Abstract

The differentiation between various surficial deposits and bedrock on Anticosti Island is difficult because of the dense and homogeneous forest cover and because of the subdued topography. Remote sensing allows us to solve this problem by making use of the physical characteristics of Quaternary deposits and the weathered bedrock, which influence internal drainage and the availability of soil moisture to the vegetation. A spectral simulation of LANDSAT-4 was made using an airborne Daedalus 1260, 11-channel scanner. Several supervised classifications of the digital images were made using test sites studied in the field. Using the raw data from Thematic Mapper bands TM2, TM3, TM4, and TM7, the geologic environments and the ecodynamic units could be distinguished with 70% accuracy. However, the integration of bands TM2 and TM4 with the vegetation index (VI) = [(TM4 – TM3)/(TM4 + TM3)] and the algorithme (A) = [(TM7 − VI)/(TM7 + VI)] resulted in a classification accuracy of 80%. Band TM7 (2,08–2,35 μm) distinguishes itself from the other bands by having a strong reflection over bare bedrock and an absorption by water, which allow the characterization of modern alluvial deposits. The characteristics of TM7 can also be distinguished from those of the near-infrared wavelengths of TM4, which are absorbed by forest vegetation.

You do not currently have access to this article.