Landscape fragmentation in South Coast Renosterveld, South Africa, in relation to rainfall and topography

Abstract

The South Coast Renosterveld has been fragmented extensively by agriculture. The extent of this
fragmentation in terms of overall habitat loss, fragment sizes and fragment numbers has not been described previously, thereby limiting the development of conservation strategies for this vegetation type. Patterns of renosterveld loss in three sectors along a west–east gradient were described using LANDSAT imagery and a Geographical Information System-based program (FRAGSTATS) for spatial pattern analysis. These patterns were then correlated with rainfall and topography measures, which are indicators of agricultural potential. Over 80% of the South Coast Renosterveld has been cultivated. Fragmentation levels increased significantly from east to west, with 33% of natural vegetation remaining in the east and only 4% in the west. Topographical variables were the strongest predictors of patterns of renosterveld loss, with fragments being largely confined to slopes too steep
for ploughing; they therefore face little risk of future cultivation. These results have implications for conservation planning options for the South Coast Renosterveld. There is the potential for large reserves in the east, as well as corridor reserves along major river valleys, but for only small, isolated reserves in the west.

Introduction

Highly fragmented landscapes pose particularly difficult problems for conservation planning. This is because remnants of natural habitat tend to have a high conservation value and are also highly vulnerable to a wide array of threats (Saunders et al. 1991; Fahrig & Merriam 1994). Thus, many remnants qualify as priorities for conservation action, despite their lack of appeal to conservation authorities and lobby groups (Pressey 1994; Balmford 1996). The high conservation value of fragments is associated with high irreplaceability: as a result of large-scale habitat loss, each fragment potentially can make a high contribution to a reservation goal; alternatively, the loss of a fragment may greatly restrict options for attaining a representative reserve system (Pressey et al. 1994, 1995). Their location in an agricultural matrix makes fragments highly vulnerable to a wide array of processes that threaten the long-term maintenance of biodiversity (Terborgh & Winter 1980; Gilpin & Soulé 1986). Fragmentation is invariably associated with land systems where conservation competes poorly with other forms of land use such as agriculture and urbanisation. This is certainly true of renosterveld, a fire-prone, small-leaved, grassy shrubland of South Africa’s Cape Floristic Region (Low & Rebelo 1996). Renosterveld, renowned for its spectacularly rich geophyte flora (Cowling 1990; Johnson 1992), is largely associated with shale-derived and moderately fertile lowland soils, making this vegetation type highly suitable for cereal cultivation (Hoffman 1997). The West Coast Renosterveld, which covers an undulating coastal plain north of Cape Town (Fig. 1), has been extensively transformed by cereal and pasture crops: only 3% remained in 1988 as isolated fragments, largely on lands too steep for agriculture (McDowell 1988; Heydenrych & Littlewort 1995). The South Coast Renosterveld, which is the focus of this paper, is mostly confined to the semi-arid to subhumid (350–600 mm year–1) coastal forelands of the southern Cape of South Africa (Fig. 1). This vegetation type is apparently less fragmented than the West Coast Renosterveld (Cowling et al. 1986; Rebelo 1995), although no quantitative data exist.

For centuries, the South Coast Renosterveld was grazed by livestock initially belonging to Khoi-Khoi pastoralists and later to Dutch settlers (Rebelo 1995). Mechanisation after World War I facilitated large-scale, intensive agriculture, and an estimated 160 000 ha of natural vegetation was cleared for cereals and pastures between 1918 and 1990 (Cowling et al. 1986; Hoffman 1997). However, estimates of the amount of loss of natural vegetation are crude (Moll & Bossi 1984; Low & Rebelo 1996) and no account of the number and sizes of renosterveld fragments exists. Despite its critical conservation status, only 0.8% of the South Coast Renosterveld is formally conserved (Rebelo 1992). An overview of patterns of renosterveld loss, together with a knowledge of smaller-scale patterns and processes (Kemper 1997) can provide useful information for developing strategies for renosterveld conservation.

The establishment of a predictive understanding of transformation of natural vegetation will allow estimates of future threats to the remaining vegetation (Pressey et al. 1996). Reliable winter rainfall is a prerequisite for cereal cultivation in the warm temperate Western Cape. Rainfall patterns across the distribution of the South Coast Renosterveld change from west to east, with winter rainfall in the west and a bimodal spring–autumn regime in the east (Deacon et al. 1992). Accordingly, we predict less fragmentation in the eastern sector where winter rainfall is less reliable. In addition, slope strongly influences agricultural practices. In guidelines set by the Agricultural Resources Act, land can be worked on slopes of <1.8°; on slopes between 1.8 and 8.1°, land can be worked but measures to prevent erosion, such as the construction of contours, must be taken; land with slopes of >8.1°
cannot be worked. Slope also changes from west to east, with gently rolling, almost flat country in the west being gradually replaced by steeper country in the east. Therefore, we predict a strong relationship between steep topography and patterns of renosterveld loss.

In this paper, we examine the South Coast Renosterveld landscape fragmentation patterns in terms of overall habitat loss, fragmentation sizes and fragmentation numbers. Our objectives were: (i) to provide a descriptive, broad, regional overview of fragmentation patterns in the South Coast Renosterveld, (ii) to describe how these patterns change from west to east; and (iii) to model the extent of habitat loss in terms of rainfall regime and topographical features. Finally, we discuss the implications of our results in terms of future options for conserving this vegetation type.