Dados geomorfométricos como subsídio ao mapeamento da vegetação / Geomorphometric data as subsidy for vegetation mapping

AUTOR(ES)
DATA DE PUBLICAÇÃO

2007

RESUMO

The potentials of geomorphometric variables derived from SRTM (Shuttle Radar Topographic Mission) for the mapping of vegetation types were investigated. Four published vegetation maps for study areas inside three Brazilian biomes (Amazonian forest, Savanna and Steppe) served as reference for testing the discriminatory potential of the geomorphometric variables. The study sites differed in the feasibility of applying topographic data to vegetation mapping according to the variation of geomorphometric conditions and vegetation reference mapping scale and detail level. The analysis indicated the vegetation class groups prone to classification in contrast to groups occurring in similar topographic conditions. The variables that presented strongest relationships with the vegetation classes were elevation, horizontal and vertical curvature. Although geomorphometric data had shown potential for discriminating vegetation classes, the achieved classification could not reach the thematic detail level of the reference maps due to co-occurrence of classes with similar vegetational structures. Based on discriminant analysis, it was possible to experimentally map to the sub-phytophysiognomic level. The best results were found for the RPPNSA and PNCA sites, and the sub-physiognomies of the latter. PNB site showed the worst results at the phytophysiognomy level, while its tested sub-physiognomies performed satisfactorily. It was found that discriminant analysis of geomorphometric data could subsidy vegetation mapping.

ASSUNTO(S)

análises discriminantes modelo digital de elevação vegetation rremote sensing digital elevation model topography vegetação topografia sensoriamento remoto discriminant analysis

ACESSO AO ARTIGO

Documentos Relacionados