Estudo comparativo entre diferentes modelos digitais de elevação na extração da rede de drenagem em área submetida ao manejo florestal sustentável

Abstract

Currently, Digital Elevation Models (DEM) have become an indispensable tool for the planning of forest management activities, extraction of drainage networks, identification of trees in a permanent preservation areas, etc. Currently, free available DEMs are ALOS/PALSAR, ASTER/GDEM, SRTM and TOPODATA. In this context, the present work aims to compare a drainage network elaborated through fieldwork and GPS measurements with those extracted from the four above mentioned DEMs. First, the drainage vector was vectorized from the trees with a diameter at the breast height (DBH) ≥ 40 cm. In addition, drainage contribution area ranging from 10 to 150 was applied for each DEM. After, a comparison between each generated drainage network with the ground reference was compared each other using Kappa index and Z-test (Global Accuracy). Results showed that the drainage network extracted from the SRTM and ASTER with a value of contribution area 150 (13,56 km and 12,97) were the most similar one with those extracted obtained from the field network (9,65 km). Statistically, the DEMs showed an overall accuracy higher than 94%. Therefore, all models performed well. It is recommended to evaluate other environments as well as airborne LiDAR data.