PREDICTING THE GENERAL YIELD CLASS OF SITKA SPRUCE ON BETTER QUALITY LAND IN SCOTLAND

The ability to predict accurately the potential site productivity of unplanted land is extremely important for forestry investment appraisal, production forecasting and land use planning at regional and national levels. This paper describes a model for predicting the General Yield Class (GYC) of Sitka spruce (Picea sitchensis (Bong.) Carr.), the most important commercial forestry species, on better quality land in Scotland. Using principal component analysis and multiple step-wise regression techniques 36.8 per cent of the total variation in GYC was explained by variation in 10 site and crop variables. The F statistic from the analysis of variance was significant at the 1 per cent level. Site factors most highly correlated with GYC were those related to climatic exposure (elevation and topex), soil moisture status (site drainage class and soil drainage class), crop age and soil type. Mean estimated deviation of predicted GYC from actual GYC ranged from 1 m3 ha-1 yr-1 for average sites (that is sites where the dependant variables were close to their average values) to 2.7 m3 ha-1 yr-1 for extreme sites. The predicted GYC for 15 independent sample sites was 18.5 m-3 ha-1 yr-1 compared to a true mean of 18.9 m3 ha-1 yr-1. This compares favourably with predicted means of 22.1 and 13.6 m3 ha-1 yr-1 from two earlier models.