Abstract:

Based on the data of forest resource inventory in the west Dongting Lake region, a classification method, suitable for vectorized data, was proposed and developed for multi-scale landscape components. The dominant tree species, age and crown closure, altitude, slope, slope aspect and slope position were selected as the influencing factors for the classification. The small-scale forest landscape classification mainly considers the grading of vegetation factors and environmental factors, however the large-scale classification needs to take consideration of the coupling effects among the elements of forest landscape types. These concerns act as a rule to merge the small forest landscape patches into their adjacent units. In this study the attribute characteristics of space unit in the vector graphics was based to analyze its semantic proximity relation, and the spatial analysis function of ArcGIS was used to realize the merger of space unit. According to the edge effect and coupling strength between different forest landscape types, small patches was eliminated, and merged into patches of landscape types with high environmental coupling. Thus, multiscale landscape classification database was established, and multi-scale forest landscape classification were realized.

Key words: coupling effects, multi-scale, forest landscape, landscape classification, forest resource inventory data for management