Permanent plot selection in moutainous ecosystems / Seleção de parcelas permanentes em ecossistemas montanhosos

AUTOR(ES)
DATA DE PUBLICAÇÃO

2003

RESUMO

A methodology for the selection of an optimal set of drainage basin with regional representativeness of the environmental variability of the studied area is presented. The methodology is based on the cluster analysis of the second order drainage basins parameterized by environmental characteristics related to the spatial pattern of the land cover, to microclimatic properties indicated by the irradiance regime and to the morphology of the basins. The study area is in the Serra da Mantiqueira, Southeastern Brazil, with an area of 5.600 km2, bounded by the parallels 22o15 S and 22o45 S and the meridians 44o30 W and 45o30´ W, which presents an altimetric amplitude of 2.477m (min=305m, max=2782m). A Digital Elevation Model (DEM) was constructed through linear and geostatiscal interpolations of altitude samples digitized from 1:50,000 topographic maps with contour lines at every 20m. The study site was stratified into drainage basins with the program TauDEM. Parameters related to the land cover were attributed to each basin with the use of zonal operators applied to an existing land cover map taking the basin boundaries as reference for zone definition. By the same way it was attributed other parameters. Those related to the basins morphologies were extracted from the topography (DEM) and drainage network grid data. Parameters related to the microclimatic conditions, or the clear sky irradiance, was simulated with the program TOPORAD that requires, besides the DEM, a characterization of the optical properties of the atmosphere and a grid data representing the surface albedo. Optical properties of the atmosphere were obtained with the use of the 6S program and from a global aerosol database. The albedo grid was derived from Landsat data. A total of 24 variables were attributed to each basin, 6 related to the land cover spatial distribution, 4 related to the basin morphology and 14 related to the irradiance regime. The total set of 5368 basins was reduced to 605 basins by the deletion of inadequate basins as those with less than 10ha of forest cover. The analysis of the collinearity among the variables allowed the selection of five uncorrelated variables. Agglomerative and divisive approaches for the cluster analysis were applied to the complete and to the selected environmental variables sets and different classifications were obtained. For each classification result the analyzed set was stratified into classes based on the adoption of different similarity thresholds. The drainage basin closest to the center of gravity of the cluster in the attribute space was selected to represent each class. The representativeness of the selected basins in relation to the regional environmental variability was investigated by the analysis of the distribution of the values of the environmental variables applied of each basin in the whole dynamic range of each variable. A set of twelve basins obtained by the divisive cluster analysis applied to the five uncorrelated variables and stratified with similarity threshold of 90 was chosen as the most parsimonious and representative of the regional environmental conditions. The presented methodology can be improved by the introduction of new environmental variables. It is expected that this methodology can be applied in ecological studies related to the analysis of the response of terrestrial ecosystems to global change or to the development of sustainable forest management techniques that require the location and allocation of a system of permanent plots with regional representativeness.

ASSUNTO(S)

ecossistema de montanha sensoriamento remoto irradiância solar permanent plots mountain ecosystem ecology ecologia serra da mantiqueira solar irradiance mantiqueira range parcelas permanentes remote sensing

ACESSO AO ARTIGO

Documentos Relacionados