The expansion of bamboo forests has been assessed for various regions in, mainly, western Japan, but the actual conditions of the distribution of bamboo forests are not clear over a broad scale. We estimated the probability of bamboo forest distribution (PBD) throughout Japan using vegetation data derived from natural environmental information GIS and digital national land information etc. We also examined several environmental factors that may affect the distribution of bamboo forests. The accuracy of PBD was evaluated in several regions using detailed vegetation maps. PBD was estimated using a generalized liner model based on bamboo forest distribution data and environmental variables (warmth index, precipitation, maximum snow depth, surface geology, slope inclination, proportion of land-use types). As a result, a predictive model with high goodness of fit was constructed (AUC=0.815). To assess the applicability of the PBD map, we evaluated the accuracy of PBD predictions in Hiroshima Prefecture. The results indicated that PBD was strongly correlated with both the area of bamboo forest per 1000 ha (R² =0.872) in six regions and the annual bamboo forest expansion rate (R² =0.982) in four regions. Third meshes with PBD values greater than 0.48, which value was found as cutoff point from ROC analysis, were found primarily along coastal plains and in hilly zones of western Japan, in areas south of Miyagi Prefecture on the Pacific coast, and south of Niigata Prefecture along the coast of the Sea of Japan. The proportions of land-use types, including forest, farmland, and inhabited areas exhibited non-linear, convex relationships with PBD, suggesting that PBD tends to be high in the "Satoyama" area, which is characterized by a mosaic of these land-use types. Finally, we proposed an example of regional characteristics evaluating method for consideration of land monitoring and management, according to PBD and population change pattern.

View full abstract

In this report, we investigated the wood species of the woodland which is located in the suburbs of Shiroi city, Chiba prefecture, in order to understand the species composition of the woodland, and estimate the renewal of the forest in the future. The correlation between species and areas was positive and had significant differences in all investigation sections, and the correlation coefficient of deciduous trees was higher than that of evergreen trees. Comparing to the appearance rate of wood species in the sub tree - tree layer, shrub layer, both layers (the same species which were found in both layers of each investigation section), in deciduous trees, Carpinus tschonoskii, Quercus serrata, Quercus acutissima were the highest at the sub tree - tree layer, and Magnolia kobus, Celtis sinensis, Aphananthe aspera, Zelkova serrata were the highest at the shrub layer. In evergreen trees, Quercus myrsinaefolia was high at the sub tree - tree layer, shrub layer and common species, Neolitsea sericea, Ligustrum lucidum were the remarkably high at the shrub layer. We guessed that evergreen trees at the shrub layer grow and the forest floor becomes covered, so the composition will be changed if the abandonment of the woodland is allowed to advance in the future. We consider the woodland where the forest floor is regularly managed to be limited, and the woodland where the abandonment period is short, the forest floor was bright has extended comparatively in the extant woodland in Shiroi city.

View full abstract