Ten years of regeneration dynamics in an unexploited lime-hornbeam forest in the Bialowieza National Park (Poland): an assessment of the variability of the forest mosaic

Changes in tree species composition and the forest mosaic over the 1982û1992 period are described for a one-hectare plot in an unexploited lime-hornbeam forest in the Bia�owie¿a National Park (Poland). For trees 10 m the total numbers declined slightly, basal and crown projection area remained stable, and crown volume increased, resulting in a denser canopy. Picea abies had the highest annual mortality rate (3%), Carpinus betulus the lowest (0.2%). For trees with height ú 10 m, Carpinus had the lowest mortality rate and showed the highest increase in both crown projection area (133%) and crown volume (215%). Regeneration occurred irrespective of crown cover. The percentage of the area occupied by tree phases, based on the height of the upper canopy layer (1: ú 10 m; 2: 10û20 m; 3: 20û30 m; 4: 30 m) remained rather stable. The gap phase slightly increased. Phase 4 expanded most, at the cost of intermediate height phases 2 and 3, and had the highest probability (0.75) of remaining stable. The highest probability of a transition was for phase 3 to become phase 4 (0.34), reflecting the growth of the highest canopy layer. Developmental phases overlapped considerably. Pair-wise overlap (1 + 2, 1 + 3, etc.) amounted to 15û20% for all pairs of phases. Double (1 + 2 + 3, etc.) and triple overlaps added ca 15% for phase 1 with the other phases. Phase 1 was not confined to former gaps but showed a net pattern (regeneration network) earlier recognized by Koop (1989). Between 1982 and 1992 this pattern became more pronounced. It was concluded that tree regeneration outside gaps is an important and probably general phenomenon in old-growth forests, and that differential regeneration beneath declining or otherwise heterogeneous canopies may produce spatial patterns and vertical structures not easily explained by gap dynamics. An important consequence is that studies on the dynamics of forest mosaics must deal explicitly with overlapping developmental phases