Quantification of repeated gap formation events and their spatial patterns in three types of old-growth forests: Analysis of long-term canopy dynamics using aerial photographs and digital surface models

Abstract

Treefall gaps are important elements of forest dynamics, as they contribute substantially to regeneration and forest turnover rates. Repeated gap formation, i.e. the formation of new gaps in areas where gaps have previously occurred then closed within decadal timeframes, is also believed to occur frequently and to contribute significantly. However, most gap studies have focused on the relationships among gap size, associated light levels and regeneration processes. In contrast, little reliable, quantitative information on the frequency of repeated gap formation events and their spatial patterns is available. To quantify repeated gap formation events in old-growth forests, we re-analyzed our previously published data on long-term canopy dynamics in three forest types (warm-temperate, cool-temperate and subalpine coniferous forests) in Japan, using digital surface models developed from aerial photographs taken in different years. Estimated repeated gap formation rates were 2.6%/32years (0.081%/year), 3.8%/43years (0.088%/year) and 8.3%/43years (0.193%/year) in the warm-temperate, cool-temperate and subalpine forest, respectively. The size distributions of repeated gap areas were skewed towards small areas, did not differ among the three forest types and closely fitted a power-law function. The mean repeated gap area also did not differ among forest types. The repeated gap areas were generally distributed along the edges of old gaps in the warm-temperate and cool-temperate broadleaved forests, according to a torus-translation test. Within the old gaps, continuous and/or rapid height growth was detected in sites with repeated gap events, and dramatic changes in canopy height were observed in sites adjacent to the edges of old gaps, which could have been due to the lateral expansion of branches and their subsequent destruction. These analyses revealed that the temporal and spatial canopy dynamics of the studied forest communities were very diverse and more dynamic than previously reported. Hence, future analyses of data obtained from Large-scale and Long-term Monitoring sites (LLMs) established in various types of forest communities around the world should pay attention to repeated gap formation events and their spatial patterns.