Assessing the accuracy of dynamic yield models (Growfor) in highly productive stands of Sitka spruce

Abstract

Dynamic yield models are a useful tool to allow forest managers to make informed decisions regarding the management of their forests. In particular, they allow for the prediction of stand development and timber production based on just the current state of the stand, with no information on previous management needed. Work began on the first set of dynamic yield models in Ireland in 2003 (Growfor). Sitka spruce (Picea sitchensis (Bong.) Carr.) was the obvious choice as it is the dominant species in Irish forestry, currently comprising c. 50% of the total forested area. Due to recent developments in the forest estate, with afforestation of better land by private landowners now dominating, these dynamic models, especially for conifers such as Sitka spruce, have been found to inadequately cover the full range of yield classes now observed in Irish forestry.For the first time, Growfor’s performance in highly productive stands is assessed. Detailed measurements taken from felled sample trees enabled reconstruction methodologies to be employed, recreating stand attributes three years previously as well as obtaining detailed information on thecurrent state. This allowed investigation of Growfor’s ability to model these stands’ growth over this period. This study has shown that Growfor struggled to accurately predict the attributes of top height and basal area over the three-year period. It is recommended that Growfor’s developmental dataset for Sitka spruce be updated to reflect the ever-increasing proportion of stands which now find themselves classified as highly productive.