The true response of Fagus sylvatica L. to disturbances: A basis for the empirical inference of release criteria for temperate forests

• Dendrochronology and photogrammetry provide different views on disturbance history.
• Canopy position has a large effect on growth response magnitude.
• A high proportion of false releases implies bias in spatial analysis.
• New release thresholds are suggested for weak and major disturbance detection.

Because of some arbitrary decisions in recent dendrochronological research on forest disturbance histories, describing the actual growth responses of trees to disturbance events is still an issue of great importance. This is even more important in temperate beech-dominated old-growth forests driven by fine-scale disturbances, where a majority of growth pulses occur close to the arbitrary threshold of release detection. Recognizing this limitation, we provide a new empirically-based release detection criterion on the basis of actual reactions of trees to independently dated disturbance events interconnecting three data sets – tree censuses, dendrochronology and historical aerial photographs. The growth response of Fagus sylvatica L. was studied in detail in relation to the canopy position of surviving trees as well as regional climate responses, using 280 increment cores extracted in 8 old-growth forests in the Czech Republic.Cluster and bootstrap analysis resulted in distinguishing 2 ecoregions with site-specific growth patterns and plasticity. A boundary line (BL) method was applied to the identification and quantification of release, providing a new curve for F. sylvatica regionally valid in the mountainous regions of Central Europe. During the period investigated (1970s–1990s) 7% of trees experienced no reaction, despite clear evidence of canopy disturbances in both aerial photographs and tree censuses. 25% of growth changes were attributed to regional-scale environmental factors, in particular climate and air pollution. The magnitude of the growth response varied significantly among canopy positions, with subcanopy trees released from suppression experiencing the most intense growth rates, and individuals under permanent suppression showing a negligible response. Considering these differences we empirically inferred new criteria, subdivided into two categories according to disturbance intensity. Pulses falling within the 25–63% of the boundary line are classified as weak releases while pulses exceeding 63% of the BL are identified as major. Still, our results imply relatively high uncertainty in the detection of disturbance using dendrochronology, which can be troubled particularly in spatial analysis of past disturbances. We believe that our empirically-derived criteria represent a substantial contribution to further dendroecology research, with broad applicability to a wide range of beech-dominated temperate forests.