Classification and Forest Parameter Extraction of Patagonian Lenga Forests with Aster and Landsat Etm+ Data

The goal of this project was to improve the mapping and monitoring system of the Directorate of Forest and Parks of the Province of Chubut, Patagonia. In providing updated quantitative and qualitative knowledge of Patagonian vegetation, mainly native forests, sustainable forest resource management shall be improved. In this paper object-oriented classification using eCognition software is presented as well as the feasibility of modeling measured forest parameters of Lenga (Nothofagus pumilio) stands by relating them to vegetation indices derived from ASTER and Landsat ETM+ data. Forest parameters such as leaf area index (LAI), diameter at breast height (DBH), basal area, volume among others were measured and related to the satellite data by using simple and multiple linear as well as non-linear regression analysis. The findings of forest parameter determination showed that LAI can be estimated with ASTER data at a relative root mean square error (RRMSE) around 12%. The best determination results were achieved for the forest parameters DBH and basal area with relative RMSE around 26% and 30% respectively. Volume was estimated with lower accuracy. a relative RMSE of 45% was achieved.