and abundance-based diversity indices). The Simpson index and therophyte cover were the only variables that were not explained by pH

Table 4 ANOVA results for differences between heuweltjie zones with pooled other categories in soil chemical, vegetation and structural parameters for soil surface. Df residual ¼ 118, N ¼ 121 (NM ¼ 24, NF ¼ 67, NC ¼ 30). For calculation of cumulative palatability refer to text. Asterisks (*) indicate application of significance level p ¼ 0.01 due to heterogeneity of variances and skewed or non-existent normal distribution. Otherwise significance level was p ¼ 0.05. Significant p-values are in bold. Superscript letters (a), (b) and (c) indicate homogenous groups according to Tukey's HSD at p ¼ 0.05, respectively p ¼ 0.01. Zones Parameter

Soil parameters pH EC [mS/cm] CaCO3 [%] OC [%] N [%] C/N ratio Vegetation parameters Species richness Shannon index Evenness Simpson index Average palatability Vegetation cover [%] Cover life form [%] Cover annuals Chamaephytes Geophytes Hemicryptophytes Phanerophytes Structural parameters Cover fine material [%] Cover biotic crust [%] Cover dead wood [%] Cover soft litter [%] Cover dung [%] Cover bioturbation [%] Cover sheet erosion [%] Cover rill erosion [%]

Matrix

Fringe

Centre

p

Mean

SE

Mean

SE

Mean

SE

5.97a 987 0.04a 0.87a,b 0.066a 13.3a

±0.16 ±171 ±0.27 ±0.07 ±0.005 ±0.41

6.87b 880 0.31a 0.71a 0.061a 12.2b

±0.09 ±102 ±0.16 ±0.24 ±0.003 ±0.25

8.10c 685 3.00b 1.06b 0.099b 13.7a

±0.14 ±153 ±0.24 ±0.06 ±0.004 ±0.37

<0.001 0.392 <0.001 <0.001 <0.001 0.002

23.7a 1.70a 0.539 0.686 27 28.7a

±1.21 ±0.099 ±0.030 ±0.037 ±4 ±2.31

20.3b 1.52a,b 0.508 0.646 25 26.6a

±0.72 ±0.059 ±0.018 ±0.037 ±2 ±1.38

15.1c 1.29b 0.491 0.581 23 20.2b

±1.08 ±0.089 ±0.027 ±0.033 ±3 ±2.06

<0.001* 0.008 0.501 0.093 0.724 0.013

0.9 25.2a 0.1 0.7a 2.1

±0.71 ±2.45 ±0.07 ±0.17 ±0.50

0.6844 24.2a 0.1099 0.2b 1.4

±0.42 ±1.47 ±0.04 ±0.10 ±0.30

4.8597 14.7b 0.168 0.0b 0.5

±0.63 ±2.19 ±0.06 ±0.15 ±0.44

<0.001* <0.001 0.75 0.009 0.06

95.7 58.9 2.2 2.3 0.06 0.5a 0.8 0.7a,b

±0.82 ±6.55 ±0.50 ±0.59 ±0.02 ±2.71 ±0.15 ±0.16

97.6 60.8 2.1 2.6 0.04 1.1a 1 0.8a

±0.49 ±3.92 ±0.30 ±0.36 ±0.01 ±1.62 ±0.09 ±0.10

99.1 56.2 3.7 1.3 0.04 14.8b 0.5 0.3b

±0.73 ±5.86 ±0.45 ±0.53 ±0.02 ±2.42 ±0.14 ±0.15

0.011* 0.811 0.009 0.15 0.562 <0.001 0.018* 0.023

a

Transformation Transf. Type

p (transf.)

Rank

<0.001

Log Log

<0.001 <0.01

Log

0.013

Log

<0.001

Log

0.007

Rank

0.073

Rank

<0.001

U. Schmiedel et al. / Acta Oecologica 77 (2016) 27e36

at all, but at least partly by exposition (northing). The results of the Generalized Linear Models for the effect of historic and recent grazing intensity on the soil and vegetation of the heuweltjie zones are listed in Table 6. Historic grazing intensity had significant effects (p < 0.05) on the cover of chamaephytes and on the palatability in the heuweltjie centres, associated with a negative correlation of both variables with the grazing intensity. However, effects on the heuweltjie fringes were much more pronounced and exerted by a larger number of variables. In these cases, grazing intensity was negatively correlated with vegetation (i.e., total and chamaephyte cover) as well as OC. For both zones, fringes and the matrix, historic grazing intensity showed a negative effect on soil C:N ratio (p < 0.05). Recent grazing intensity had only effects of statistical significance (p < 0.05) on vegetation cover on heuweltjie centres and species richness in the matrix. Both were positively correlated with recent grazing intensity.

4. Discussion 4.1. General characterisation of heuweltjies in the landscape With 12% of the surface area, heuweltjies contribute considerably to the grazing area in Lowland Namaqualand. The heuweltjie density of 166 ± 118 mounds per km2 in the study area represents the lower end of the range determined by Picker et al. (2007) for 34 randomly selected sites in the winter rainfall region of South Africa. For areas with mean annual rainfall below 345 mm year1, which applies to our study, Picker et al. (2007) found a significant decrease of heuweltjie density with decrease of rainfall. Their findings are in Table 5 Results of multiple linear regressions on the influence of abiotic environmental parameters on vegetation; for abbreviations of response variables see Table 2. Standardized coefficient

SEM

t

33

Table 6 Results of the GLM-Analysis. Only models with p < 0.1 are shown. Different significance levels are denoted with asterisks (* ¼ p < 0.05; ** ¼ p < 0.01; *** ¼ p < 0.001). HGI ¼ historic grazing intensity; RGI ¼ recent grazing intensity; Zone: C ¼ heuweltjie centre; F ¼ heuweltjie fringe; M ¼ matrix vegetation; Model: G ¼ gamma distribution of parameter values; P ¼ Poisson distribution of parameter values; for abbreviations of response variables see Table 2. Response

Predictor

Zone

Model type

t-Value

p-value

Cover chamaephytes (%) Palatability OC (%) N (%) C:N ratio Cover vegetation (%) Cover chamaephytes (%) IC (%) C:N ratio N (%) Cover vegetation (%) OC (%) Cover vegetation (%) Cover chamaephytes (%) Species richness (100 m2) Bioturbation (%) IC (%) OC (%) N (%) Species richness (100 m2)

HGI HGI HGI HGI HGI HGI HGI HGI HGI RGI RGI RGI RGI RGI RGI RGI RGI RGI RGI RGI

C C F F F F F M M C C F F F F F M M M M

G G G G G G G G G G G G G G P G G G G P

2.32 2.18 4.35 1.93 5.14 2.67 2.51 1.79 2.82 1.8 2.22 1.82 1.85 1.7 1.94 1.84 2.03 2.1 2.06 2.71

0.031* 0.041* <0.001*** 0.059 <0.001*** 0.01*** 0.015* 0.096 0.014* 0.089 0.070* 0.075 0.07 0.095 0.053 0.071 0.065 0.058 0.062 0.007***

line with Lovegrove and Siegfried (1986) who also found an increase in heuweltjie density by more than 100% from the arid parts in the north-west to the less arid south-east of the Western Cape. As suggested by Picker et al. (2007), the density of heuweltjies might be related to distances between termite populations, which tend to be narrower with higher plant productivity due to higher rainfall and soil fertility at sites with similar vegetation and soil types.

p

Vegetation cover, N¼101. Adjusted R2¼0.350, F (3,97)¼18.980, p¼<0.001 pH 0.417 ±0.081 5.168 <0.001 OC 0.462 ±0.085 5.428 <0.001 Fringe zone (F) 0.198 ±0.085 2.327 0.022 Cover chamaephytes [%], N¼101. Adjusted R2¼0.330, F (3,97)¼17.426, p<0.001 pH 0.425 ±0.082 5.190 <0.001 OC 0.408 ±0.086 4.725 <0.001 Fringe zone (F) 0.254 ±0.086 2.937 0.004 Cover therophytes (log10), N¼101. Adjusted R2¼0.226, F (2,98)¼15.569, p¼<0.001 Centre zone (C) 0.454 ±0.088 5.153 <0.001 Northing 0.177 ±0.088 2.013 0.047 2 Cover geophytes (rank) [%], N¼101. Adjusted R ¼0.095, F (2,98)¼6,248, p¼0.003 pH 0.295 ±0.097 3.034 0.003 C:N ratio 0.234 ±0.097 2.403 0.018 Species richness, N¼101. Adjusted R2¼0.266, F (3,97)¼13.088, p<0.001 pH 0.517 ±0.089 5.796 <0.001 C:N ratio 0.441 ±0.115 3.843 <0.001 OC 0.325 ±0.112 2.894 0.005 Simpson index (square root), N¼101. Adjusted R2¼0.035, F (1,99)¼4589 p¼0.035 Northing 0.210 ±0.098 2.142 0.035 Soft litter cover [%], N¼93. Adjusted R2¼0.045, F (1,91)¼5,359, p¼0.023 pH 0.236 ±0.102 2.315 0.023 Dead wood cover [%] (square root), N¼101. Adjusted R2¼0.122, F (3,97)¼ 5,631, p¼0.001 OC 0.231 ±0.094 2.452 0.016 Northing 0.223 ±0.095 2.355 0.021 pH 0.210 ±0.094 2.221 0.029 Bioturbation cover [%] (rank), N¼101. Adjusted R2¼0.210, F (2,98)¼14.256, p<0.001 Centre zone (C) 0.294 ±0.106 2.773 0.007 pH 0.246 ±0.106 2.321 0.022

4.2. Characterisation of heuweltjie zones The ANOVAs of soils and vegetation variables revealed strong differences among the three heuweltjie zones. Heuweltjie centres in the study area were characterised by increased soil pH, calcium carbonate, total organic carbon and total nitrogen content compared to the heuweltjie fringe and matrix soils. Increased pH and comparatively high calcium carbonate content on heuweltjie centres have been shown by several other studies on heuweltjies in southern Africa before (Francis et al., 2007; Kunz et al., 2012; Midgley and Musil, 1990; Midgley et al., 2012; Petersen, 2008). The relatively high pH and the occurrence of calcium carbonate in heuweltjie centres in an otherwise largely carbonate-free landscape (Midgley et al., 2012) can be related to the lentil-shaped calcrete under the surfaces of the heuweltjies (Francis et al., 2007; Petersen, 2008). The development of these calcretes is facilitated by the ratio of rainfall and evaporation in arid regions, as well as by soluble carbonates from respiratory carbon dioxide, that results in increased precipitation in reaction with calcium from organic decomposition processes (Midgley et al., 2012; Mujinya et al., 2011). Burrowing animals excavate this crust and bring fragments to the surface. With the higher nutrient content, the heuweltjies provide more favourable growing conditions than their surroundings. Also the significantly higher content of total N and OC found in the heuweltjie centres is in line with other studies on heuweltjies in Namaqualand (Kunz et al., 2012) and the Worcester-Robertson Karoo (Midgley and Musil, 1990). The nutrient richness of heuweltjie centres has been explained by their zoogenic origin:

34

U. Schmiedel et al. / Acta Oecologica 77 (2016) 27e36

termites harvest organic material and transport it into the heuweltjie (Midgley and Musil, 1990; Petersen, 2008). This reallocation of organic material leads to nutrient accumulation of the heuweltjies and relative impoverishment of their environment. The preferred utilization of the heuweltjies by burrowing animals enhances the nutrient accumulation. The three heuweltjie zones showed strong differences with regard to vegetation cover and structure. Total vegetation cover decreased centripetally. This was mainly due to the strong decrease of the chamaephyte cover on heuweltjie centres compared to the matrix, where chamaephytes are the dominant life form (LutherMosebach et al., 2012). The decrease of chamaephyte cover was only partly compensated by the increase of therophyte cover. A higher therophyte cover could be expected in years with seasonal rainfall at or above long-term average, whereas the annual rainfall in Soebatsfontein in the year of our data collection (2008, 113 mm rainfall) was well below the annual mean of 130 mm (Haarmeyer et al., 2010). Previous studies have explained the difference in life-form composition between zonal vegetation and heuweltjies by the increased soil nutrients and water supply on heuweltjies (Knight et al., 1989; Kunz et al., 2012; Midgley and Musil, 1990). Midgley and Musil (1990) found that the increased soil nutrient content (N, Ca), together with a better water supply, lead to a higher frequency and relative cover of deciduous plants and a higher nitrogen content in the leaves of plants growing on heuweltjies than in the matrix. The leaves of deciduous plants have been shown to be softer and more palatable (Skarpe et al., 2000), which again contributes to the attractiveness of heuweltjies to herbivores (Milton and Dean, 1990; Rahlao et al., 2008), which again maintain the nutrient richness on heuweltjies through defecation. Selective grazing can lead to overgrazing on heuweltjies, which in turn may result in bare soil patches in the landscape, which are only covered by annuals after sufficient rainfall. This pattern can be observed on large parts of the Heuweltjieveld vegetation of Namaqualand (Knight et al., 1989; Schmiedel, own observation). The replacement of perennial shrubs by annuals has been described as typical effects of overgrazing in the Succulent Karoo (Mucina et al., 2006; Todd and Hoffman, 1999) and for other biomes (Grime, 2001; McIntyre and Lavorel, 2001). The attractiveness of heuweltjies has not been reflected by the mean palatability values per plot in our study, since the ranking of the farmers mainly focussed on perennial species, which were largely absent from the heuweltjies. In addition to the larger herbivores, fossorial small mammals and the aardvark (Orycteropus afer) also prefer heuweltjies for their dens, due to their deep soils and low density of stones. Their activities together with the herbivory make heuweltjies a “hotspot of disturbance” (Esler and Cowling, 1995). The high disturbance intensity on heuweltjies is described by numerous studies (Esler and Cowling, 1995; Knight et al., 1989; Kunz et al., 2012; Lovegrove and Siegfried, 1986; Midgley and Musil, 1990). Esler and Cowling (1995), for instance, found in the southern Karoo differences in life-history traits of plants on and off heuweltjies, with the heuweltjie vegetation being characterised by traits attributed to opportunistic species (i.e. seed retention, slow dispersal spread over the year and low germination rate, see also Grime, 2001). These opportunistic plants colonize bioturbated heuweltjie centres successfully, where they complete their life cycle within a few weeks and are able to satisfy their relatively high nutrient requirements (Knight et al., 1989). 4.3. Impact of historic and recent grazing intensities on the three heuweltjie zones Due to its attractiveness to herbivores, heuweltjie vegetation

has been suggested as a suitable indicator for overgrazing (Stokes et al., 2009). We thus tested whether we can identify the effects of different historic grazing intensities on the three heuweltjie zones, based on life-form composition and diversity indices, as well as soil variables. We compared these effects with the effects of the deviating patterns of recent grazing intensity. Among the three heuweltjie zones, fringes were most responsive to long-term historic grazing intensity: total and chamaephyte cover, as well as OC and C:N ratio of heuweltjie fringes, decreased with increasing historic grazing pressure. Decrease of the total and perennial plant cover has been described as an effect of selective grazing and an indicator for heavy grazing in the Succulent Karoo of the winter rainfall zone (Todd and Hoffman, 1999), as well as the Savanna (Wesuls et al., 2013), and the shrub-dominated Karoo (Du Toit et al., 2011) of the semi-arid summer rainfall zone of southern Africa. The loss of OC in the heuweltjie fringe zone can be related to leaching at the steep slopes of the fringes, due to the intensive livestock trampling under high grazing intensity, which leads to visible signs of erosion particularly downslope of the heuweltjie centres (Schmiedel, own observation). OC loss can lead to a narrower C:N ratio in the fringe. However, other studies of grazing impact on soil chemistry found that both OC and total N increased with grazing intensity (Manley et al., 1995; Smet and Ward, 2006). These studies focussed on the grazing impact within 100e200 m from artificial water points, the so-called piospheres, in African Savannas and North American mixed-grass rangelands. Piospheres accumulate nutrients through the longer period the livestock stay around the water points. Heuweltjies in our study represent analogues to piospheres with high grazing impact and nutrient accumulation in the centre. Following this analogy, heuweltjie centres represent the highly disturbed area closest to the water points, also referred to as “sacrifice zone” (Andrew, 1988). Due to the attractiveness of the heuweltjie centres, they can become sacrifice zones even under moderate grazing intensity. Even though fringes form part of the heuweltjies and are also attractive for grazing, disturbance of the fringes is less intense than of the centres and therefore soil and vegetation variables on fringes show a more differentiated response to differences in grazing intensities. In contrast to the historic grazing, recent grazing intensity showed fewer effects on soil and vegetation. The only identified responses were increase of total vegetation cover at heuweltjie centres and of species richness in the matrix. Fringes showed the same effects but with low statistical significance. These seemingly positive responses to grazing can be attributed to the nearly inverse grazing intensity patterns of the historic and recent management. Several camps that have been overgrazed in the past have been rested after 2000, and others that have been underutilized in the past have been used intensively after 2000. The seemingly positive effects of recent high grazing intensity, with a stocking density of up to 194% of the assumed carrying capacity, are therefore more likely the still visible effect of the lower grazing intensity in these camps during the historic period. The absence of any observable negative effect of even very high grazing intensity can be explained by the relatively short period the vegetation was exposed to this high grazing intensity. When the farmland was handed over to the Soebatsfontein community in 2000, the farmers had very few animals. The herds grew slowly over the years due to lack of capital to buy large numbers of livestock. Hence, the grazing intensity assessed for the years 2008/2009 does not reflect the situation throughout the past eight years. Considering the change in grazing patterns, one could expect signs of vegetation recovery (e.g., higher total or chamaephyte vegetation cover) under low recent grazing intensity, particularly on heuweltjie centres that provide suitable conditions for fast

U. Schmiedel et al. / Acta Oecologica 77 (2016) 27e36

vegetation growth, due to their nutrient and water supply. This was not the case for our study. Rahlao et al. (2008) could indeed show a change from dominance of stem-succulent and deciduous shrubs to evergreen shrubs and trees on heuweltjies under complete protection from grazing. The recovery, however, was recorded after a resting period of 67 years and under more humid conditions of the southern Succulent Karoo (mean annual rainfall of 269 mm). Recovery in semi-arid and arid systems is very slow (Noy-Meir, 1973; Rahlao et al., 2008) and eight recent years of low grazing intensity did not have observable effect on the vegetation. 4.4. Conclusions The special habitat conditions of heuweltjie centres have a strong effect on the vegetation, which make them attractive to herbivores and a hotspot of disturbance with little response to land use intensity e even if grazing intensity is below the recommendation. Once transformed to a disturbed state, heuweltjie centres recover very slowly. Heuweltjie fringes were less strongly impacted than centres and showed the most differentiated responses in soil and vegetation among the three zones. Heuweltjie fringes are therefore more suitable as indicator patches for land-use effects than the centres. Acknowledgements The manuscript is based on the Diplom thesis of I.U.R. at the University of Hamburg under supervision of U.S., J.D. and A.G. Fieldwork, soil analysis, and data processing were conducted by I.U.R., and J.L.M. J.O. helped with the GLM. The paper has been conceptualized and drafted by U.S. and I.U.R., with contributions by all authors. We are grateful to Floors Brandt for providing the detailed data on stocking density during the years 1986e1999, to the BIOTA paraecologist Reginald Christiaan from Soebatsfontein for gathering the recent livestock numbers. Timo Labitzki provided the classification of landscape units and the grid point raster for the study area. We would like to thank the participants of the paperwriting seminar of the Biodiversity, Evolution and Ecology of Plants working group at the Biocentre Klein Flottbek of the University of Hamburg, for providing helpful comments on earlier versions of the manuscript. The Department of Environmental Affairs (DEA) of the Northern Cape Province, South Africa, kindly provided the research permits. The Soebatsfontein community generously hosted us during the fieldwork and shared their knowledge on the veld and information with us. We would also like to thank Annelise le Roux for her help with plant identification. We acknowledge the financial and organisational support by BIOTA Southern Africa, funded by BMBF under promotion number 01LC0624A2. The field work was financially supported by the Deutsche Kakteengesellschaft. The drafting of the paper was part of U.S.’s work within task 159 of Southern African Science Service Centre for Climate Change and Adaptive Land Management (SASSCAL) sponsored by the German Ministry for Education and Research (BMBF) (BMBF Reference no. 01 LG0905A). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.actao.2016.08.012. References Andrew, M.H., 1988. Grazing impact in relation to livestock watering points. TRENDS Ecol. Evol. 3, 336e339.

35

Bivand, R., 2014. Spdep: Spatial Dependence: Weighting Schemes, Statistics and Models. R package version 0.5-71. Bivand, R., Keitt, T., Rowlingson, B., 2014. Rgdal: Bindings for the Geospatial Data Abstraction Library. R package version 0.8-16. Bjornstad, O.N., 2013. ncf: Spatial Nonparametric Covariance Functions. R package version 1.1-5. Cramer, M.D., Midgley, J.J., 2015. The distribution and spatial patterning of mimalike mounds in South Africa suggests genesis through vegetation induced aeolian sediment deposition. J. Arid Environ. 119, 16e26. Dean, W.R.J., Milton, S.J., du Plessis, M.A., 1995. Where, why, and to what extent have rangeland in the Karoo, South Africa, desertified. Environ. Monit. Assess. 37, 103e110. Desmet, P.G., 2007. Namaqualand - a brief overview of the physical and floristic environment. J. Arid Environ. 70, 570e587. Diaz, S., Lavorel, S., McIntyre, S., Falczuk, V., Casanoves, F., Milchunas, D.G., Skarpe, C., Rusch, G., Sternberg, M., Noy-Meir, I., Landsberg, J., Zhang, W., Clark, H., Campbell, B.D., 2007. Plant trait responses to grazing - a global synthesis. Glob. Change Biol. 13, 313e341. DIN, 1995. ISO 10694 Bodenbeschaffenheit - Bestimmung von organischem Kohlenstoff und Gesamtkohlenstoff nach trockener Verbrennung (Elementaranalyse). Deutsches Institut für Normung e.V.: Beuth Verlag. DIN, 1997. ISO 11265 Bodenbeschaffenheit - Bestimmung der spezifischen elek€higkeit. Deutsches Institut für Normung e.V.: Beuth Verlag. trischen Leitfa Du Toit, G.v.N., Snyman, H.A., Malan, P.J., 2011. Physical impact of sheep grazing on arid Karoo subshrub/grass rangeland, South Africa. South Afr. J. Animal Sci. 41, 280e287. Esler, K.J., Cowling, R.M., 1995. The comparison of selected life-history characteristics of Mesembryanthema species occurring on and off Mima-like mounds (heuweltjies) in semi-arid southern Africa. Vegetatio 116, 41e50. ESRI Inc. 2005. ESRI® ArcMapTM. [9.1]. 2005. Francis, M.L., Fey, M.V., Prinsloo, H.P., Ellis, F., Mills, A.J., Medinski, T.V., 2007. Soils of Namaqualand: compensations for aridity. J. Arid Environ. 70, 588e603. Google Earth, 2003. “Soebatsfontein” 30
13’05.84’’ S and 17
34’46’’ E. Google Earth, 25 November 2003. 3 July 2007. Gotelli, N.J., Ellison, A.M., 2004. A Primer of Ecological Statistics. Sinauer, Sunderland, Mass. Grime, J.P., 2001. Plant Strategies, Vegetation Processes, and Ecosystem Properties. John Wiley & Sons Ltd., Chichester. € ngro €ft, A., Mager, D., Medinski, T., Mills, A., Petersen, A., Eschenbach, A., 2010. Gro Evaluation of the soil degradation state along fence-line contrasts. In: Schmiedel, U., Jürgens, N. (Eds.), Biodiversity in Southern Africa, Patterns and € ttingen & WindProcesses at Regional Scale, vol. 2. Klaus Hess Publishers, Go hoek, pp. 207e213. Haarmeyer, D.H., Luther-Mosebach, J., Dengler, J., Schmiedel, U., Finckh, M., Berger, K., Deckert, J., Domptail, S.E., Dreber, N., Gibreel, T., Grohmann, C., € ngro € ft, A., Haensler, A., Hanke, W., Hoffmann, A., Husted, L.B., Gro Kangombe, F.N., Keil, M., Krug, C.B., Labitzky, T., Linke, T., Mager, D., Mey, W., €pper, M., Muche, G., Naumann, C., Pellowski, M., Powrie, L.W., Pro Rutherford, M.C., Schneiderat, U., Strohbach, B.J., Vohland, K., Weber, B., Wesuls, D., Wisch, U., Zedda, L., Büdel, B., Darienko, T., Deutschewitz, K., Dojani, S., Erb, E., Falk, T., Friedl, T., Kanzler, S.-E., Limpricht, C., Linsenmair, K.E., Mohr, K., Oliver, T., Petersen, A., Rambold, G., Zeller, U., Austermühle, R., €sing, B.M., Classen, N., Dorendorf, J., Dorigo, W., Esler, K.J., Etzold, S., Bausch, J., Bo Graiff, A., Grotehusmann, L., Hecht, J., Hoyer, P., Kongor, R.Y., Lang, H., €wer, I.U., September, Z.M., Sop, T.K., van Lieckfeld, L.A.B., Oldeland, J., Peters, J., Ro Rooyen, M.W., Weber, J., Willer, J., Jürgens, N., 2010. The BIOTA observatories. In: Jürgens, N., Haarmeyer, D.H., Luther-Mosebach, J., Dengler, J., Finckh, M., Schmiedel, U. (Eds.), Biodiversity in Southern Africa, Patterns at Local Scale - the €ttingen & Windhoek, BIOTA Observatories, vol. 1. Klaus Hess Publishers, Go pp. 6e801. Hoffmann, B.D., 2010. Using ants for rangeland monitoring: global patterns in the responses of ant communities to grazing. Ecol. Indic. 10, 105e111. Illius, A.W., O'Connor, T.G., 1999. On the relevance of nonequilibrium concepts to arid and semiarid grazing systems. Ecol. Appl. 9, 798e813. Jackman, S., 2012. pscl: Classes and Methods for R Developed in the Political Science Computational Laboratory, R Package Version 1.04.4. Department of Political Science, Stanford University, Stanford, California. Jürgens, N., 2013. The biological underpinnings of Namib Desert fairy circles. Science 339, 1618e1621. Kiage, L.M., 2013. Perspectives on the assumed causes of land degradation in the rangelands of Sub-Saharan Africa. Prog. Phys. Geogr. 37, 664e684. Knight, R.S., Rebelo, A.G., Siegfried, W.R., 1989. Plant assemblages on mima-like earth mounds in the Clanwillian District, South Africa. South Afr. J. Bot. 55, 465e472. Kunz, N.S., Hoffman, M.T., Weber, B., 2012. Effects of heuweltjies and utilization on vegetation patterns in the Succulent Karoo, South Africa. J. Arid Environ. 87, 198e205. Linke, T., 2009. Ethnographische Rahmenbedingungen und rezente Entwicklungen der pastoralen Landnutzung in Soebatsfontein, Südafrika. Anthropology. Uni€t Hamburg, Hamburg, pp. 1e151. versita Loredo, I., Vuren, D.V., 1996. Reproductive Ecology of a population of the California Tiger Salamander. Copeia 4, 895e901. Lovegrove, B.G., Siegfried, W.R., 1986. Distribution and formation of Mima-like earth mounds in the western Cape Province of South Africa. South Afr. J. Sci. 82, 432e436.

36

U. Schmiedel et al. / Acta Oecologica 77 (2016) 27e36

Luther-Mosebach, J., Dengler, J., Schmiedel, U., Roewer, I.U., Labitzky, T., Groengroeft, A., 2012. A first formal classification of the Hardeveld vegetation in Namaqualand, South Africa. Appl. Veg. Sci. 15, 401e431. Manley, J.T., Schuman, G.E., Reeder, J.D., Hart, R.H., 1995. Rangeland soil carbon and nitrogen responses to grazing. J. Soil Water Conserv. 50, 294e298. McIntyre, S., Lavorel, S., 2001. Livestock grazing in subtropical pastures: steps in the analysis of attribute response and plant functional types. J. Ecol. 89, 209e226. Midgley, G.F., Musil, C.F., 1990. Substrate effects of zoogenic soil mounds on vegetation composition in the Worcester- Robertson valley, Cape Province. South Afr. J. Bot. 56, 158e166. Midgley, J.J., Harris, C., Harington, A., Potts, A.J., 2012. Geochemical perspectives on origins and consequences of heuweltjie formation in the southwestern Cape, South Africa. South Afr. J. Geol. 115, 577e588. Milton, S.J., Dean, W.R.J., 1990. Mima-like earth mounds in the southern and western Cape: are the origins so mysterious? South Afr. J. Sci. 86, 207e208. Milton, S.J., Yeaton, R.I., Dean, W.R.J., Vlok, J.H.J., Cowling, R.M., Richardson, D.M., Pierce, S.M., 1997. Succulent Karoo. Vegetation of Southern Africa. Cambridge University Press, Cambridge, pp. 131e166. Mucina, L., Jürgens, N., Le Roux, A., Rutherford, M.C., Schmiedel, U., Esler, K.J., Powrie, L.W., Desmet, P.G., Milton, S.J., 2006. Succulent Karoo biome. In: Mucina, L., Rutherford, M.C. (Eds.), The Vegetation of South Africa, Lesotho and Swaziland. South African National Biodiversity Institute Pretoria, pp. 221e229. Mujinya, B.B., Mees, F., Boeckx, P., Bode, S., Baert, G., Erens, H., Delefortrie, S., Verdoodt, A., Ngongo, M., van Ranst, E., 2011. The origin of carbonates in termite mounds of the Lubumbashi area, D.R. Congo. Geoderma 165, 95e105. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853e858. Noy-Meir, I., 1973. Desert ecosystems: environment and producers. Annu. Rev. Ecol. Syst. 54, 25e51. Okullo, P., Moe, S.R., 2012. Large herbivores maintain termite-caused differences in herbaceous species diversity patterns. Ecology 93, 2095e2103. Pebesma, E.J., 2004. Multivariable geostatistics in S: the gstat package. Comput. Geosci. 30, 683e691. Petersen, A., 2008. Pedodiversity of Southern African Drylands. Soil Science. University of Hamburg, Hamburg, pp. 1e375. Picker, M.D., Hoffman, M.T., Leverton, B., 2007. Density of Microhodotermes viator (Hodotermitidae) mounds in southern Africa in relation to rainfall and vegetative productivity gradients. J. Zool. 271, 37e44. Prieto-Benitez, S., Mendez, M., 2011. Effects of land management on the abundance and richness of spiders (Araneae): a meta-analysis. Biol. Conserv. 144, 683e691. PSA, 2002. Procedures for Soil Analysis. International Soil Reference and Information Centre, Wageningen. Quinn, G.P., Keough, M.J., 2002. Experimental Design and Data Analysis for Biologists. Cambridge University Press, Cambridge.

Rahlao, S.J., Hoffman, M.T., Todd, S.W., McGrath, K., 2008. Long-term vegetation change in the Succulent Karoo, South Africa following 67 years of rest from grazing. J. Arid Environ. 72, 808e819. R Core Team, 2013. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project. org. Reynolds, J.F., Grainger, A., Stafford Smith, D.M., Bastin, G., Garcia-Barrios, L., Fern andez, R.J., Janssen, M.A., Jürgens, N., Scholes, R.J., Veldkamp, A., Verstraete, M.M., Von Maltitz, G., Zdruli, P., 2011. Scientific concepts for an integrated analysis of desertification. Land Degrad. Dev. 22, 166e183. Reynolds, J.F., Smith, D.M.S., Lambin, E.F., Turner II, B.L., Mortimore, M., Batterbury, S.P.J., Downing, T.E., Dowlatabadi, H., Fernandez, R.J., Herrick, J.E., Huber-Sannwald, E., Jiang, H., Leemans, R., Lynam, T., Maestre, F.T., Ayarza, M., Walker, B., 2007. Global desertification: building a science for dryland development. Science 316, 847e851. Roig, V.G., Loyarte, M.M.G., Rosi, M.I., 1988. Ecological analysis of mound formation of the mima type in Río Quinto, Province of Cordoba, Argentina. Stud. Neotropical Fauna Environ. 23, 103e116. Skarpe, C., Bergstrom, R., Braten, A.L., Danell, K., 2000. Browsing in a heterogeneous savanna. Ecography 23, 632e640. Slotow, R., Rothstein, S.I., 1995. Influence of social status, distance from cover, and group size on feeding and vigilance in White-crowned Sparrows. Auk 112, 1024e1031. Smet, M., Ward, D., 2006. Soil quality gradients around water-points under different management systems in a semi-arid savanna, South Africa. J. Arid Environ. 64, 251e269. StatSoft Inc, 2007. STATISTICA für Windows (Software-System für Datenanalyse). [8.0]. Accessed 03.09.09. Stokes, C.J., Yeaton, R.I., Bayer, M.B., Bestelmeyer, B.T., 2009. Indicator patches: exploiting spatial heterogeneity to improve monitoring systems. Rangel. J. 31, 385e394. Todd, S.W., Hoffman, M.T., 1999. A fence-line contrast reveals effects of heavy grazing on plant diversity and community composition in Namaqualand, South Africa. Plant Ecol. 142, 169e178. Turner, J.S., 2003. Trace fossils and extended organisms: a physiological perspective. Palaeogeogr. Palaeoclimatol. Palaeoecol. 192, 15e31. Vetter, S., 2005. Rangelands at equilibrium and non-equilibrium: recent developments in the debate. J. Arid Environ. 62, 321e341. Wesuls, D., Pellowski, M., Suchrow, S., Oldeland, J., Jansen, F., Dengler, J., 2013. The grazing fingerprint: modelling species responses and trait patterns along grazing gradients in semi-arid Namibian rangelands. Ecol. Indic. 27, 61e70. Winslow, M.D., Vogt, J.V., Thomas, R.J., Sommer, S., Martius, C., Akhtar-Schuster, M., 2011. Science for improving the monitoring and assessment of dryland degradation. Land Degrad. Dev. 22, 145e149.