Autoregressive spatial analysis and individual tree modeling as strategies for the management of Eremanthus erythropappus

Abstract

The objectives of this study were to apply statistical techniques to discriminate fertilization treatments of Eremanthus erythropappus (DC.) MacLeish. through autoregressive modeling, and to develop individual tree models for diameter and crown area (CA) projection to define management strategies for candeia plantations subjected to different fertilization treatments. This is an important tree species originating from the Brazilian Atlantic Rain forest and Savannah biomes, intensively used in the cosmetic industry. Nonetheless, to date, research has not addressed the management of natural stands or plantations of the species. Our experiment was located in Baependi, Minas Gerais, Brazil, and comprised of four randomized blocks and 13 treatments. The treatments consisted of 12 different regimes of fertilization plus a control. Each sample plot was composed of 50 plants plus two border plants in a planting spacing of 2.5 × 2.0 m and undergoing pruning at 5 and 6 years of age. Starting in the second year, total tree height (H) and circumference (at 1.30 m from the ground or breast height, CBH) were measured every 6 months. Starting in the fifth year CA was measured. Tree growth varied by fertilization strategy. Differences were detected by using an autoregressive approach, considering that standard statistical methods were not powerful enough to detect significant differences. Three growth groups were formed, and maximum growth was obtained for treatment 10 (NPK, 8-28-16). Management guidelines are provided based on individual tree models for different fertilization levels.