A study of habitat fragmentation in Southeastern Brazil using remote sensing and geographic information systems (GIS)
Nenhuma Miniatura disponível
Garcia, G. J.
Título da Revista
ISSN da Revista
Título de Volume
The purpose of this work was to study fragmentation of forest formations (mesophytic forest, riparian woodland and savannah vegetation (cerrado)) in a 15,774-ha study area located in the Municipal District of Botucatu in Southeastern Brazil (São Paulo State). A land use and land cover map was made from a color composition of a Landsat-5 thematic mapper (TM) image. The edge effect caused by habitat fragmentation was assessed by overlaying, on a geographic information system (GIS), the land use and land cover data with the spectral ratio. The degree of habitat fragmentation was analyzed by deriving: 1. mean patch area and perimeter; 2. patch number and density; 3. perimeter-area ratio, fractal dimension (D), and shape diversity index (SI); and 4. distance between patches and dispersion index (R). In addition, the following relationships were modeled: 1. distribution of natural vegetation patch sizes; 2. perimeter-area relationship and the number and area of natural vegetation patches; 3. edge effect caused by habitat fragmentation, the values of R indicated that savannah patches (R = 0.86) were aggregated while patches of natural vegetation as a whole (R = 1.02) were randomly dispersed in the landscape. There was a high frequency of small patches in the landscape whereas large patches were rare. In the perimeter-area relationship, there was no sign of scale distinction in the patch shapes, In the patch number-landscape area relationship, D, though apparently scale-dependent, tends to be constant as area increases. This phenomenon was correlated with the tendency to reach a constant density as the working scale was increased, on the edge effect analysis, the edge-center distance was properly estimated by a model in which the edge-center distance was considered a function of the to;al patch area and the SI. (C) 1997 Elsevier B.V. B.V.
landscape ecology, fragmentation of natural forest formations, edge effect, fractals, fractal dimension, fractal models, remote sensing, geographic information system, vegetation indices
Forest Ecology and Management. Amsterdam: Elsevier B.V., v. 98, n. 1, p. 35-47, 1997.