Modelling the variation of bark thickness within and between European silver fir (Abies alba Mill.) trees in southwest Germany

This study examined bark thickness variability of Silver fir (Abies alba Mill.) in southwest Germany over time and space by comparing a dataset of bark measurements from the 1970s with more recent assessments. Within-tree variability of bark thickness was analysed to estimate the required number of sampling locations per tree for defined accuracy levels. A range of models from the literature that predict bark thickness were compared for their predictive performance and Monte-Carlo simulations were used to estimate the effect of the number of sample trees and plots on the precision of the predictions. In addition, net log volume after bark subtraction was compared for logs of varying lengths to assess the influence of log assortments on calculated sales volume. Results show that several sampling locations are needed per tree and that at least five trees from at least 35 plots should be selected for measurements in the study region. For practical applications, diameter outside bark and breast height diameter are suggested as explanatory variables for models that predict double bark thickness. Additionally, relative tree height and age – and therefore growth rate – significantly improved predictions; however, environmental factors could not explain the variation between stands. Log lengths from 5 to 21 m only slightly influenced bark thickness equations that were fit on measurements at log midpoints. The findings highlight the need to consider bark thickness variability at different levels when developing bark thickness equations. In general, bark thickness was found to be smaller in more recent assessments and this indicates the need to regularly review existing bark equations for their validity.