With progressing climate change, increasing weather extremes will endanger tree regeneration. Canopy openings provide light for tree establishment, but also reduce the microclimatic buffering effect of forests. Thus, disturbances can have both positive and negative impacts on tree regeneration. In 2015, three years before an extreme drought episode hit Central Europe, we established a manipulation experiment with a factorial block design in European beech (Fagus sylvatica L.)-dominated forests. At five sites located in southeastern Germany, we conducted three censuses of tree regeneration after implementing two different canopy disturbances (aggregated and distributed canopy openings), and four deadwood treatments (retaining downed, standing, downed + standing deadwood and removing all deadwood), as well as in one untreated control plot. In addition, we measured understory light levels and recorded local air temperature and humidity over five years. We (i) tested the effects of experimental disturbance and deadwood treatments on regeneration and (ii) identified the drivers of regeneration density as well as seedling species and structural diversity. Regeneration density increased over time. Aggregated canopy openings supported species and structural diversity, but reduced regeneration density. Tree regeneration was positively associated with understory light levels, while maximum vapor pressure deficit influenced tree regeneration negatively. Deadwood and browsing impacts on regeneration varied and were inconclusive. Our study indicates that despite the drought episode regeneration in beech-dominated forests persisted under moderately disturbed canopies. However, the positive effect of increased light availability on tree regeneration might have been offset by harsher microclimate after canopies have been disturbed.