Testes de identidade de modelos aplicados na construção de curvas de índices de local e na produção de povoamentos de Eucalyptus grandis

Abstract

ResumoEste trabalho teve como objetivo aplicar testes de identidade de modelos na construção de curvas de índi-ces de local e na produção da área basal e do volume por hectare em povoamentos de Eucalyptus grandis W. Hill. Ex. Maiden para seis fazendas de uma região do estado de São Paulo. O teste de Regazzi (2003) foi aplicado para analisar a possibilidade de utilização de um único modelo de altura dominante para todas as fazendas. A curva média de produção em área basal e volume ajustada pelo modelo de Clutter (1963) foi projetada para o índice de local médio e aplicado o teste de identidade de modelos, por meio do teste de Chow (1960). Uma única equação não é suficiente para representar o crescimento da altura dominante para as seis fazendas. Ao aplicar o teste separadamente entre as fazendas observou-se que os dados de algumas delas podem ser agrupadas, sem perda de precisão. Com as projeções do modelo de Clutter (1963), o teste de Chow (1960) indicou que algumas fazendas podem ser agrupadas e outras não deveriam ser agrupadas em termos de produção de área basal e volume.Palavras-chave: crescimento, parcelas permanentes, comparação de curvas, manejo florestal. AbstractThis study aimed to apply identity model tests in the construction of site index and on the production of basal area and volume per hectare in stands of Eucalyptus grandis W. Hill. Ex Maiden growing in six farms in Western São Paulo State. The Ragazzi test (2003) was applied to analyze the possibility of using a single model for dominant height for all farms. The average production curve in basal area and volume was adjusted by the Clutter (1963) model and projected for the mean local index, and the model identity test was applied through the Chow test (1960). A single equation is not efficient to represent the dominant height growth for the six farms. When applying the test, separately between the farms, we observed that some of the data may be grouped without loss of accuracy. With the projections of the Clutter model (1963), the Chow test (1960) indicated that some farms can be grouped and others should not be grouped for volume and basal area production.

Identity Test of Models Generated in Neutron Probe Calibration in Latosol Cultivated With Coffee and Signalgrass

“…In Figure 2 (b), considering only layer 2 in relation to the two management systems (T and Sg), it was observed that the regression distances resulted in overlapping confidence intervals only at the ends of the equation where the experimental error is larger, indicating that there is a difference between the regressions in this soil layer. Roveda et al (2016) evaluated eucalyptus growth in six different locations, and observed that two models did not show statistical difference through the Model Identity method, indicating the possibility of using a single plant growth curve for these locations. For the others, as there was a statistical difference, it was concluded by using a specific model for each location.…”

“…Variations in exploitation form and depth of the root system of plants used in covers and their capacity of biomass production, in relation to the main crop, allow the recycling of nutrients, several models are observed. This procedure has been widely used in forestry engineering due to the large number of models generated for administrative planning and prediction of yield capacity and production (ROVEDA et al, 2016). Thus, this procedure could be used in the comparison of models that describe the behavior of the calibration of a neutron probe under different types of soil and / or handling.…”