2018 Volume 44 Issue 2 Pages 340-351
Buried seeds in forest soils have attracted attention as a natural greening material, making predictions of their number and spatial distribution important. We aim to assess the degree of variation in forest seed bank characteristics based on a nested sampling approach. The study was conducted in a secondary forest that consists of four patch types (conifer, canopy gap, deciduous broadleaf tree, and evergreen broadleaf tree) in a 100-m slope. Three 5×5m plots containing three 20×20cm quadrats were established under each patch type. We obtained soil samples from a depth of 0―5cm, and seedling emergence was monitored over 5 months. In total, 116 seeds germinated (6.4/L), and twenty-five species were identified, with a dominance of the shrub Eurya japonica and the conifer Chamaecyparis obtusa. Variations in seed bank density were small between patches but were large between plots within the same patch type, and the latter accounted for approximately half of the total variances. The seed bank density tended to decrease in high-light environments (Indirect Site Factor＝0.12―0.16) and increase in plots with high densities of acorn-producing overstory trees, although there were also substantial within-plot variations in seed bank density, and further studies are necessary to confirm the results. Our simple method of nested sampling and statistical analysis for evaluating spatial heterogeneity of forest seed bank and for analyzing correlations with vegetation and micro-environmental factors are useful for considering effective sampling strategies of forest topsoil for revegetation practices.