Peeking at ecosystem stability: making use of a natural disturbance experiment to analyze resistance and resilience.

To determine which factors contribute most to the stability of species composition in a beech forest after profound disturbance, we made use of a natural experiment caused by a severe windthrow that occurred at a permanent monitoring site in an old beech forest in Lower Saxony (Germany). The floristic composition was recorded for the succeeding five years after the disturbance and used to derive measures of resistance and resilience for plots as well as for individual species. Due to the existence of previously established randomly distributed permanent plots, we had precise information of the pre-disturbance state, including initial cover of the herb layer, species richness, and species composition. Variables describing the floristic change, resistance, and resilience were derived from correspondence analysis allowing for partitioning the effects of variation among plots from those of temporal change. We asked to which degree these variables could be predicted by pre-disturbance state and disturbance intensity. We found that both the pre-disturbance state and the disturbance intensity were good predictors for floristic change and resistance, while they failed to predict resilience. Among the descriptors of the pre-disturbance state the initial cover of the herb layer turned out to be a useful predictor, which is explained by a high vegetation cover buffering against losses and preventing establishment of newcomers. In contrast, species number neither showed a relationship to floristic change nor to resistance. Putative positive effects of species number on stability according to the insurance hypothesis might have been counterbalanced by a disruption of niche complementarity in species-rich communities. Among the descriptors of disturbance intensity, the loss in canopy cover and the change in photosynthetically active radiation after the storm were equally good predictors for the change in floristic composition and resistance. The analysis of the responses of single species on initial vegetation cover and disturbance intensity revealed that a wide range of different mechanisms were involved. Resistance and resilience did not depend on the presence of particular species or on specific traits.