MODELS TO PREDICT THE GENERAL YIELD CLASS OF DOUGLAS-FIR, JAPANESE LARCH AND SCOTS PINE ON BETTER QUALITY LAND IN SCOTLAND

Recent changes in forestry incentives mean that there is potential for an increase in afforestation on better quality agricultural land in S cotland. As a result improved information is required about timber yie lds from a range of species on better quality sites for production for ecasting, financial appraisal, and planning at the local, regional and strategic levels. This paper describes the development of models that enable General Yield Class (GYC) to be predicted from site factors fo r Douglas fir, Japanese larch and Scots pine. Temporary sample plots w ere established in stands below 350 m on Land Capability for Forestry class I to V sites. At each location GYC was assessed, as well as the soil, climate and topographic factors which had been demonstrated to i nfluence forest productivity in earlier studies in Scotland. The model s, based on a step-wise multiple regression procedure, indicate that m ean spring temperature, geomorphic shelter (topex), and crop age are m ost important in determining the productivity of Douglas fir and Japan ese larch. For Scots pine, mean spring temperature, mean winter temper ature, and crop age are the most important factors. The models account ed for between 34 and 45 per cent of variation in General Yield Class and are sufficiently precise for estimating mean productivity at regio nal and national levels.