Site quality of pedunculate oak (Quercus robur L.) stands in Galicia (northwest Spain)

Abstract

Accurate estimates of forest productivity are required for sustainable forest management. Height of dominant and codominant trees at a reference age is often used as a measure of site productivity. Eight algebraic difference equations based on the models proposed by Sloboda, Bertalanffy-Richards, Korf, Hossfeld, and McDill-Amateis were tested on the transformed, longitudinal data structure that considers all possible growth intervals. Autocorrelation was modelled by expanding the error term as an autoregressive process according to this data structure. Generalized nonlinear least squares methods were used for model fitting. Several numerical and graphic analyses were used to compare the different candidate models. The relative error in dominant height prediction was used to select 60 years as the best reference age. An algebraic difference equation based on the model proposed by Korf provided the best compromise between biological and statistical aspects and produced the most adequate site index curves. This model is therefore recommended for height growth prediction and site classification of pedunculate oak stands in Galicia. The model is polymorphic and base-age invariant, having only one asymptote. Predictions of height for age intervals between t1 and t2 of more than 15 years should be considered with caution because of the associated critical error.