Proxy variables to represent the productive capacity in a forest growth and yield model for eucalypt hybrid plantations in north-eastern Bahia, Brazil

Abstract

The present study aimed to evaluate proxy variables to represent the productive capacity of the sites in a forest growth and yield model for plantations of Eucalyptus urophylla × Eucalyptus grandis hybrid, considering them in a linear and a linear mixed model structure. Data from permanent plots in stands located in north-eastern Bahia (Brazil) were used, where there is a high spatial variability of rainfall and different soil types. The variables site index (SI), based on a height-age relationship; mean height of dominant trees (Hd); thirteen soil classes; and six annual rainfall classes (ranging from 850 to 1 200 mm) were used as alternatives to represent the productive capacity. The replacement of the site index (SI) by the mean height of the dominant trees in the Schumacher model resulted in an equation somewhat less accurate than that obtained by using site index (SI), but in the validation model's process provided more accurate estimates. The inclusion of the rainfall or soil classes as a random effect, in a mixed model, instead of the site index (SI), also resulted in somewhat less precise equations as well as less accurate estimates in the validation process. However, despite such results, all the evaluated variables were shown as alternatives to represent the productive capacity of the stands in a forest growth and yield model due to the accuracy of volume estimates in the fitting and validation process.