No abstract

No abstract

Individual-tree-selection (ITS) is a continuous cover forestry method thataims to utilize the uneven-sized stand structure through optimized individualtree selections. ITS is an interesting management alternative to rotationforestry in terms of production of multiple benefits and forest ownerprofitability. With regard to Fennoscandia, ITS is currently only practiced to alimited extent, partly because of a scarcity in knowledge and skills among allstakeholders, but also due to the shortage of scientific guidelines for howoptimized tree selections are performed in practice. Thus, the aim of thethesis is to develop tools for optimizing practical ITS recommendations forNorway spruce.Tree position data for model fitting were collected from five trial sites insouthern Sweden, one of which was documented with terrestrial laser scanningto get crown shape data for knot size modelling. A method that optimizes thetarget diameter dependent on the discount rate and the degree of thelocal competition exposed to the subject tree, was developed. Technically, adefined adaptive control function was optimized with full system stochasticsimulation and with expected present value as the objective functionvariable. Since the quality of the optimization is dependent on the qualityof the models in the simulation, priority was given to the evaluation anddevelopment of the central models. Available individual tree growth modelswere tested and evaluated with tree data from long term uneven-sizedtrial plots. Furthermore, new individual-tree distance-dependent growthmodels were developed for uneven-sized Norway spruce. A modellingframework, for prediction of the largest knot size per stem height section,was developed to provide means of simulating log quality depending on theimpact from local tree competitors.The results indicated that the thinning intensity in uneven-sized standstructures may be used to actively control the transition point where smallerdiameter classes are subjected to suppressed growth. The evaluation ofthe Swedish individual-tree growth models which are included in theHEUREKA forest simulator, showed underestimation of large trees andoverestimation of small trees. The underestimation is smaller or absent indenser forests for diameter classes above 40 cm. The new distance-dependentgrowth models are presented in two versions, one with and one without crownratio as a predictor. The new growth models facilitate the testing of spatialdecision criteria in ITS optimization. From the results of the ITSoptimization, field operators can choose either a fixed level target diameteror a flexible target diameter that is adaptive to the degree of localcompetition surrounding the subject tree. Optimal fixed target diametersat breast height range from 33 cm to 23 cm depending on discount rate(1–3 %). A decision support structure for ITS is proposed.