The Weighted Least Squares Method Affects the Economic Rotation Age of Loblolly Pine - Two Planting Density Scenarios

ABSTRACT

Managers are typically interested in using equations that provide accurate volumes for large, merchantable trees while modelers are often more interested in equations that meet statistical assumptions. Therefore, modelers often use weighted (WLS) rather than ordinary least squares (OLS) to fit individual tree volume equations since the random error variance is not constant. Since OLS and WLS produce different parameter estimates, perhaps the method chosen will impact the predicted optimum economic rotation age. To address this concern, individual tree volumes were obtained from plots established in an unthinned loblolly pine plantation in southeastern Arkansas. Parameters of the combinedvariable function were then estimated using OLS and WLS for the same dataset. Stand-level projections and predicted diameter distributions for two planting densities (748 and 2,692 seedlings per hectare) were then obtained from a growth and yield model (PTAEDA 3.1) and individual tree volumes were predicted using the two parameter estimation approaches. When conducting economic analyses, we found the optimum rotation age between OLS and WLS can differ by as much as 11 yrs (i.e. 10 yr clearcut vs 21 yr clearcut).