Ecological Change on Kalahari Rangelands: Permanent or Reversible?: Implications for sustainable agricultural use

Andrew Dougill, School of the Environment, University of Leeds, Leeds, LS2 9JT, UK;

Abstract

Ecological changes in savannas are increasingly seen as an inevitable consequence of the pastoral agricultural intensification required to sustain continued economic development of semi-arid regions. Patterns of vegetation change on the formalised TGLP ranches of the Kalahari, notably increased bush dominance, are clearly linked to intensified cattle grazing. However, contentious debates remain revolving around whether these ecological changes are permanent, and therefore represent land degradation, or whether changes are reversible and simply a consequence of sustainable agricultural use. To assess the permanence of ecological changes, studies reported here investigate the impact of intensified cattle grazing on the key determinants that affect vegetation growth, namely soil water and nutrient availability patterns. Studies on a range of scales show no evidence of a link between soil and vegetation changes suggesting that natural events such as drought or fires, or management actions such as prescribed burning may enable a return of bush encroached areas to a grass dominant state. However, the expansive nature of bush encroachment under cattle ranching and the rapid increases in borehole provision seen in the last decade imply that the coalescence of bush dominant zones is a very real threat, and one which is greatly reducing the fodder heterogeniety upon which opportunistic livestock management strategies depend. The implications with regard to the future sustainable use of Kalahari rangelands are far-reaching and need to be incorporated into future land use decision-making processes. In particular, the threat to future sustainable livestock production occurs as a function of the establishment of more borehole watering points and the fragmentation of management units as a result of fencing policies, rather than as a function of permanent ecological changes.

Key words:ecological change, Kalahari, livestock management, resilience, bush encroachment, TGLP.

Introduction

Agricultural intensification of domestic livestock production, especially on designated cattle ranches, throughout the Kalahari of Botswana has been experienced at an increasing rate in the last thirty years. Changes in land use practices have led to a series of ecological changes that remain the focus of debate regarding the implications for continued sustainable agricultural development and future management strategies required to support the maximum agricultural productivity for development of Botswana's rural economy. This paper analyses these debates and reviews recent studies undertaken to resolve key environmental uncertainties. In particular discussions aim to:-

Summarise the results of a range of environmental studies investigating ecological changes on Kalahari rangelands following the introduction of cattle grazing,

Detail the spatial characteristics of ecological changes (on ranch and micro scales) which provide ecosystem resilience to permanent changes and land degradation,

Outline environmental explanations of ecological changes in relation to determining factors of soil water and nutrient availability, direct herbivory impacts and changes in fire regimes,

Present a conceptual model for ecosystem dynamics in Kalahari rangelands,

Discuss the implications of environmental studies to the development of sustainable agricultural management strategies in the future Kalahari.

To achieve these aims, this paper will draw on the full range of previous and ongoing environmental studies undertaken in the Makoba Ranch Blocks, Central District, Botswana (Figure 1) since 1988. The information presented represents the most detailed analysis of environmental changes occurring on a select area of the Kalahari, synthesising ranch-scale ecological studies (Perkins and Thomas 1993a, b), micro-scale ecological studies (Dougill, Trodd and Shaw 1998b; Dougill and Trodd forthcoming), ranch-scale soil studies (Dougill and Cox 1995; Dougill, Thomas and Heathwaite 1999) and micro-scale soil studies (Dougill, Heathwaite and Thomas 1998a) and summarising ongoing remote sensing studies detailing spatial patterns of ecological change (Trodd and Dougill 1998). Discussions serve to highlight the environmental understanding gained through a single site monitoring project analysing changes in the natural resource base upon which rural livelihoods depend and examining the agricultural management implications of ecological community changes.

Background: Kalahari Environment and Agriculture

The Kalahari sandveld covers 2,500,000 km2 of southern Africa, including the western 75 % of Botswana (Figure 1). The area, despite its regular classification as a desert, experiences mean annual rainfall figures of between 150 and 600 mm (increasing eastwards), and is typified by mixed savanna grass and bush vegetation communities. Throughout Botswana, the Kalahari has extremely low relief and is dominated by deep aeolian sand deposits (over 5 m depth - Thomas 1988) with very low nutrient availability (Buckley, Gubb and Wasson 1987a; Buckley, Wasson and Gubb 1987b; Thomas and Shaw 1991). Despite the absence of even seasonal surface water, the Kalahari has long been recognised by agricultural developers as an untapped potential grazing resource (Debenham 1952). Livestock numbers in Botswana have continually increased since the early 20th century with the overall trend only occasionally disrupted by drought events (White 1993). The most dramatic increase in livestock numbers has occurred since independence in 1966 as the Botswana Government, supported by western development aid and the favoured access to European markets provided under the Lomé Convention, has pursued policies of intensifying livestock production. This has been made possible by the sinking of deep boreholes to utilise ancient groundwater resources. The agricultural policies adopted during this time, from the Livestock Development Project (LDP1) in 1970, through the Tribal Grazing Lands Policy (TGLP) in 1975, to the 1991 National Policy on Agricultural Development (Republic of Botswana 1991), despite some subtle differences (see White 1993 for discussion), remain centred on the intensification of livestock production based on a privatised cattle ranching framework. The continued adoption of agricultural development policies based on constrained privately-owned ranch blocks (Republic of Botswana 1991) implies that if we are to understand and predict the future sustainability of agricultural development of the Kalahari ecosystem we must investigate the environmental impacts of this management strategy. It is only through such a study, that environmental arguments can be added to the socioeconomic and cultural debates on the suitability of the cattle ranching system to the Kalahari (Tsimako 1991; White 1993; Thomas and Sporton 1997; Duraiappah and Perkins 1999). It is to this end that this paper is aimed, to allow rational incorporation of environmental understanding into the wider debates on the applicability of the ranching management strategies adopted in many parts of the Kalahari.

Makoba Ranches and Ranch Scale Patterns of Ecological Change

The Makoba Ranch Block (Figure 1) was chosen for these studies as an area typical of the structural adjustments occurring in the Kalahari following independence. It is one of twenty TGLP ranch blocks demarcated in 1975 and has experienced continued expansion in borehole provision since the early 1970s. Consequently, the ranch block provided a range of boreholes of different ages on which to analyse the ecological changes occurring as a result of the intensified cattle rearing.

Ecological surveys around eight boreholes (Figure 2) were initially undertaken during the 1988-89 wet season (Perkins 1991). Studies were based on a 'piosphere' approach (Georgiadis 1987) designed to investigate the link between the herbivore use intensity (HUI), which declines exponentially with distance away from the borehole, and changes in vegetation communities. Detailed vegetation analyses were undertaken for 25 m by 25 m quadrats at set distances of 0, 25, 50, 200, 400, 800, 1500, 3000 and 5000 m from the borehole on each study ranch. In depth analysis of these studies has been provided by Perkins and Thomas (1993a,b) and is simply summarised here. The most obvious and widespread ecological change recorded was the shift from grass to bush dominance, a process termed bush encroachment, which has been noted in many other parts of the Kalahari (Cole and Brown 1976; Cooke 1983; Skarpe 1990; Ringrose et al. 1990, 1996). Perkins and Thomas (1993a,b) found that bush encroachment generally remains confined to an area within 2 km of the borehole (Figure 3). Given the initial adoption of an 'eight kilometre rule' for distances between boreholes within the TGLP, there remained, at this time, substantial areas largely unaffected by such ecological changes. These unencroached areas remained predominantly grass-dominated, and were termed the 'grazing reserve' (Perkins and Thomas 1993a). Field ecological evidence suggests that such a bush-encroached - grazing reserve duality is the major ecological pattern on TGLP ranches. Consequently on a typical 6400 ha TGLP ranch block, ecological and fodder diversity remains within the ranch when only one central borehole is used to supply water to cattle.

Ecological heterogeneity in terms of such a mix of bush-dominant and grass-dominant areas and therefore of fodder supplies has been clearly linked to sustainable pastoral production in semi-arid areas by Scoones (1995), who stresses the important role played by bush fodder in drought years. Such fodder can limit cattle losses and consequently retain larger livestock herds from which to enable post-drought restocking as part of opportunistic management strategies (Westoby, Walker and Noy-Meir 1989; Behnke and Scoones 1993). Therefore, it is not surprising that studies investigating the impacts of ecological changes on pastoral production figures have, to date, found no significant decline in pastoral production through time (Vossen 1990; White 1993) with variations linked solely to rainfall histories. These arguments have been used to argue that Kalahari rangelands have not been subjected to major degradation (White 1993). However, Perkins and Thomas (1993b) also noted that bush encroachment is an expansive process with its greatest spatial extent occurring on the older boreholes. Given this recognition, combined with the increasing development pressures that has caused a relaxation of the 8 km spacing between boreholes (Tsimako 1991), there is now both an ecological and management potential for the coalescence of bush-dominated areas. Coalescence would have a major impact on the fodder diversity of rangelands and therefore sustainable pastoral production, a factor now being recognised by farmers on the Makoba TGLP ranches.

Further environmental studies have investigated further one of these initial study ranches, Uwe Aboo (borehole 1 in Figure 2). This focus on one ranch followed the realisation of the similarities in patterns of ecological changes between boreholes of all ages (Perkins and Thomas 1993a, b) and the need to focus on the causal links between cattle grazing, ecological changes and the determining factors of soil water and nutrient availability. The findings from these follow up studies are used here to present a conceptual model describing ecosystem dynamics in Kalahari rangelands, and to assess whether bush encroachment can be viewed as a permanent or reversible ecological change.

Ecological Theory and Research Implications

Prior to detailed analysis of the findings from these studies it is essential to understand wider debates concerning the functioning of semi-arid ecosystems and consequently the sustainability of agricultural systems adopted in these areas. Recent debates have highlighted weaknesses in traditional ecological equilibrium theories that had previously been applied to all global ecosystems, including highly dynamic semi-arid systems (e.g. Stoddart, Smith and Box 1975). Alternatively, it is now acknowledged that variability far exceeds any equilibrial tendencies in semi-arid ecosystems (Ellis and Swift 1988; Westoby, Walker and Noy-Meir 1989; Behnke and Scoones 1993). In addition, the weaknesses of the pastoral management strategies, such as fixed carrying capacities and constrained ranch blocks, adopted as a result of the acceptance of equilibrium theories are increasingly also being realised (Scoones 1995; Scoones et al. 1996; Warren 1995). The shift in ecological paradigm from equilibrium to non-equilibrium theories has also been accompanied by the closing of the divide between environmental science and socioeconomic and cultural studies (Warren 1995). This shift importantly has led to the recognition that indigenous pastoral strategies are carefully adapted to the spatial and temporal variability which characterise semi-arid ecosystems (Scoones 1995).

These changes in ecological understanding have many implications to environmental studies in the Kalahari. Non-equilibrium thinking and interdisciplinary frameworks have been applied throughout studies in this area, though as stated earlier environmental studies have maintained the focus of assessing the impacts of the TGLP demarcation into privately-owned ranch blocks to maintain the relevance to present agricultural management policies (Republic of Botswana, 1991). The transition in ecological thinking can explain much of the confusion in the debates on whether ecological changes following grazing intensification represent land degradation (see Cooke 1983; Skarpe 1990; White 1993; deQueiroz 1993; Dougill and Cox 1995 and Adams 1996 for a potted history of these lengthy debates). The principal confusion stems from the association between vegetation changes, such as the bush encroachment seen on the Makoba ranches, and land degradation. This direct association between vegetation change and land degradation goes against non-equilibrium ecological understanding, where rainfall variability plays a greater role in controlling plant growth than variations in grazing regimes (Ellis and Swift 1988; Friedel, Pickup and Nelson 1993). This rainfall variability implies that vegetation changes measured at a given time are often reversible and therefore not necessarily indicative of degradation which (by definition) implies an 'effectively permanent' decline in the rate at which an ecosystem can support agricultural production (Abel and Blaikie 1989; UNEP 1997).

Studies on the Makoba ranches progressed beyond the recognition of ecological changes to investigate the causes of the transition from grass-dominant to bush-dominant vegetation communities. Only through such investigation of the causes of ecological changes can questions be answered regarding the permanence of changes and therefore their significance in terms of land degradation and sustainability criteria.

Causes of Ecological Change in the Kalahari

Changes in semi-arid ecosystem structure (ratio of bush to grass cover) and ecological productivity are governed principally by soil water and soil nutrient availability, direct herbivory effects and fire regimes (Scholes and Walker 1993; Belsky 1994). Studies after the initial ranch-scale vegetation surveys therefore focused on the link between cattle grazing intensity and changes in these ecological determining factors. This has been achieved through detailed analysis of changes in soil hydrochemical characteristics and local-scale vegetation patterning studies to assess the relative effects of direct grazing impacts on vegetation, and the likely impacts on fire regimes of intensive grazing. As some direct management of herbivory pressures and fire regimes is possible, ecological changes can theoretically be classed as reversible provided soils have not changed greatly. Consequently, the recent shift in emphasis of land degradation studies to consider the impacts of management strategies on soil factors (e.g. Stocking 1995; UNEP 1997) is important, and provides a framework for the studies reported here.

Soil System Studies

To develop an understanding of the role assigned to soil water and nutrient availability in affecting ecosystem structure, and contributing to bush encroachment, a range of soil studies have been conducted at Uwe Aboo (Figure 2). These studies have focused on assessing the applicability of the 'two layer model of environmental change' proposed by Walker and Noy-Meir (1982), which has often been presented as the best explanation of bush encroachment in the Kalahari (e.g. Skarpe 1990, 1991; Perkins and Thomas 1993a, b). In this model, the balance between grass and bush production is determined by the relative availability of soil water and nutrients (principally inorganic nitrogen and phosphorus) in different rooting zones. Grasses out-compete bushes for water and nutrients in the topsoil layer (0 - 0.5 m depth), while bushes have the competitive advantage in the subsoil (below 0.5 m depth) (Walker et al. 1981; Belsky 1990). According to the two-layer model, cattle’s grazing changes this balance by suppressing grass growth and promoting soil water movement into the subsoil. At the same time, increased mineralization of organic nitrogen in cattle dung (which is more readily decomposed than residual plant litter) into nitrate (NO3--N), enhances the leaching of this vital plant nutrient into the subsoil. Consequently, the two layer model predicts that areas of intensive grazing experience significant increases in moisture content and inorganic nutrient concentrations in their subsoil, with bush encroachment being the inevitable (and therefore permanent) ecological consequence (Figure 4).

The applicability of the proposed link between grazing intensity, ecological change and changes in soil hydrochemical characteristics has been investigated using an integrated program of ranch-scale monitoring of soil water and nutrient availability and controlled small scale experimental studies investigating the impacts of vegetation removal and cattle dung additions on soil water movement, nutrient mineralization and nutrient leaching processes. Ranch scale studies were carried out on a 3 km transect north of the 1973 borehole on Uwe Aboo Ranch (Figure 5), with soils sampled at 25, 100, 400, 800, 1600 and 2800 m from the borehole and in an ungrazed grass-dominant control site at the neighbouring cordon fence. The analytical methodology followed and in depth analysis of these studies has been outlined by Dougill and Cox (1995) and elaborated by Dougill, Thomas and Heathwaite (1999). The methods followed and findings from the controlled experiments are reported in detail by Dougill, Heathwaite and Thomas (1998a). Consequently, the analysis of the soils results presented here simply provides an integrated summary of the soil-based research and the implications for discussions on the permanence of bush encroachment.

Ranch-scale studies highlight that for all the soil hydrochemical characteristics investigated (soil moisture contents, and extractable NO3--N, NH4+-N and PO43--P concentrations in both topsoil and subsoil layers) significant changes are only experienced within a distance of 100 m from the borehole (Dougill and Cox 1995). Vitally, detailed comparisons between soil water and soil nutrient availability at an intensively grazed bush encroached site and the ungrazed grass-dominant control site showed no significant differences in any of the soil hydrochemical characteristics investigated (Tables 1 and 2). These studies combine to provide strong evidence that intensive cattle grazing, and associated ecological changes, are not linked to (or caused by) changes in soil water and soil nutrient availability in Kalahari soils. These findings directly oppose the hypotheses proposed in the two-layer model (Figure 4).