MAPPING ENVIRONMENTAL CARRYING-CAPACITY USING AN ARTIFICIAL NEURAL-NETWORK - A 1ST EXPERIMENT

The economic development activities of an increasing world population threaten the assimilative capacity of our environment and have stimula ted interest in the concept of environmental carrying capacity. While the pace of land transformations has encouraged the refinement of info rmation technologies such as satellite remote sensing to provide a syn optic view of earth-system processes, the volume of information these systems generate and the high level of expertise required to translate these data retard effective and timely land-management decision makin g. This paper introduces a methodology that employs an artificial neur al network trained to recognize categories of population support capac ity from satellite data acquired from the NOAA-AVHRR. The network, fun ctioning as an 'intelligent' mapping tool, achieved a clasification ac curacy of 77.5 per cent for the study site and points to the potential role a model of this type may play in land degradation monitoring.