Error-in-variable method to estimate parameters for reciprocal base-age invariant site index models

Difference equations are widely used in forest productivity studies, especially in the construction of base-age invariant site index curves. A major problem associated with these equations is that several structures derived from original stem analysis data have to be considered for parameter estimation, and the compatibility of the estimated parameters using the ordinary least squares method is not guaranteed. The error-in-variable method is a parameter estimation method that ensures the compatibility of estimated parameters for the reciprocal base-age invariant models. In this study, we employed this method to estimate parameters for the base-age invariant models, based on several data structures consisting of two heights at two separate ages, and this method and the ordinary least squares method were compared. The error-in-variable method rendered the parameters compatibility without the need to specify dependent and independent variables. It also reduced the number of data structures such that the construction of the site index curves could be simplified.