The two storms Vivian (1990) and Lothar (1999) left an area of roughly 9000 ha of fully damaged protection forests in Swiss mountain regions. Given this huge dimension, questions arose on how to manage these areas to keep the protection gap, i.e. the time period with reduced overall protection against natural hazards, short. Quantifications are presented for the stability of lying logs left in place, the frequency of post-disturbance mass movements, and the tree regeneration in windthrow areas. The average height above ground of unsalvaged lying logs decreased from 2.1 m shortly after disturbance to 0.8 m 20 years later. In the period 1990–2014, the number of avalanches in windthrow areas was markedly small, and annual rates of shallow landslides and debris flows in windthrow areas did not differ from rates in comparable undamaged forested areas. Regeneration density rarely exceeded 4000 stems ha^–1 20 years post-windthrow at elevations above 1200–1500 m a.s.l. Mean height of tallest trees reached 5.6 m in areas that were cleared and 6.5 m in those left unsalvaged. Trees planted post-windthrow were 1.0–2.4 m taller than naturally regenerated ones. Practitioners rated the protective effect to be acceptable 24 years post-disturbance in only 5 out from 16 observed windthrow areas (31%), with planting trees as the main cause of success. We conclude that in protection forests the regeneration speed after disturbance rarely meets practitioner's expectations in terms of both stem density and stand structure. However, leaving woody debris from wind disturbance in place proved to replace protective effects for an astonishingly long time. An intensive management with salvage logging, planting and even technical constructions seems therefore only inevitable on windthrown areas where risks seem too high based on hazard, damage potential and possible spread of bark beetles to nearby protection forests. A management alternative applicable to many other cases of windthrown protection forests is to plant trees between lying stems.