Research on Knowledge-based Predictive Natural Resource Mapping System

Knowledge-based predictive mapping is a kind of mapping approach which links the existing geo-spatial data and knowledge to predict the status of unknown geographical phenomena. The knowledge reflects the relationships between the known environmental conditions and the phenomenon to be predicted. In the field of natural resource inventory, such a mapping approach is very valuable. However, standard GIS Software can not effectively support knowledge management, which is crucial to knowledge-based predictive natural resource mapping, so new tools and methods should be developed. In this paper, we present our approach to manage knowledge and how the knowledge base is integrated with environmental database to produce natural resource maps. We first introduce basic concepts of knowledge-based natural resource mapping. Then, we investigate the needs and limitations of current systems. Our own approach that utilizes fuzzy logic, object-oriented method and hierarchical structure to represent knowledge and organize knowledge base is then put forward. This paper also presents our prototype knowledge-based predictive natural resource mapping system and its application in soil mapping which is conducted in a small watershed in Northeast China. Finally, the paper also discusses the existing problems and research challenges. * [email protected]; phone 010-64889461