Formation Mechanisms of Community Boundaries in Subalpine Coniferous Forests

研究了日本北部北海道亚高山森林中边界清晰的林分类型镶嵌景观的形成和动态。主要研究地点为北海道东部知道子半岛安内贝素山的西坡。描述了油松纯林和冷杉、油松、白桦混交林的镶嵌景观。对这两种林分类型的核心和边缘的详细普查表明，不连续的林分边界与沿边界逐渐变化的土壤条件无关。边界动态顺序为云杉-纯林-混交林。这种置换过程可能与树种对树冠积雪和风害的耐受性不同有关。另一个被考察的地点是北海道南部小马阁山山坡上的一处连续不断的景观。利用序贯航片分析了兴安落叶松入侵的动态格局，提出了一种新的模拟器。该模型基于林分层次大小结构的地理扩展模型(Kohyama和Shigesada，1995)和林隙动态景观中林分大小结构的移位斑块镶嵌模型(Kohyama，1993)。利用新的模拟器重新检验了森林纬度区划的响应。与没有斑块镶嵌的情况(即Kohyama和Shigesada)相比，竞争相互作用形成的森林边界移动速度略快，而百年全球变暖后仍有数千年的时间滞后。

This study was carried out to examine the formation and dynamics of the mosaic landscape of forest stand types with clear boundaries in subalpine forests of Hokkaido, northern Japan.Main research site was set on the western slope of Mount On'nebetsu of Shiretoko Peninsula, eastern Hokkaido. The mosaic landscape made by pure stands of Picea glehnii, and mixed stands of Abies sachalinensis, P. glehnii and Betula ermanii was described. Detailed census of core of, and margin between these two stand types showed that the discontinuous stand boundaries were not related to the edaphic conditions that changed gradually along the boundaries. Dynamic sequence around boundaries was from Picea -pure stands to mixed stands. The replacement process was hypothesized to be related to species difference in tolerance against aggravation by snow deposit on crown together with wind. Another site investigated was a successional landscape on the slope of Mount Komagatake in southern Hokkaido. The dynamic pattern of the invasion of Larix kampferi was analysed using sequential aerial photographs.As a generalized modelling tool, a new simulator was proposed through this study. The model is based on the geographically extended model of stand-level tree size structure (Kohyama and Shigesada 1995) and on the shifting patch mosaic model of stands with tree size structure in gap-dynamic landscape of forest (Kohyama 1993). The response of latitudinal forest zonation was re-examined using the new simulator. Compared to the case without patch mosaic (I.e. Kohyama and Shigesada), the forest boundaries made by competitive interaction was predicted to move slightly faster, while still millennia time lag was the case after centennial global warming.