Strategic forestry planning : evaluation of variable spatial aggregations and forest landscapes

This thesis deals with strategic forestry planning. The outcomes from strategic planning, such as potential harvest volumes, growth, net revenues from harvesting, and areas clearcut, should provide ready guidelines for planners and managers and form a base for tactical and operational planning. Methods estimating costs for regeneration, logging, and off-road extraction, and the distribution on assortments of felled volume based on data from Swedish National Forest inventory (NFI) sample plots in a system for long-term forecasts of timber yields and potential cut (the Hugin system) are developed and evaluated. As a base for a forest management programme in the Hugin system, results from a questionnaire study of the public preferences concerning the forest environment in a northern Swedish county were used. The results were translated into forest data useful in the Hugin system for selection of appropriate NFI sample plots. A forest scenario with high consideration given to peoples’ preferences regarding different forest attributes was compared with a scenario with 1990s levels of cutting and use of silviculture methods. The management programmes differed substantially, with much more single tree selection, natural regeneration and higher shares of broad-leaved trees in the first scenario. In this scenario, which considered the people’s preferences, calculations indicated an approximately 20% reduction in logged timber volumes, which corresponds to a 10% loss of profit. Generally, the basic unit in tactical and operational planning is predefined stands. With small homogeneous stands the opportunities to apply optimal management at the local level are better than with large heterogeneous stands, but the costs for inventories, planning, and harvesting single stands will be higher. The economic consequences of different stand delineation strategies were analysed on an area of 5900 ha having a continuous description of forest variables in raster elements estimated using field and satellite data combined by the k nearest neighbour (kNN) method. The results indicate that differences in net present values (NPVs) are small: 0.8 % and 2.0 % higher NPV for small stands than for medium and large stands (3.6, 7.8, and 21.2 ha), respectively. When including entry costs, that is, the cost of moving machines between treatment units, large stands have 2.1% and 3.2 % higher NPV than medium and small stands, respectively. In the final study, the formation of treatment units was included as a part of the planning process. Dynamic treatment units were formed and selected for harvest such that NPVs were maximised for a forest area of 558 ha in Northern Sweden. The forest variables were described in raster elements (30 m x 30 m) estimated by the kNN method. The degree of clustering was controlled by an entry cost for logging. Simulated annealing was used as solution technique. The results show that the net present values are at least as high as forestry with predefined stands even if entry costs are included. The NPV from felling decreases by about 4% when the degree of clustering is increased to about the same level as with predefined stands