Landscape diversity in a conservation area and commercial and communal rangeland in Xeric Succulent Thicket, South Africa

Abstract

Nature reserves in Xeric Succulent Thicket of South Africa contain a greater diversityof wildlife and correspondingly a greater diversity of disturbance agents thanadjacent, unconserved freehold and communal rangeland. Although more lightlystocked, it is unknown whether protected areas contain a higher diversity oflandscape patches (i.e., sub-landscape features such as bush clumps,termitariums, bare patches or animal wallows) which could influence thereflectance value of a single pixel depicting a 20 × 20 marea in a SPOT satellite image, than unconserved land. Our key questions were:How does patch diversity in a nature reserve compare with that on commercial andcommunal rangeland? Can pixel diversity in a SPOT satellite image be used toquantify these differences? And, is there a correlation between reflectancediversity in a SPOT image and patch diversity on the ground? As a first step,the coefficients of variation (CV) for 10 groups of 12 picture-element (pixel)values of a dry season SPOT satellite image were calculated for two commercialfarms and a communal rangeland. The same data were collected on a naturereserve, 50 to 100 m inside the boundary between the reserve andthe freehold or communal rangeland. Next, we recorded the variety of 20 ×20 m plots on the ground, also in groups of 12 plots, at the samegeographical coordinates as the satellite-based measurements. The means of thesatellite-based and ground-based indeces were significantly and positivelycorrelated. In addition, the nature reserve displayed significantly higher pixelCVs than the communal rangeland, and also contained significantly higherground-based diversity indeces than the freehold, and possibly the communal,rangeland. We postulate that the higher landscape patchiness in the naturereserve is a result of the diversity of disturbances caused by wildlife(especially megaherbivores) coupled with naturally low stocking rates, while thelower diversity in the communally managed rangeland is the result of continuousheavy grazing coupled with intensive fuelwood harvesting. The satellite-basedtechnique is useful for identifying potential sites of high biodiversity,wherein more intensive sampling at a finer scale can be undertaken. It is,however, important to use dry season imagery because of the temporary‘masking’ effect of ephemerals during the wet season.