Norway spruce bark thickness models based on log midspan diameter for use in mechanized forest harvesting in Czechia

Abstract

ABSTRACTForesters need to estimate bark thickness accurately in order to know the merchantable timber volume. Therefore, we developed a precise model for double bark thickness estimation that can be used with harvesters. We fitted a line to data from the Czech standard for bark deductions of Norway spruce (Picea abies (L.) Karst.) based on over-bark midspan diameters. Our model provided smaller deviations from the currently used polynomial model than other methods of bark thickness estimation. Only the diameter class method provided similar results. The mean absolute error (MAE) of the double bark thickness was 0.37 mm for our model or 0.48 mm for the diameter class method. The root means square error (RMSE) in this dataset was 0.43 mm or 0.56 mm respectively. We tested the deviations from the polynomial model on data from 22,442 spruce logs, gathered during forest harvesting in Czechia. Our model provided MAE of 0.34 mm and RMSE of 0.40 mm. Diameter class method provided MAE of 0.38 mm and RMSE of 0.43 mm. We concluded that our model is suitable for practical use in the Czech Republic. Its implementation will allow precise, automated log scaling, reducing at the same time the need for manual scaling.