Spatio-temporal pattern analysis of managed forest landscapes: A simulation approach.

摘要

This study dealt with research problems at the landscape level. The objectives of this thesis were to develop tools to study and characterize landscapes and to interface with a geographic information system (GIS), to evaluate landscape indices, and to examine development of forest cutting patterns under different cutting methods and explore alternative forest management strategies.Many extant landscape indices were reviewed, and some new indices proposed. Each was evaluated in terms of its ability to distinguish four test synthetic landscapes with distinct spatial patterns. Fractal dimension, patchiness index, dispersal index, and two fragmentation indices (i.e., the forest interior area and the largest forest patch size) appeared to be most sensitive to spatial variations among the test landscape mosaics, and may be most useful to study and quantify the landscape pattern. On the other hand, some commonly-used landscape indices, contagion and dominance, could not distinguish variations in distinct landscape patterns.The simulation program and the landscape indices were then used to study landscape patterns generated by different forest cutting methods. The results indicated that different cutting designs may produce landscapes with distinct characteristics. Landscapes were clearly less fragmented when larger sizes of cut-units were used. When a stream system was included in the landscape structure, the behavior of many landscape characteristics changed. The results suggested that simple landscape models (i.e., the checkerboard model and random model) may lead to misleading interpretations of landscape patterns.A computer program was developed for simulation and analysis of landscape patterns. The primary applications of the computer program were (1) to quantify spatial patterns of landscapes, (2) to perform experiments with different silvicultural strategies and forecast the consequences of management activities, (3) to examine the behavior of landscape indices without having a large number of landscape samples, (4) to interface with and to complement GIS in terms of ecological analysis, and (5) to serve as a base on which GIS-related landscape models could built.