Tactical forestry planning. Common sense assisted by computers, and models linked to the strategic plan

摘要

This thesis is based on seven papers. Paper 1 presents a method to predict the present net value of timber production based on satellite data and subjectively inventoried forest stand data. In paper 2 it was tested whether satellite data can improve the estimation of cutting priority functions. The results indicate that satellite data are useful, as a complement to low quality stand data. However, for the stand database used in these studies, satellite data did not significantly improve the precision of the estimations. In paper 3, a forest-road valuation model and an optimization algorithm are presented. The combination of stand data and Landsat TM data was useful as value predictors in the road valuation. The combination of human knowledge and the computer calculation capacity was crucial in the road planning concept presented. The dynamics of a forest landscape is forecasted in paper 4. Static nonlinear regression functions were used to predict the mean digital numbers in a Landsat TM image at stand level. A future forest landscape was then predicted and presented as color composite images. In paper 5 digital high-altitude aerial photos were tested as an objective data source. Regression analysis was applied to estimate volumes and harvest priorities. When estimations were aggregated to stand-level, the precision was comparable with customary field survey methods. Paper 6 proposes a method whereby forestry planning parameters are objectively estimated in a spatially continuous manner without being limited by stand boundaries. The economic loss resulting from increased harvest ages was studied in paper 7. A function was presented that predicts the inoptimality loss as a function of stand data and extension time from optimal harvest time. Only the problem of economic loss, not the utility gain, from extending the rotation age was analyzed. 79 refs, 3 figs, 1 tab