Lacunarity as a texture measure for a tropical forest landscape

摘要

Fragmentation and loss of tropical forest cover alters terrestrial plant and animal population dynamics, reduces biodiversity and carbon storage capacity, and, as a global phenomenon could affect regional and global climate patterns. Lacunarity as a texture measure can offer a simple solution to characterize the texture of tropical forest landscape and determine spatial patterns associated with ecological processes. Lacunarity quantifies the deviation from translational invariance by describing the distribution of gaps within a binary image at multiple scales. As lacunarity increases, the spatial arrangement of tropical forest gaps will also increase. In this study, we used the Spatial Modeler in Imagine as a graphic programming tool to calculate lacunarity indices for a tropical forest landscape in Southern Mexico and Northern Guatemala. Lacunarity indices were derived from classified Landsat MSS images acquired in 1974 and 1984. Random-generated binary images were also used to derive lacunarity indices and compared with the lacunarity of forest patterns derived from the classified MSS images. Tropical forest area declined about 17%, with most of the forest areas converted into pasture/grassland for grazing. During this period, lacunarity increased about 25%. Results of this study suggest that tropical forest fragmentation could be quantified with lacunarity measures. The study also demonstrated that the Spatial Modeler can be useful as a programming tool to quantify spatial patterns of tropical forest landscape by using remotely sensed data.