Uni and bivariate modelling of the spatial variability of Pinus taeda L.

Abstract

The objective of this study was to assess spatial variability of Pinus taeda L. growth due to soil clay contents a function of the soil obtained from a detailed mapping, in the scale 1:10.000, in an area located in the municipality of Rio Negrinho, Santa Catarina State, Brazil. Data on forestry growth was collected from plots of continuous forest inventory, obtained from a population growing on an area of 2.252,11 ha. Soil properties and Pinus taeda crop yields were analyzed through geoestatistical methods using semivariograms and interpolation by kriging and cokriging in order to obtain maps describing the attribute of main interest over the study region. Semivariograms were used to evaluated the spatial structure of the dendrometric property annual average increment (IMA) which is of primary interest and has shown stronger correlation with soil attributes. The comparison of the maps obtained by the uni and bivariate models suggests the latter provides a more detailed description of the main variable over the area, being the resolution dependent on the availability of data obtained geographically for the study.