During timber harvest trees of the residual stand may suffer damages through bark lesions. This may happen associated either with felling or transportation processes. Although qualitative information on the causal factors is available, quantitative interdependences between the major factors and the resulting degree of damage are largely unknown so far. To address this gap, two different quantitative models on the occurrence probability of bark damages were developed exploiting a substantial database of 183 harvest operations through GLMM approaches (generalized linear mixed effects models). The answering variables in the two models were either the occurrence of bark damages at the lower stem section (below 1.3 m height) or of bark damages along the whole stem. Among the multitude of potential causal predictors, the following major factors were found to impact on the occurrence of bark damages (sorted in descending order): mean skidding distance, distance of trees to skid trails, intensity of removals, tree species, tree resp. stand height, harvesting system (table 2, table 3). The models were subjected to a rigorous external validation based on the independent dataset that had been gathered previously by Meng (1978a). The validation clearly corroborated the models' capability to predict the occurrence probability of bark damages during harvesting operations (table 4, table 5 and table 6).