Volume prediction of young improved Sitka spruce trees in Great Britain through Bayesian model averaging

Abstract More and earlier thinning operations are expected in Sitka spruce planted forests in Great Britain as a result of an increased demand for biomass and faster growth driven by breeding. It is however unknown whether the current models, which were designed to predict volume in adult trees, can provide unbiased volume predictions for the young individuals that are likely to be harvested in future thinning operations. The primary objective of this study was to answer this question. To do this, we used retrospective data from a destructive experiment originally aimed at assessing timber properties to reconstruct the taper and volume of 12 improved Sitka spruce trees at different ages. These volumes were then compared against the predictions from the current methods, which were found to be from moderately to strongly biased. The second objective was to provide proof of concept that a combination of existing volume models and other theoretical volume models could yield less biased predictions. We successfully addressed this objective through the Bayesian model averaging approach. The method, albeit tested with limited data, proved to be a promising alternative until new volume models are released. Further data from other available destructive experiments can be used to refine our calibration.