An Object-Based Classification and Fragmentation Analysis of Land Use and Cover Change in the Keiskamma Catchment, Eastern Cape, South Africa

Land use/cover trends in the Keiskamma catchment were investigated using Multi-temporal Landsat satellite imagery from 1972 to 2006. Object-oriented post-classification comparison and fragmentation analyses were performed; and the Landscape-Function analysis was used to assess the current rangeland condition of proximal hillslopes and the riparian zone. Temporal land use/cover analyses revealed that intact vegetation has undergone a significant decline from 1972 to 2006. Increases in degraded vegetation and fragmentation, as well as in bare eroded soil, were identified particularly in the communal villages of the central Keiskamma catchment. The Landscape Organisation Index revealed very low vegetation connectivity in the communal rangelands that have weak local traditional institutions. Fragmentation analyses on the riparian and proximal hillslopes revealed evidence of increasing vegetation fragmentation from 1972 to 2006. Given the worsening degradation trends in the communal areas, a systematic re-allocation of state land in sections of the catchment is recommended. Strengthening local institutions that effectively manage vegetation resources will be required, in order to maintain optimum flow regimes in rivers and curb thicket degradation. Such measures should be sensitive to the socio-economic challenges facing the people in the communal areas.