Potential wood stock in brazilian savanna using geotechnologies

Abstract

Monitoring natural resources of our planet is essential to gather information to support strategies for conservation and a sustainable use of these natural resources. This monitoring is even more important in endangered biomes, as is the case of brazilian savannas, one of the most threatened and richest, in terms of biodiversity, savannas in the world. Usually, forest monitoring is accomplish by measuring the trees on the field, the so-called forest inventories, which are expensive, costly and extremely difficult to be performed periodically in extensive forests, as is the case of the Brazilian savannas. Therefore, this study aimed to update the information about volume and carbon stocks in savanna fragments in order to reduce tree measurements in the field. We used data from 61 savanna remnants, where 25 of them had 2 measurements with 5 years interval (monitoring). Multiple linear models, based on field data and reflectance values of Landsat images, were fitted and validated. Subsequently, the best models were applied to remnants that had only one measured, and then volume and carbon estimates were obtained for all 61 remnants in the second year of measurement. Additionally, maps updating the productivity of these remnants were generated.

Keywords:
Remote sensing; Forest inventory; Geostatistics