Multi-temporal Remote Sensing Land-cover Change Detection for Biodiversity Assessment in the Berg River Catchment

Abstract

Due to the intimate relationship that exists between land cover and biodiversity it is possible to draw inferences on the current state of the biodiversity of an area, assess the likely future pressures and plan accordingly based on an analysis of land-cover change. As a means of assessing the state of biodiversity in the Cape Floristic Region, two land-cover maps (1986/7 and 2007) were developed and demonstrated for the Berg River catchment in the Western Cape province of South Africa using multispectral Landsat Thematic Mapper (TM) data. The land-cover maps were produced to an accuracy of 85% using an object-orientated nearest neighbour supervised classification. The existing vegetation types of South Africa data set were superimposed on the newly classified remnants of natural vegetation to model changes in biodiversity. It was found that the area occupied by natural vegetation increased by more than 14%, suggesting an increase in biodiversity from 1986/7 to 2007. Considerable variation between vegetation types was, however, recorded. The land cover mapping, change analysis and biodiversity modelling methods employed by this study show that land-cover change analysis provides an ideal platform from which to initiate more intensive analyses of biodiversity change and conservation. Some limitations to the use of Landsat imagery for biodiversity monitoring are discussed.