Abstract

Much effort is being made around the world to classify the native forest so as to better know its composition, structure and distribution in order to exercise a better management over them. The Araucaria araucana forests have ecologic relevance because they are millenary and its distribution in Chile and Argentina is quite reduced, which justifies the execution of the present investigation. This study proposes as its objective to classify at fine and medium scales the Araucaria and Nothofagus ecosystems of the south of Chile. To do this, remote sensors were used, applying a biunivocal-combined classification and the tree method, obtaining a detailed description. Auxiliary information used was: cartography of the Conguillio National Park, Digital Elevation Model and Vegetation Indexes. Also, Lansat-8 OLI Images from 2013, with spatial resolution of 30x30m, were used. Two hundred and sixty forestry inventory plots were carried out and the data were analyzed with the Idrisi Kilimanjaro, ArcGis 8.0 and SPSS programmes. The categories obtained from this supervised classification were 23, which had 75.0% of global accuracy. The accuracy for forests was 74.6%, for grass and bushes 80.5%, for soil with scarce vegetation 84.95%, for snow 95,1%, for volcanic lava 73.3%, and for water was 73.0%. Thus, the Araucaria and Nothofagus ecosystems were classified using this joined application of biunivoca-combined method and the tree method. This was done at a fine and medium scales. The ,final map allows to carry out a much better management of these ecosystems.