Abstract:As a basic unit to coordinate the relationship between human activities and natural resources, county is an important area to carry out afforestation project and forest landscape resources management. The case study site was Yixian County of Hebei Province, which is well known for its landscape diversity, more forestation and high forest coverage. With the support of geographic information system (GIS) and principal component analysis (PCA), and based on the date of forest resources inventory and land use, the forest landscape patterns and its hierarchical characteristics were studied at county scale. The results showed that woodland landscape accounted for more than 58% of the total area, poplar, Pinus tabulaeformis, oak trees, apricot and other economic forest were the dominant forest landscape;splitting index (SPLIT), Shannon’s diversity index (SHDI) and patch density (PD) exhibited a strong response to the changes in three levels of organizational hierarchy. With the refinement of the organization level, the degree of landscape fragmentation was increased, the area and quantity of the patch were increased gradually. The higher the organization level in the landscape level and patch type level was, the more obvious the trend of forest landscape contagion and cohesion was, in contrast, the more fine the organization level was, the more obvious the forest landscape aggregation and clumpiness was, and the more prominent the edge effect of the landscape patch was. Moreover, it proved that many factors affected forest landscape patterns and coupled with each other, such as terrain, land use type and living environment. From mountain to plain area, with the enhancement of the human activities and forest landscape, the forest landscape fragmentation, complexity of patch shape and its spatial difference were increased gradually. Therefore, the health management of forest landscape resources should be focused on the key indicators such as landscape diversity, spatial segmentation, core patch area and number, and so on
