Divisão de Observação da Terra e Geoinformática

Essa página está em modo somente de leitura. Você pode visualizar a fonte, mas não alterá-la. Informe-se com o administrador do Wiki, caso você ache que isso está incorreto.
====== Abstract ======
Mapeamento de áreas susceptíveis a escorregamentos no município de Caraguatatuba (SP) usando técnicas de sensoriamento remoto

LANDSLIDE SUSCEPTIBILITY MAPPING IN CARAGUATATUBA (SP) MUNICIPALITY USING REMOTE SENSING AND GIS TECHNIQUES

E. V. Marcelino

The objective of the present research was the development of a quantitative approach
for landslide susceptibility mapping using remote sensing and GIS techniques. In order
to determine the best products for identifying landslide scars, six imagery fusion
techniques (Brovey, HSV, PCA, IHS and two Wavelets) were evaluated and many
geographical inference approaches were applied to obtain the landslide susceptibility
map. The study area was Caraguatatuba municipality, São Paulo, Brazil, due to
relatively recent catastrophic events of landslide that occurred there. ETM+ Landsat 7
(multispectral and PAN) and HRV Spot 4 (PAN) data were used for obtaining maps in
the 1:100,000 scale. IHS fusion was the technique that better preserved the
spatial/spectral quality and the photointerpretative potential from the studied imagery
and thus it was used for identifying landslide scars of the study area. By Maxver
classification and visual analysis a total of 497 landslide scars were mapped. From this
total, 397 scars were utilized for ponder the environmental variables and thematic
classes considered; the remaining 100 scars were kept apart for the validation phase of
the obtained susceptibility maps. Boolean, Index Overlay, AHP and Conditional
Probability were the used techniques of geographical inference. Boolean technique
provided the worst result due the excessive generalization of the favorable areas of
landsliding. Index Overlay and AHP provided intermediate similar results, mainly for
the class Very High susceptibility of landslides (r2 = 62%). Conditional Probability was
the better approach, providing a correlation of 95% for the Very High susceptibility
class. Because their high slopes and shallow soils (Inceptisols), Serra-do-Mar medium
hillsides were the more susceptible areas for the occurrence of landslides events.