CASE STUDY No. 1
BIODIVERSITY AND PERFORMANCE OF GRASSLAND ECOSYSTEMS IN COMMUNAL AND COMMERCIAL FARMING SYSTEMS IN SOUTH AFRICA

RESPONSIBLE TECHNICAL DIVISION

Plant Production and Protection Division Grassland and Pasture Crops Group

Stephen Reynolds

AUTHOR

David Hoare

Agricultural Research Council,

Range and Forage Institute,

South Africa

[email protected]

ABSTRACT

The Eastern Cape province of South Africa is home to a large human population that is rural, poor and has experienced little development. The major pressures on these communal rangelands are from intense herbivory, wood fuel harvesting and shifting cultivation. Eight study sites in communal and commercial farming systems were selected and a number of diversity indices were performed on floristic data from these sites. It was found that mean species richness, total number of species and species evenness were higher in the commercial than communal grasslands although the variation in species composition (internal heterogeneity) was similar. Analysis using plant functional traits indicated that there was a shift in the communal areas towards pioneer woody species that had little browsing, fuel or timber value, grass species with lower palatability and forb species that were tolerant of intense herbivory and disturbance to the soil surface. There were more exotic species per sample in the communal than commercial grasslands and these contributed a higher proportion of the total vegetation cover. Landcover data indicates a loss of woody vegetation cover and transformation by shifting cultivation. Production data from livestock numbers at a district level indicate no reduction in long-term stocking rates, but stock composition has changed. Data from stocking rates is problematic as a measure of production and must be considered to be inconclusive since other factors could maintain high stock numbers. There is evidence of improving rangeland condition in the area around towns, where animal biomass is also lower than was historically recorded.

INTRODUCTION

South Africa has a unique natural environment and biological diversity, for which it has been recognized at a national level that good management is essential for sustainable development. The wise use of resources requires a good understanding of the ecological processes that maintain the resource base and it is essential that the complex relationships between the social order and natural environment are well understood. The nature and intensity of resource use in South Africa has not been spatially uniform and different social structures have been imposed on the environment in different areas and at different times.

The study area is situated in the Eastern Cape Province of South Africa (Figure 1). The Eastern Cape comprises an area of approximately 170500 km². It is an area of extraordinary complexity and diversity, encompassing three regional biodiversity “hotspots”. Although literature on vegetation studies has been accumulating recently, it has in the past been understudied and to date, diversity patterns have never been adequately described. Environmental gradients in this area are very steep and complex leading to a rich mixture of floristic elements. This provides an ideal natural laboratory which should be taken advantage of for understanding patterns across a variety of scales. An understanding of diversity patterns in the region can lead to an understanding of ecosystem structure and function, resulting in better management of the area, especially with respect to understanding patterns across a variety of scales. An understanding of diversity patterns in the region can lead to an understanding of ecosystem structure and function, resulting in better management of the area, especially with respect to rangeland conservation planning. The region has vast economic potential from the perspectives of eco-tourism and agriculture (commercial and subsistence). Different management practices have an effect on species composition and thus diversity and it is important to understand this relationship in order to predict their effects on diversity patterns.

FIGURE 1

Southern Africa with the study area in South Africa shaded

The Eastern Cape is home to 15.5 percent (6.3 million) of South Africa’s total human population. Forty-nine percent of the province’s population is unemployed compared to the national figure of 25 percent (South African Institute for Race Relations, 1991). The majority of people of the Eastern Cape are more rural, significantly poorer and less developed than in other parts of South Africa with a large proportion of the population being reliant to some degree on natural resources for direct subsistence use or indirectly as a form of income generation. A rapidly growing population coupled with increasing poverty and urbanization have a compounding impact on the resource base.

The land use of the region is divided into communal and freehold tenure systems: communal tenure areas are heavily populated (56 people per km²) whereas in regions characterized by freehold tenure, where commercial farming generally takes place, population density is lower (3–6 people per km²) (Palmer et al., 1999). In communal areas livestock represents wealth and is a form of currency, and these areas are heavily stocked. In contrast, commercial farming systems in the freehold areas are characterized by land stocked at economically sustainable levels.

The north-eastern part of the province has been under communal land ownership for in excess of 100 years, whereas the westerly regions have been commercially farmed as stock ranches for close to 100 years (Figure 2). In the central regions there has been a recent shift from commercial to communal land ownership (Figure 2), thus providing an opportunity to compare communal impacts of different time periods. In the communal area the major pressures on the landscape and on biodiversity are from intense herbivory, fuel wood harvesting and shifting cultivation. Herbivory by domestic stock is expected to cause a change in species abundance and composition, wood harvesting changes vegetation structure and can lead to the loss of certain woody species and shifting cultivation disturbs the soil surface so that opportunities arise for invasion by exotic plant species and changes species composition by resetting the landscape to an earlier successional stage. The objectives of the study were to determine whether there are any significant differences in natural plant species and functional type diversity in communal and commercial grasslands of the Eastern Cape, South Africa.

FIGURE 2

Study area showing predominant farming types in the Eastern Cape, South Africa.

Sites for biodiversity analysis are indicated by circles (stripes: commercial, dots: communal.

P1: Aliwal North. P2: Stormberg.

P3: Amatola.

P4: Smaldeel.

C1: Mount Ayliff. C2: Tsolo.

C3: Umtata.

C4: Sada).

Physiography

The mountain regions of the study area are dominated by grassland, with numerous small patches of Afromontane forest on southern aspects, whereas lower-lying areas are covered by grassland, savanna and dwarf-shrub (Nama-karoo) vegetation. The incised river valleys of the study area are characterized by subtropical thicket, a vegetation type endemic to the Eastern Cape and containing a number of succulent species. This study concentrates on the grasslands in the mesic parts of the Eastern Cape for which detailed recent floristic survey data are available.

The dominant geological group in the study area is the Karoo Supergroup, comprising alternating bands of fine-grained sandstone, shale and mudstone (Maud, 1996). The soils in the study area may be divided into mountain and plain types. The soils of the mountain areas are generally shallow and weakly developed lithosols whereas on the plains soils may be shallow and poorly drained with high-clay subsoils or, in the dryer areas, sandy loams containing many boulders and gravel (Hartmann, 1988; Werger, 1980). Dolerite outcrops may yield more fertile, clay-rich soils. This study was located in formations of the Karoo Supergroup.

The climate according to the Köppen classification is Cfa (C = warm temperate climate with the coldest months18°C to +3°C, f = sufficient precipitation during all months, a = warmest month is over +22°C) (Schultze, 1947). The lowland regions (< 800 m) have hot summers and frost prone, cold winters. The average daily minima for the coldest months are below freezing for the whole study area. Winter frost is common and especially severe at high altitudes. The whole study area experiences maximum rainfall, in summer although a weak bimodal (spring and autumn) pattern exists in the central regions. Data from weather stations indicate a gradient of increasing annual rainfall from west to east ranging from 419 mm at Venterstad to 664 mm at Lady Grey. Surface response models (Dent et al., 1989) suggest that median annual rainfall in excess of 1000 mm occurs at high altitudes in the Witteberg and Drakensberg ranges and that rainfall increases with altitude and from west to east in the study area.

METHODOLOGY

Grasslands of the commercial farming regions of the study area have been floristically described in detail (Hoare & Bredenkamp, 1999, 2001), but no published descriptions are available for the communal rangelands. Floristic data from these areas has been accumulated from a number of unpublished surveys, including the author’s own unpublished data. These have already been collated into a single database (Mucina et al., 2000; McDonald, 1997) as an initial phase of a recent national vegetation mapping project (Mucina & Rutherford, 2002). The criteria for selecting from this database were that the data must contain full-floristic information, be geo-referenced to within 100 m accuracy (the basic expected accuracy using GPS with selective availability), and the plot-size must be known.

Eight study sites were chosen, four each within communal and commercial grazing systems (Figure 2). The sites were selected to be in grassland vegetation and, as far as possible, similar with respect to environmental conditions (rainfall, topography, elevation, geology, soils). Study sites could not be exactly matched with respect to these environmental conditions so it was necessary to measure species’ richness along gradients to determine whether there were any directional changes with respect to environmental gradients. The two major environmental gradients in the study area are elevation and rainfall (Hoare, 1997; Hoare & Bredenkamp, 1999, 2001). Species richness from a total of over 500 floristic samples in mixed communal and commercial grasslands covering a range of environmental conditions was compared to rainfall and elevation gradients.

The selected sites were a circular region with a radius of 10 km (area: 314 km²) within which 10 random samples of 10 x 10 m were selected in open grassland vegetation. A complete checklist of species was available for each sample, as well as a cover estimate for each species using a modified Braun-Blanquet cover-abundance scale (Table 1). The cover-abundance classes were converted to a central class value for the purposes of further analysis (Table 1).

From the 10 samples for each site the following indices were calculated:

1. Total number of species per site,
2. Mean number of species per sample,
3. Jackknife estimate of species richness (S’ = S+(n-1/n)k) per site,
4. Whittaker’s beta diversity index (ß = (S/)-1) as a measure of the variation in species composition among localities within each site, which distinguishes it from species turnover along gradients (Vellend 2001),
5. Simpson’s evenness index ( (1- p_i²) / (1-(1/s)) ), a Type II index sensitive to changes in abundant species (Krebs, 1989),
6. The proportional cover of the five most dominant species per sample.

TABLE 1

Braun-Blanquet cover-abundance classes used for estimating aerial cover of individual plant species.

A simple functional type analysis was performed on the species by grouping them into major life-form groups (Table 2). Life form has been used as an easily recognizable plant functional trait (PFT) (e.g. Tilman et al., 1997), but most previous studies have found that life form alone is too broad a classification by which to determine the relative sensitivity to environmental variability so the life-form categories were further sub-divided using life-cycle, morphological and utilization traits (Table 2). Traits were selected that are easy to measure (Box, 1996), but were linked to ecological function, e.g. response to disturbance (Kindsher & Wells, 1995; Lavorel et al., 1997; Campbell et al., 1999; McIntyre et al., 1999). At a global scale it has been possible to identify only a few traits that are consistently associated with disturbance (e.g. Lavorel et al., 1997; Diaz & Cabido, 1992). A multivariate procedure was used for deriving plant functional types from functional traits (Landsberg et al., 1999; McIntyre et al. ,1999).

The invasibility of ecosystems in the study area was evaluated on the basis of the presence and relative dominance of exotic species within floristically sampled vegetation. The causes and mechanisms of invasion were not investigated here and no data is available on the spatial extent of exotic invasions in the study area.

Land cover information for the evaluation of cultivation impacts was obtained from a national land cover database (Fairbanks et al., 2000). However, although shifting agriculture was observed on the ground, it is difficult to quantify from remotely-sensed data as fallow lands develop secondary grassland-type vegetation cover that within a single or two seasons is indistinguishable spectrally from adjacent natural grasslands. These data were therefore supplemented using field observations of land cover.

TABLE 2

Hierarchical system of plant functional traits used in determining plant functional types for an analysis of impacts of grazing, browsing, trampling, soil disturbance and wood harvesting in a comparison between communal and commercial farming areas of the Eastern Cape.

RESULTS

Diversity/environment relationships

There was no relationship between species’ richness per 100 m² sample and elevation (Figure 3a, r² = 0.004) or species’ richness per 100 m² sample and rainfall (Figure 3b, r² = 0.015) . However, there was a very weak positive relationship between species’ richness and environmental heterogeneity for commercial grasslands (r² = 0.164), measured as the standard deviation in elevation around each sample (Figure 4). The magnitude of this relationship with environmental heterogeneity is different for communal and commercial grasslands, with communal grasslands tending to have lower species’ richness for the same amount of environmental heterogeneity as the commercial grasslands. Using normalized difference vegetation index (NDVI) as a surrogate for production (Figure 5) there appeared to be similar species richness at low production levels for communal and commercial grasslands and an apparent differential in species richness at higher production levels with commercial grasslands having more species than communal grasslands. This difference requires further study at other locations.

FIGURE 3

Relationship between species’ richness and environment (A: elevation and B: median annual rainfall) for 679 plots of 100 m² in grasslands of the Eastern Cape.

FIGURE 4
Relationship between environmental heterogeneity, measured as standard deviation in elevation, and species’ richness for communal and commercial grasslands of the Eastern Cape.

FIGURE 5

Relationship between the Normalized Difference Vegetation Index from the NOAA AVHRR sensor and species’ richness for communal and commercial grasslands of the Eastern Cape.

Species diversity

There were an average of 33.8 species (32.6 without exotics) in the commercial grasslands and 26.3 (23.5 without exotics) in the communal grasslands per 100 m² sample (Table 3). There was a total of 145 species per site in the commercial and 126 species per site in the communal grasslands. Using a jackknife estimate of species’ richness this can be extrapolated to 209 and 190 species in the commercial and communal grasslands respectively (Table 3). There was a similar number of unique species per sample in communal and commercial grasslands suggesting that rates of compositional turnover amongst rare species were similar in both grassland systems.

TABLE 3

Diversity indices for sites in communal and commercial grasslands of the Eastern Cape. The values given here are mean values from 10 samples of 10 x 10 m per site.

This is supported by the measure for variation in species composition among localities within each site (Vellend, 2001), which is similar for communal (mean ß = 3.84) and commercial grasslands (mean ß = 3.30) (Table 3). Species’ dominance, measured as the proportional cover of the top five species, was on average 88 percent in the communal areas and 66 percent in the commercial areas (Table 4), indicating that fewer species are contributing to the proportional cover in communal than commercial grasslands. Species evenness, using Simpson’s index (Table 4), was 0.685 in the communal grasslands and 0.843 in the commercial grasslands indicating that proportional cover was more evenly distributed amongst species in the commercial grasslands than in the communal grasslands.

TABLE 4

Measures of species’ evenness and species’ dominance in communal and commecial grasslands of the Eastern Cape

There was a strong negative relationship between species’ dominance and species richness (Figure 6) in commercial grasslands (r² = 0.66), but this relationship was very weak in communal grasslands (r² = 0.07). The weakness of this relationship in the communal grasslands is probably due to the high levels of dominance in most samples thus providing fewer points lower down in the dominance scale against which to run a regression.

FIGURE 6

The effect of species’ dominance on species richness in communal and commercial grasslands of the Eastern Cape.

Functional types

Multivariate analysis of functional traits based on ecological responses to disturbance and utilization resulted in the definition of 20 plant functional types (Table 5). There were significant differences in the proportional representation of functional types amongst woody species, grasses and forbs. Communal areas had higher proportional cover of woody species in the functional groups representing low-growing, spreading trees or shrubs, often pioneers with thorns (PFT 1), and the group representing shrubs that are inedible for browsing (PFT 5; Figure 7). The commercial areas had higher proportional cover of single-stemmed upright trees with hard wood (PFT 2), trees with soft wood or that are succulent (PFT 3) and woody species favoured for browsing (PFT 4; Figure 7). There were six grass functional types of increasing palatability and grazing value (Table 5). Communal and commercial grasslands showed opposite trends with the proportional cover of grasses with lower grazing value increasing in the communal areas and decreasing in the commercial areas (Figure 7). The exception was perennial stoloniferous grasses (PFT 8), which were, unexpectedly, better represented in the commercial areas. They were replaced in the communal areas by prostrate forbs. All forbs were better represented in commercial than communal areas, except prostrate, non nitrogen-fixing forbs with underground storage organs (PFT 19; Figure 7), a group which includes the typical rosette-shaped, prostrate herbs with tap roots that are often found in areas with high grazing impact (Lavorel et al., 1997). Of note was the much higher representation of forbs in the commercial grasslands that belong to the class that lacks any apparent defence against herbivores (secondary compounds, prickles, etc. – PFT 17). The number of species in each of the major functional classes (trees/shrubs, grasses and forbs) was consistent between communal and commercial areas with approximately 70 being forbs, 25 percent percent grasses and the remainder woody species (Figure 7).

FIGURE 7

Proportional aerial cover of plant functional types in communal and commercial grasslands of the Eastern Cape divided into trees/shrubs, grasses and forbs. Plant functional type numbers are according to Table 5.

TABLE 5
Plant functional types (PFTs) derived from a list of functional traits occurring in species of communal and freehold grasslands of the Eastern Cape.

Species’ composition

There were 12 species, mostly grasses, that were identified as having a high frequency and percentage cover in either the communal or commercial grasslands or both. The composition and grazing value of these are provided in Table 6. Only two decreaser species appear on the list, namely Themeda triandra, one of the most important grazing grasses in southern Africa, and Hypparhenia hirta, which, even though it decreases under grazing pressure, is an indicator of disturbance and is only of moderate grazing value when the plants are still young. Themeda triandra was moderately better represented in the commercial grasslands, but was absent from a number of communal and commercial samples. The species that have a high importance value in the communal areas all have poor grazing value or are indicators of disturbance (Table 6). In contrast the important grass species in the commercial grasslands have moderate grazing value and there is a higher diversity of different grass species at the different sites that are able to be grazed relative to the communal areas.

TABLE 6

Species with a high importance (cover and frequency) in commercial and communal grasslands of the Eastern Cape

Exotic species

There were an average of 11.3 exotic species per site and 2.8 species per sample in the communal grasslands versus 7.0 and 1.2 species per site and sample for the commercial grasslands (Table 3). This is 8.9 and 4.8 percent of the relative floras of these two regions. Exotic species contributed 20 percent to the proportional cover of communal grasslands and only 1 percent to the cover of commercial grasslands (Table 4).

There was a weak positive relationship between total species richness and number of exotic species for communal grasslands (r² = 0.26), but not in commercial grasslands (r² = 0.09). There was a weak positive relationship between exotic species’ cover and species’ richness in communal grasslands (r² = 0.22).

Land cover

Land cover data derived from Landsat TM satellite data for the communal portions of the study area indicated approximately 21 percent cover by cultivation and 69 percent natural vegetation (National Land cover Database), but observations in the field suggest that transformation is much greater than this (Table 7). A random sample of 250 sites visited in the field in the communal areas during 2002 indicated that only 33 percent are still natural vegetation, the remainder being transformed by cultivation, urbanization, plantations and other land cover categories.

TABLE 7

Land cover statistics (percentage cover) for communal areas from two different sources.

The major discrepancy between satellite and ground-based statistics are the fact that the satellite imagery appears unable to detect secondary grassland on previously cultivated lands as a separate class to natural grassland. This indicates that large parts of the cultivated areas are left fallow in any one year. Evidence of landscape transformation can be observed from satellite imagery (Figure 8) as well as from observations in the field, where woodland remnants containing species not able to be utilized for timber or fuel may often be seen in areas where woodland previously occurred (Figure 9).

FIGURE 8
Landsat TM image of portion of study area showing extent to which woodland loss has occurred. The expectation is that the majority of this image be covered by woodland, but much of this has been transformed to secondary grassland.

FIGURE 9
Photograph of small settlement within the boundaries of the Landsat TM image shown in Fig. 8 showing remnants of woodland. The species remaining are stem-succulent tree species that have poor wood quality.

Production

The only available data to evaluate whether system rundown is occurring comes from livestock numbers on a district level. Total livestock biomass reveals a degree of variability between years, but no indication that carrying capacity is declining (Figure 10). Although there appears to be more variability in later years and some downward trend for the Mount Ayliff district after 1990, this trend may be attributed to inter-annual rainfall variability or social factors. One indicator of potential degradation is the fact that the livestock composition has changed from predominantly cattle to predominantly goats in the last decade, goats being able to utilize components of the vegetation that are unavailable to cattle. However, the control of stock numbers and the human factors explaining stocking rates are relatively complex and are not investigated further here. Whether vegetation transformation can be invoked as an explanation for stocking rates and composition is not known, at this stage.

FIGURE 10
Total numbers of livestock units (LSUs) for the districts of Tsolo and Mount Ayliff in the communal grasslands of the Eastern Cape.

DISCUSSION

The communal grasslands of the Eastern Cape have come under severe utilization pressure from a number of factors. It is clear from land cover information that shifting cultivation and betterment planning has modified a large area of the communal landscape and that much of what is considered natural grassland is, in fact, secondary. In addition, expectations in terms of woody vegetation cover are much lower than would be expected from vegetation distribution models and pictorial evidence indicates that there are woodland remnants far away from any existing woodlands. Add to these two factors the high stocking rates measured for almost a century and a picture emerges of the severe pressure on all components of biodiversity in communal grasslands of the Eastern Cape.

This case study has been able to demonstrate that the particular land-use pressures in the communal parts of the study area have led to a reduction in indigenous species richness and increasing dominance by fewer species, although internal rates of species’ turnover are unchanged. Functional type analysis demonstrates that those woody species most susceptible to over-harvesting or intense browsing have been reduced as a proportion of the flora in communal areas. Similarly, those grasses most susceptible to overgrazing are found in lower proportions in the communal than commercial grasslands and those forbs most tolerant of soil disturbance and high grazing pressure dominate communal grasslands.

The different grazing systems has a severe impact on the forb component of the flora, with only 12 percent of forb species shared between the Amatola commercial and Umtata communal sites. A greater degree of floristic congruency had been expected because the two sites are geographically close to one another, environmentally more similar than any other two sites and are within identical grassland vegetation types (Mucina & Rutherford, 2002). Almost 75 percent of the species are found in the forb functional groups and it is in this group that the most species change that may affect overall diversity appears to be occurring.

One of the major effects of the grazing pressure in the communal grasslands has been a reduction in species evenness (Table 4). Few studies have evaluated the effect of reduced biodiversity on energy flow and nutrient cycling, but Wilsey and Potvin (2000) determined that higher levels of evenness resulted in higher total biomass irrespective of the identity of the dominant species. Evenness may change with little effect on richness, as is the case here, but it is still regarded as a negative response to environmental stress.

There are more exotic species per sample in the communal areas and exotic species contribute a large proportion of the aerial cover of the vegetation in communal grasslands. Tilman (1999) states that invasibility is equally dependent on species composition, disturbance and other factors as on species’ richness. However, this study was unable to support this hypothesis as there was no positive effect of reduced species richness on species invasibility. Although there were more exotic species in the communal than the commercial grazing areas, no causal link can be established between reduced species richness and ecosystem invasibility in this type of correlative study (Rosenzweig, 1995). It is probably the higher disturbance regime that is contributing to the increased occurrence of exotic species in communal grasslands. High exotic cover is known to affect indigenous species cover and species richness also declines, but it is impossible to determine whether it was reduced species richness (relative to potential species richness) that led to invasion in the first place in this study (as per Tilman, 1997). Exotic species appeared to make a positive contribution to species richness and, under the intense grazing pressure, to be behaving biologically as indigenous species, i.e. have become naturalized.

There was no relationship between rainfall or elevation and richness across the broad study area so it is assumed that these factors had no influence on the results presented here. However, a comparison between communal and commercial areas for elevation heterogeneity and vegetation activity (NDVI) as determinants of species richness produced positive results, although the relationship was very poor. Environmental heterogeneity measured as the standard deviation in elevation showed a partially hump-shaped curve, but this curve was lower in the communal areas relative to the commercial areas, i.e. species richness was lower in communal areas than commercial areas for the same levels of environmental heterogeneity. It may be that the intense utilization pressure is homogenizing the species distribution patterns to some degree.

There was no apparent reduction in ecosystem production despite intense grazing pressure across many years. This is based on stocking rates (Figure 10), which appear not to have decreased despite many years of high grazing pressure. However, this is a problematic measure to use and does not take into account a host of human factors that could play a role in determining stocking rates including other inputs that could support animal numbers. There is some evidence from satellite data that production in the communal areas is less than expected for the same amount of rainfall, but the results are not conclusive and further research is currently being undertaken. The results for grassland production, as presented here, are, therefore, problematic and inconclusive. It appears that in these highly productive communal grasslands that production levels (in terms of stock numbers) are able to be maintained at high levels because functional diversity has not changed (see Tilman et al., 1997) and that the only risk is that there is no standing biomass to buffer environmental uncertainty. The social response to drought is usually a reduction in stock numbers by translocation, harvesting or mortality. Of more concern from the type of information emanating from the current study is creeping environmental degradation - loss of environmental quality in imperceptible amounts that eventually emerges as a biodiversity issue.

The grasslands at the Umtata site provided some evidence that recovery of these grasslands may be possible. The grass component of this site has a number of species of high grazing value, although the forb component of the grasslands is still severely transformed. This site has undergone a reduction in stocking rates as urbanization levels have increased. The response is that overgrazed grass species have recovered and those of poor grazing value have become proportionately lower. Further reduction in grazing pressure could result in the prostrate forbs becoming overgrown, which would initially lead to a reduction in biodiversity, but indigenous forbs of other functional classes would eventually become re-established depending on their availability in the landscape. This trend in improving rangeland condition has been observed around other towns, e.g. Butterworth and Lusikisiki, where NDVI values are higher and animal biomass is lower (Palmer, pers. comm.).

Currently there is a weak national policy framework for dealing with the kinds of biodiversity/production management issues outlined here. Land redistribution occurs at a national level to address some land tenureship issues, but implementation has been slow. There are a number of research projects managed by the Agricultural Research Council and funded by the National Department of Agriculture that are attempting to quantify some of the land degradation problems in communal areas. The intention is that outcomes from this type of research will guide the National Department in developing a new legislative framework, but this process is still in its very early stages and has not yet developed into a structured programme although there are dedicated personnel at a national level working on such a programme. The current research falls partially within the sphere of this type of policy development research framework, although it is sometimes not clear to researchers what the influence of the particular research programme may be. A third aspect of management policy development is initiatives managed by the provincial government, such as the erection of fencing, development of skills and further research projects centred on management issues. There are also extension officers in the local areas, but they have poor training and little infrastructural support so they are largely ineffectual in modifying management behaviour. Another national level intervention is the so-called Landcare project, which is an attempt to develop partnerships between the private sector and local communities with the aim of improving land management. These projects often have an economic focus and attempt to instil some entrepreneurial spirit in local communities, i.e. show local communities that they can benefit financially by managing their land in a more sustainable way. It appears that the major stumbling block for many of these projects is social inertia and customs.

The methods applied here have not been widely used in research of this nature in southern Africa and also form only a small part of the range of methodologies employed for investigating the stability of production systems. However, the particular suite of methodologies used here specifically addresses biodiversity patterns and can be used in any other natural forage production system with equal chance of successfully extracting underlying ecological processes. The intention is to attempt comparisons across different natural vegetation boundaries using this set of methods to see if consistent results can be obtained. A number of organizations made an input, directly or indirectly, into this project, namely three universities, two ARC institutes, the Eastern Cape provincial government, the national government via the Department of Agriculture, and the FAO. There are strengths and weaknesses in this grouping since it brings together a variety of different research and management approaches, but appears to lack a single purpose, which maybe it does. A lesson learnt from fieldwork exercises is that there are conflicting views and constraints on what data to collect due to time and financial constraints and differing research objectives, but the collection of basic ecological and floristic data appears to be able to yield some very powerful results.

Sustainable agriculture relies on ecosystems with integrity and this ecological integrity is dependent on the composition and interaction of its constituent organisms. Natural biodiversity is one component of this integrity and, based on the type of work that ecologists such as Tilman have performed, it appears to play an important role in the stability of natural systems. The challenge is to determine how important biodiversity is as a factor on its own and whether it is an indicator of ecosystem health at a higher level.

REFERENCES

Box, E.O. 1996. Plant functional types and climate at the global scale. Journal of Vegetation Science 7: 309–320.

Campbell, B.D., Stafford Smith, D.M. and Ash, A.J. 1999. A rule-based model for the functional analysis of vegetation change in Australasian grasslands. Journal of Vegetation Science 10: 723–730.

Dent, M.C., Lynch, S.D. & Schulze, R.E. 1989. Mapping mean annual and other rainfall statistics in southern Africa. Department of Agricultural Engineering, University of Natal. ACRU Report No. 27. Massachusetts: Clark University.

Diaz, S. and Cabido, M. 1992. Morphological analysis of herbaceous communities under different grazing regimes. Journal of Vegetation Science 3: 689–696.

Fairbanks, D.H.K., Thompson, M.W., Vink, D.E., Newby, T.S., van den Berg, H.M and Everard, D.A. 2000. The South African Land-Cover Characteristics Database: a synopsis of the landscape. South African Journal of Science 96: 69–82.

Hoare, D.B. 1997. Syntaxonomy and synecology of the grasslands of the southern part of the Eastern Cape. M.Sc. thesis, University of Pretoria, Pretoria.

Hoare, D.B. and Bredenkamp, G.J. 1999. Grassland communities of the Amatola/Winterberg mountain region of the Eastern Cape, South Africa. South African Journal of Botany 64: 44–61.

Hoare, D.B. and Bredenkamp, G.J. 2001. Syntaxonomy and environmental gradients of the grasslands of the Stormberg / Drakensberg mountain region of the Eastern Cape, South Africa. South African Journal of Botany 67: 595–608.

Hartmann, M.O. 1988. The soils of the Eastern Cape. In: M.N. Bruton & F.W. Gess. (eds.) Towards an environmental plan for the Eastern Cape. Rhodes University, Grahamstown.

Kindsher, K. and Wells, P.V. 1995. Prairie plant guilds: a multivariate analysis of prairie species based on ecological and morphological traits. Vegetation 117: 29–50.

Krebs, C.J. 1989. Ecological methodology. University of British Columbia.

Landsberg, J., Lavorel, S. and Stol, J. 1999. Grazing response groups among understorey plants in arid rangelands. Journal of Vegetation Science 10: 683–696.

Lavorel, S., McIntyre, S., Landsberg, J. and Forbes, T.D.A. 1997. Plant functional classifications: from general groups to specific groups based on response to disturbance. TREE 12: 474–478.

Low, A.B. & Rebelo, A.G. (eds.) 1996. Vegetation of South Africa, Lesotho and Swaziland. Companion to the vegetation map of South Africa, Lesotho and Swaziland. Dept. of Environmental Affairs and Tourism, Pretoria.

Maud, R. 1996. The macro-geomorphology of the Eastern Cape. In: C. Lewis (ed.) The geomorphology of the Eastern Cape, South Africa. Grocott & Sherry, Grahamstown.

McDonald, D.J. 1997. VEGMAP: a collaborative project for a new vegetation map of southern Africa. South African Journal of Science 93: 424–426.

McIntyre, S. Lavorel, S., Landsberg, J. and Forbes, T.D.A. 1999. Disturbance response in vegetation—towards a global perspective on functional traits. Journal of Vegetation Science 10: 621–630.

Mucina, L, Bredenkamp, G.J., Hoare, D.B & McDonald, D.J. 2000. A National Vegetation Database for South Africa South African Journal of Science 96: 1–2.

Mucina, L. and Rutherford, M.C. (eds) 2002. Vegetation map of South Africa, Lesotho and Swaziland: an illustrated guide. Strelitzia xx, National Botanical Institute, Pretoria.

Palmer, A.R., Ainslie, A.M. and Hoffmann, M.T. 1999. Sustainability of commercial and communal rangeland systems in southern Africa. Proceedings of the VIth International Rangelands Congress, Townsville, Australia.

Rosenzweig, M.L. 1995. Species diversity in space and time. Cambridge University Press, Cambridge.

Schultze, B.R. 1947. The climate of South Africa according to the classification of Köppen and Thornwaite. South African Geographic Journal 29: 32 – 42. South African Institute of Race Relations. 1991. Fast Facts 1, 2. Johannesburg.

Tilman, D., Knops, J., Wedin, D., Reich, P., Ritchie and Siemann, E. 1997. The influence of functional diversity and composition on ecosystem processes. Science 277: 1300–1305.

Tilman, D. 1997. Community invasibility, recruitment limitation, and grassland biodiversity. Ecology 78: 81–92.

Tilman, D. 1999. The ecological consequences of changes in biodiversity: a search for general principles. Ecology 80: 1455–1474.

Velland, M. 2001. Do commonly used indices of ß-diversity measure species turnover? Journal of vegetation Science 12: 545–552.

Werger, M.J.A. 1980. A phytosociological study of the upper Orange River valley. Memoirs of the Botanical Survey of South Africa No. 46. Botanical Research Institute, Pretoria.

Wilsey, B.J. and Potvin, C. 2000. Biodiversity and ecosystem functioning: importance of species evenness in an old field. Ecology 81: 887–892.