Mind the gaps: Comparison of representative vs opportunistic assessment of tree regeneration in Central European beech forests

Tree regeneration (TR) in canopy gaps is a key process to understand how forest ecosystems might adapt to future environmental changes. Since successful TR is the result of a complex interplay of several stochastic events such as gap formation, seed production, ungulate pressure and diseases, some of the processes have been neglected or strongly reduced for model applications. Most empirical data on TR originate from (national) forest inventories based on statistically optimized designs for timber resource estimations and use geographically fixed permanent plots. Consequently, these representative samples record TR both in gaps as well as under closed canopy conditions. In this study, we compared TR in 63 representative plots of the Swiss National Forest Inventory (?NFI sampling?), located in beech and silver fir-beech forests along an elevation gradient in the Jura Mts., with an opportunistic ?gap sampling? strategy targeting TR-clusters. We analyzed quantitative (e.g., number of individuals) and qualitative (e.g., species diversity, similarity to surrounding canopy) differences between the ?NFI sampling? and ?gap sampling,? as well as the environmental factors (e.g., gap size, cluster age, elevation) influencing the composition of TR in small canopy gaps. The ?gap sampling? recorded higher median TR densities (+62%, stems ha- 1) and species richness (4.0 ? 2.1, mean ? sd) than the ?NFI sampling?. More importantly, the ?gap sampling? provided much more consistent results among sites (i.e., 4x lower variance). Consequently, the ?gap sampling? revealed not only general but also species-specific patterns across TR size classes and along elevational gradients in contrast to the lack of patterns in the ?NFI sampling?. Species richness decreased as TR grew taller, and the tallest TR size class (>130 cm) was almost solely comprised of the dominant species in the surrounding canopy (direct ingrowth). The high amount of direct ingrowth in TR indicates that small canopy gaps might offer only limited management potential to alter forests and that extensive disturbances or management at early TR stages (sapling) would be necessary to promote species better suited to the expected future site conditions. We raise the question of whether a forest monitoring that is statistically optimized for timber resource estimation is best suited to detect adaptation effects to quickly changing environments best visible at the regeneration stage, or whether forest inventory tools should be extended by methods that focus on disturbed parts of forests where changes and adaptations take place. We propose the implementation of an opportunistic gap-based monitoring of TR that essentially serves to identify the best forest management strategies (e.g. gap size, intervention intervals) in forests adapting to environmental change.