DETERMINAÇÃO DE VARIÁVEIS DENDROMÉTRICAS DE Eucalyptus urograndis COM DADOS LiDAR AEROTRANSPORTADO

Abstract

This study aimed to compare biophysical parameters of a stand of Eucalyptus urograndis obtained in conventional forest inventory with those derived from profiling the Airborne Laser, with densities of Laser points of 5 and 10 points / m². The forest plantations selected for this study are located in the Paraíba Valley, State of São Paulo, Brazil, belonging to the company Fibria Celulose S.A. The sampled plots during the forest inventory are temporary and circular, having a radius of 11.28 m and area of approximately 400 m². The area was overflown during the months of January to March, for the acquisition of laser scanning density of 5 and 10 points/m², respectively, and simultaneous acquisition of aerial photographs. In the analysis of biophysical parameters, the total height of the trees was obtained with LiDAR data directly by the determination of difference between the representative points on the ground and the highest point in the crown area selected manually, and the diameters and volume were estimated according to the height, with adjusted model from field data. The means of height, diameter and volume obtained in stands with variation in age between 3 and 8 years, derived from field information were compared with data extracted from LiDAR by the Dunnett test (α ≤ 0.05). The results indicate that the means height, diameter (DBH) and volume obtained by LiDAR with densities of 5 and 10 points/m2, generally, presented results statistically identical to the data obtained in field, except for the oldest stand, considered in this research (8 years). On the other hand, the mean errors obtained were among 3.97 to 12.77%; 4.54 to 20.89% and 14.46 to 59.26%, respectively for the height, DBH and volume. The LiDAR data with 10 points/m2 of density always presented lower mean errors than the density of 5 points/m2. The deviations between the data obtained by LiDAR against those of field presented a good correlation with age, indicating that this variable can be measured through this tool. However, these same deviations had no correlation with the slope, indicating that this variable seems did not affect the height measured with LiDAR data. In summary, there is possibility to apply this LIDAR technique on a large scale for forest inventory and monitoring, maximizing the accuracy and minimizing the collection of basic information from stands by the traditional method, still necessary to adjust the diameter and volume equations.