AVALIAÇÃO DE ÍNDICES DE VEGETAÇÃO NA ESTIMATIVA DE VOLUME EM FLORESTAS DECIDUAIS: UM MODELO PREDITIVO.

Remote sensing makes it possible to understand the physical, chemical and biological characteristics of a forest system, indirectly, which has been a challenge in several studies. Correlating data from these sensors with field-acquired sample data, through forest formations collection and measurements, can be a highly effective strategy in comprehending forest structure and the entire fragment under study. A literature review on the subject enabled an analysis and discussion about methodologies used in academic studies, particularly vegetation indices extracted from satellite image data, as well as their correlation with primary forest inventory data, using the statistical model of linear regression. The practical development of the work took place in a forest fragment of a phytogeographic formation called Seasonal Deciduous Forest (Dry Forest) in Funilândia, Minas Gerais, Brasil, where 8 (eight) sampled plots were duly inventoried and subjected to statistical analyses in the office. In this research it was found a strong correlation between vegetation indices, especially the GNDVI (0,82), and the wood volumes of the samples collected in the field. This fact allowed the main objective of this research to be achieved, which was to extrapolate the volume of the entire forest fragment studied using the predictive model. [ABSTRACT FROM AUTHOR]