Habitat loss and fragmentation have been associated with the decline of endangered species. In 1987, a catastrophic fire in the northern Great Hing＇an Mountains of China, where the main habitat of sables （Martes zibellina） is located, aggravated the loss and fragmentation of the forest landscape. Due to restricted distribution and low population density, sables were listed in the national first-grade protected species in China. The objective of this paper was to identify to what extent the habitat of sables had been restored 13 years after the fire. Based on the behavioral data, which came from field survey information by radio-tracking, GPS （Global Positioning System） and forest inventory data, suitability habitat maps were derived using the Ecological Niche Suitability Model （ENSM）. In addition, the habitat structure was analyzed with selected landscape indices. Although forest cover mostly had been restored by 2000, the results indicated that, compared to the pre-fire situation, the areas of suitable habitat had been reduced significantly, especially those of less suitable, marginally suitable and moderately suitable designation. Fragmentation was aggravated, and suitable patches were found to be further isolated with the exception of those in most suitable areas. The ratio of the patch perimeter to area in unsuitable, moderately suitable and suitable areas decreased, while the ratios within other suitability types increased. Moreover, the percentage of soft boundaries decreased slightly, which can influence the redistribution of sables. The results above indicated that the suitable habitat had deteriorated, and the restoration of the sables＇ habitat remained to be done.
The restoration of forest landscape has drawn much attention since thecatastrophic fire took place on the northern slope of Great Xing'an Mountains in 1987. Forest canopydensity, which has close relation to forest productivity, was selected as a key factor to find howmuch the forest quality was changed 13 years after fire, and how fire severity, regeneration way andterrain factors influenced the restoration of forest canopy density, based on forest inventory datain China, and using Kendall Bivariate Correlation Analysis, and Distances Correlation Analysis. Theresults showed that fire severity which was inversely correlated with forest canopy density gradewas an initial factor among all that selected. Regeneration way which did not remarkably affectforest canopy density restoration in short period, may shorten the cycle of forest succession andpromote the forest productivity of conophorium in the future. Among the three terrain factors, theeffect of slope was the strongest, the position on slope was the second and the aspect was the last.
Knowledge about cumulative effects of forest management alternatives on forest landscape is required to make forest management decision. In this paper,a spatially explicit landscape model,LANDIS,was applied to simulate forest landscape changes in 200 years under four management alternatives (no cutting,clearcutting,selective cutting I and II) in Youhao Forestry Bureau located in Small Khingan Mountains. APACK was used to calculate distribution area of the representative species and species age cohort for six species. The results showed:1) timber harvest decreased area percentage of representative conifer species,Pinus koraiensis,Picea koraiensis and Picea jezoensis,Larix gmelinii to some extent compared to no cutting. The most influencing cutting mode for the area percentage of Tilia amurensis and Quercus mongolica was selective cutting II,followed by selective cutting I and clearcutting. To the contrast,the change of area percentage of Betula phatyphylla was contrary to the management alternatives; 2) As to species age cohort composition,timber harvest significantly changed age structure,that is,it decreased over-matured age cohort of representative species,and increased seedling and middle-age cohort (B. phatyphylla was not included,because its area percentage of over-mature age cohort was the highest under clearcutting than other three scenarios).
