Characterizing Landscape Composition and Structure With Image Texture Parameters

There is a need to characterize landscapes using parameters that contain information beyond that of traditional land cover and land use categories. Surface composition and structure are desirable factors for applications such as modeling soil-plant-atmosphere interactions, understanding watershed hydrologic processes, and characterizing factors affecting plant and soil productivity, among others. Species composition, surface roughness, heterogeneity of vegetation cover, foliage clumping, and the spatial arrangement of landscape structures are among the desirable factors. These factors are potentially retrievable from the spatial domain of remote sensing data via analysis of image texture. Although image texture analysis tools are widely available, an understanding of the physical meaning of texture parameters from the perspective of landscape composition and structure is lacking. Investigations into the information content of image texture parameters from high spatial resolution imagery are presented. The roles of plant species architecture, plant canopy architecture, and landscape heterogeneity as factors controlling image texture parameters, and the implications for temporal stability of texture parameters relative to spectral parameters are discussed. Directions for future empirical and theoretical studies are suggested.