• The high tree mortality during the dry and hot years 2018-2019 in Europe affected not only even-aged and monospecific conifer stands but also semi-natural Fagus sylvatica L. (European beech) forests. This has triggered concerns on the future of European beech forests under climate change and raised questions as to whether forest management may increase tree mortality rates through creating more open canopies. We compared long-term mortality rates of beech between strict forest reserves and adjacent managed reference stands for three inventory periods at 11 sites in the federal state Hesse, Germany, near the centre of the European distribution of the species.
• We hypothesized that mortality would increase with climatic water deficits during the growing season, initial stand density, and decreasing dominance (canopy status) of trees. We also hypothesized that mortality would decrease with site moisture status and the intensity of tree removals. To quantify the influence of management, initial stand density and tree removals were used as predictor variables. In addition, we analyzed the influence of the climatic water balance and competitive status of trees on mortality.
• Mean annual natural mortality rates ranged between 0.5 % and 2.1 %. Even in the drought years beech mortality rates were surprisingly low. We observed no signs of striking canopy disintegration. The significantly higher mortality (1.6 % - 2.1 %) in unmanaged stands during the drought years 2018 and 2019 was largely confined to suppressed trees. There was no significant increase of mortality in managed stands during the drought years, but a shift of mortality towards larger canopy trees.
• Our study did not confirm a general influence of management in the form of tree removals on mortality rates. Yet, we showed that during drought years management changed the distribution of mortality within the tree community. To analyze the effect of management on mortality rates more comprehensively, a wider gradient in site moisture conditions including sites drier than in this study and longer post-drought periods should be employed.

Data of the long-term development of tree populations were collected on circular sample plots spread systematically over the area of unmanged and regularly managed beech forests. Tree individuals were re-identified and inter alia diameter at breast height and tree status (live, dead) were registered over app. 30 years in three censuses.    

Mortality rates for beech were calculated and analyzed for three periods of observation. The last period spanned the drought yers 2018 and 2019. 

A README file was uploaded to explain the dataset.