Espectrorradiometria do visível e infravermelho próximo em povoamento de Eucalyptus grandis W. Hill ex Maiden

Abstract

Aiming to develop studies relating laboratory spectroradiometry with macro and micronutrients contents of vegetative materials of forest populations of Eucalyptus grandis W Hill ex Maiden, of different ages, an experimental área was selected for the study at the Fundação Estadual de Pesquisas Agropecuárias FEPAGRO FLORESTAS, at Santa Maria, Rio Grande do Sul, Brazil. Departing from the purpose of studying the spectral behaviour, the specific objectives were: 1) to evaluate the leaves of E. grandis contents of macro and micronutrients, in the four climatic seasons, under differente types of soil managements; 2) to evaluate the leaves of E. grandis spectral behaviour including the waveband of 475 to 980ηm and approach the interaction between: climatic seasonal periods, different soil managements and the cardinal points locations of the sampled trees; 3) evaluate the nutritional status of E. grandis young trees leaves through leaf analyses and spectroradiometry, and generate models capable to estimate macro and micronutrients from reflectancy data. Results indicate that the E. grandis population, of 44 months old, shows deficiencies of P, S, B; Cu and Mn showed high contents; N was the only nutrient showing significant differences; there is no significant differences between climatic seasons. For the E. grandis population of 20 months old, variations observed in the nutrients contents were equal to those of the former area; the winter sesason showed a differente behaviour from the other climatic seasons. For the different soil managements, the macro and micronutrients did not vary significantly. Evaluating the reflectancy data for the leaves of E. grandis, of 20 months old, it was concluded that: 1) the summer and spring seasons statistically interfered on the reflectancy; 2) samples of vegetative material may be collected independently of the type of soil management; and, 3) the spectral wavebands that better expresses the vegetation behaviour were the 12, 8, 7 and 3; and, 4) the evaluation of points where the leaves were sampled (East, West, North, South and central) in the no-tillage soil management treatment indicated that: a) sampling of vegetative material should be done separating the central part of the bordering population; and, b) the waveband that better characterized the spectral response of the vegetation was that of number eight. Analysis of the main components indicated the 8, 11 na 12 wavebands as responsible for a total of 98.10% of the reflectancy values under study. The elements P and K showed low adjusted correlation coefficients while generating models considering the elements as dependent variables from reflectancy and climatic seaso;n interacting with wavelengths; the r2aj. values for the remaining elements varied from moderate to high.