Representing continuous spatial variation of geographic phenomena using fuzzy logic

Variation of geographical phenomena is often gradual and continuous over space. A discrete model based on Boolean logic has been used to portray (map) this gradual and continuous spatial variation. Due to the limitation of this discrete model, spatial variation of geographic phenomena is portrayed by discrete and distinct spatial units, and changes only occur at the boundaries of these units. As a result, the gradual and continuous nature of geographical phenomena is not retained in this discrete model. The paper presents a fuzzy logic based model (referred to as a similarity model). The model consists of two components: a fuzzy representation of geographic phenomena in the attribute domain and a raster representation in the spatial domain. With the use of fuzzy logic and raster representation, a given geographic phenomenon is perceived as a continuum in both the attribute and spatial domains. A case study conducted in the area of soil survey shows that spatial variation of soils was better represented under the similarity model than under the discrete model. It was also found that information retained under the similarity model allows the quantification of uncertainty when one chooses to harden the similarity information into the conventional Boolean representation.