The importance of whole-tree harvesting of Sitka spruce (Picea sitchensis (Bong.) Carr.) in Great Britain is likely to increase rapidly in the next 30 years. Large areas of marginal agricultural land were planted with Sitka spruce from 1960 to 1980 and this is likely to be harvested on a 50-60 year rotation. Increased mechanisation of harvesting operations has led to the frequent removal of whole trees from the site and there is concern over the implications of such forest management operations for longterm site fertility. This paper describes an integrated approach to be used in Britain to develop a decision support system (DSS) to provide long-term predictions of the impact of whole-tree harvesting on site fertility across a range of site types. The approach to be taken will make use of a wide range of data available from conventional yield tables, field and laboratory experiments which have examined various aspects of growth and nutrient cycling in Sitka spruce. This information will be used to calibrate a hybrid computer simulation model capable of simulating ecosystem dynamics. The parallel development of an ecological site classification system will provide a framework within which the model can be applied to generate a series of decision rules to be incorporated into a DSS. A key feature of the strategy employed is the establishment of a series of long-term field experiments which will test model predictions across the spectrum of sites to which the DSS will be applicable. Provision is also made to ensure that Feedback will be obtained to compare research output with operational performance. Finally, an approach is suggested by which the DSS could be integrated within a geographic information system (GIS) to provide a spatial context For the output which will be relevant to the wider forest industry, planning authorities and policy makers.