Gap effects on understory structure

In beech-dominated virgin forests, the understory from the sapling to the pole-sized stage has been characterized by a clumped spatial distribution in close association with the presence of canopy gaps (Janík et al., 2016; Nagel et al., 2006). In the Kyjov forest, beech trees in sub-canopy layers were recorded on the expanded gap area of all gaps and occurred with similar mean total cover (about 60%) in older (> 10 years) and new gaps of various sizes. In many

Canopy gap dynamics and tree understory release 2.4

gaps, more than one sub-canopy layer was present in the 1–19 m height range, indicating the coexistence of several beech cohorts in vertical layering that must have established in several periods with favorable light conditions. However, a characteristic feature is a large variation in understory cover and structure in the different gaps (Fig. 2.4). Canopy gaps in other beech-dominated virgin forests have been reported to release multiple sub-canopy layers as well (Bottero et al., 2011; Diaci et al., 2005; Diaci et al., 2012; Garbarino et al., 2012; Nagel et al., 2010; Rozenbergar et al., 2007), whereas for forest reserves with past management history, gaps are often described as having a single regeneration cohort of more or less the same height (e.g.

Diaci et al., 2005; Emborg, 1998; Kenderes et al., 2008). Thus, the sub-canopy structure seems to differ between virgin forests and stands with management history, even if the human impact relates to the distant past. Therefore, observations on gap-induced understory dynamics and dependent ecosystem processes (such as nutrient leaching and nitrogen mineralization; Bartsch, 2000) are applicable to other forest ecosystems only with great care.

The impressive rate of gap closure is partly a consequence of the rapid growth of pole-sized young beech trees, which can fill a gap within a few years. Indeed, we found cohorts of pole-sized young beeches (middle layer) in 60% of the gaps formed between 2003 and 2013 that must have established long before gap formation, offering an explanation for the rapid gap closure observed in larger old gaps for the period 2003–2013. On the long run, the released understory as a whole (from saplings to pole-sized trees) is of crucial importance, as it determines the potential for vertical gap closure. This is especially true for larger gaps, where the potential for lateral gap closure is limited, as is the arrival of new seedlings.

In a beech-fir virgin forest in Bosnia-Herzegovina, Nagel et al. (2010) found the regeneration structure in gaps not to differ from that in systematically placed research plots, which covered a wide range of canopy conditions, due to a high presence of advance regeneration across the whole forest. The authors suggested that gaps do primarily alter the development of advanced regeneration, rather than promoting the establishment of new offspring. Our results are in line with this suggestion, as total cover of sub-canopy layers was similar in old and newly formed gaps in Kyjov; yet the lower sub-canopy layers (< 2 and 2–9 m in height) differed in their cover in these two distinct gap types. Janík et al. (2016) found beech saplings under closed canopy conditions to grow up to a stem diameter of 4 cm, and then requiring canopy disturbance to advance to larger size classes. Our data suggest that gap formation had a large effect on sapling height growth in their early development stages, as expressed by the shift from small-sized regeneration (< 2 m) in the new gaps (recorded in 2013) to the dominance of larger saplings (2–9 m) in the older gaps (established before 2003). The positive relation between gap size and the cover of lower tree

Discussion 2.4

layer (2–9 m) in older gaps further documents that the growth conditions improve with increasing gap size also for shade-tolerant species such as beech. This corresponds well with the increase in light intensity with increasing gap size (Canham et al., 1990; Coates and Burton, 1997;

Vilhar et al., 2015) and it matches the observation that shifts in the cover of sub-canopy layers, as well as vertical gap filling, mainly happened in medium to large gaps (> 100 m²).

Our results only partly support our hypothesis H3, as the comparison of old and new gaps evidences a positive gap effect on the vertical growth of the lower sub-canopy layers (< 2 and 2–9 m). However, we obtained no clear signs for a promotion of seedling establishment through gap creation, since total understory cover was similar in old and new gaps. In a Slovenian virgin beech forest, Diaci et al. (2012) recorded in the first years after gap formation increasing sapling numbers (≤ 20 cm in height), indicating an important role of gap formation for the establishment of F. sylvatica seedlings. The sapling density decreased in the subsequent years and the population of young beeches shifted to a larger size class, in a similar manner as was observed in Kyjov. Thus, for studying the effect of gap formation on the establishment of beech regeneration, it seems necessary to count young tree individuals in the period immediately after gap formation.

Our results also highlight a methodological issue relevant for forest dynamics research, which is related to the definition of a gap. Part of the gap area mapped by us would not have been classified as a gap by approaches, which consider the canopy to be closed when tree height is 1/3 (e.g. Hobi et al., 2015b) or half (e.g. Nagel and Svoboda, 2008) the stand height. An even smaller gap fraction would be seen by the remote sensing approaches of Garbarino et al. (2012) and Hobi et al. (2015a) that rely on detecting more or less bare ground conditions.

All the mentioned approaches may miss part of the gaps in the upper canopy, which play an important role for the turnover of tree individuals and canopy strata in virgin forests.

A key issue in the search for a more generally accepted definition of canopy gaps is the question, from what height, relative to stand height, a tree should be assigned to the (main) canopy. This threshold is separating the dominant upper canopy layer from all subordinate strata including the understory layers. As there are species-specific differences in growth characteristics, it is necessary to develop specific definitions for different forest types.

2.5 Canopy gap dynamics and tree understory release