Mediterranean beech forests: Thinning and ground-based skidding are found to alter microarthropod biodiversity with no effect on litter decomposition rate

Despite the high ecological value of Mediterranean beech forests, very little is known about the implications of forest operations on soil microarthropod biodiversity and litter decomposition rate. There is also no information concerning the amount of time needed for disturbed forest soil to recover and return to the pre-harvesting conditions. Silvicultural treatments are scheduled about every ten to fifteen years, without taking into account the amount of time necessary for recovery. The purpose of this study was to determine this information. The study started by selecting three study sites located along the Italian Apennine, each including a chronosequence of three forest parcels: one harvested in 2021, one harvested in 2012, and a control parcel which had not been harvested within the last forty years. In the harvested parcels we investigated skid trails which are classified as disturbed, and soil not affected by the passage of a machine which is classified as undisturbed. Thus in each study area there were five experimental treatments including the control area. The soil physico-chemical properties were assessed for each treatment. These included bulk density, penetration resistance, shear resistance, organic matter content, and soil microarthropod biodiversity which was assessed by the QBS-ar index (Soil Biological Quality based on microarthropods, a qualitative index measuring the quality of soil according to the biodiversity of the microarthropod community). We further established a litter decomposition experiment, using teabags as reference material, to check the differences among treatments in litter decomposition rates. We used linear mixed-effects models to investigate the effects of the experimental treatment on the physico-chemical and biological features, and relationships among QBS-ar index and soil physico-chemical features. We analysed the effects of the experimental treatment on the litter decomposition rate using generalised linear-mixed effects models. A significant effect caused by machine passage was found in the recently harvested parcels on soil penetration resistance (∼+70 % comparing to the control), shear resistance (∼+35 %), and QBS-ar index (-25 %). Effects of machine passage on soil bulk density and organic matter content were not significant. All the investigated variables in the skid trails returned to values similar to those recorded in the control after ten years. We did not find effects of any soil physico-chemical features on QBS-ar index. Finally, no effect of the experimental treatment was found on the litter decomposition rate, suggesting that a stronger magnitude of soil disturbance is needed to affect this complex biogeochemical process. Although the initial impact was highly significant, it was found that a 10-year period is sufficient for recovering the investigated soil features in the upper soil layer. This was observed in Mediterranean beech forests when forest operations were done with low-weight machines and a limited number of machine passes on the skid trails. Thus, it can be concluded that planning another operation 10–15 years after the previous thinning, as is currently done, remains in line with the time needed for the forest soil to recover.