Modelling bark thickness of Picea abies with taper curves

Abstract

Models predicting bark thickness at 14 different relative stem heights were derived, and factors affecting bark thickness were analyzed. The study material was collected from the tracts of the fifth National Forest Inventory (NFI) during the years 1968–1971. It consists of sample tree data measured from 1864 Norway spruce ( Picea abies) stems. Bark thickness was measured with bark gauges at the following 14 different relative heights: 1, 2.5, 5, 7.5, 10, 15, 20, 30, 40, 50, 60, 70, 80, and 90%. Linear regression models predicting bark thickness at each relative height were derived. These models were combined into an equation group in which double bark thickness at each relative height was predicted with both endogenous variables (bark thickness at neighboring relative heights) and exogenous variables (tree age, tree height, and breast height diameter). This equation group was solved with the Thomas algorithm. Predicted bark thickness values were finally joined with spline functions to form a bark curve model. The models gave unbiased estimates for both bark thickness and bark volume. The relative error for bark thickness varied from 13.7 to 27.9% and the standard error for stem volume under bark varied from 5.27 to 7.79 dm 3 depending on the amount of input data. Bark thickness was found to correlate with diameter at breast height, tree height, tree age and stem tapering. Climatic zone also clearly affected the amount of bark. By combining the models derived in this study with existing stem curve models, it is possible to accurately calculate stem volume both under and over bark for any arbitrary portion of the stem.