Habitat Associations Of Dominant Species Spatial Distribution In Old Growth Broad-leaved Korean Pine (Pinus Koraiensis) Forest

Recent studies have documented local-scale associations between tree species and habitat factors in forests worldwide. A30-ha plot (500m×600m) was established in the Jiaohe Forestry Experimental Zone Bureau of Jilin province to study the long term dynamics of a broad-leaved Korean pine mixed forest. Within the plot, all free-standing trees and shrubs with a DBH of at least1cm were identified, tagged and mapped, and their geographic coordinates recorded following a standard field protocol. One objective was to study mechanisms of the formation of spatial distribution patterns of understory tree species and to explore how environmental gradients influence these patterns in a broad-leaved Korean pine mixed forest. Based on the data from the30-ha plot, all trees with a DBH not greater than5cm were defined as under storey trees. Only species represented by at least300individuals were chosen.14tree species representing94.32%of all understorey trees satisfied this condition. First, the spatial patterns of the14selected species were analysed at scale of20m*20m cells, using Morisita’s index(Iδ). The results indicate that all14species show a clumped spatial distribution. Second, the spatial autocorrelations of the14species were studied at15distance classes using the spatial autocorrelation coefficient Moran’s I. The results show a significant positive correlation at small distance classes and significant negative correlations at large spatial distance classes in most species were. Third, the partial Mantel test was used to find the relationship between the abundance of the14tree species and certain habitat factors. The results show a significant correlation with spatial distances for6species which reprtesent42.86%of the tested species.7species (50%) were significantly correlated with habitat factors.4species showed neither a significant correlation with spatial distance nor with habitat factors.3species showed a significant correlation with both distance and at least one habitat factor.5of7dominant species with more than1000individuals had significant correlation with spatial distances.5species had significant correlation with habitat. The partial Mantel tests suggested that the mass effect from neighboring quadrats and habitat association could explain most of the tree species abundance distribution in a quadrat in the broad-leaved Korean pine forest community. At last, the Pearson correlation coefficient between species abundance in a quadrat and microtopographic attributes of the quadrat was calculated to test the effects of microtopography (elevation, slope, aspect, convexity) on species distribution. The results showed that Acer mono, Carpinus cordata preferred lower elevations and southern aspects. Acer barbinerve, Acer ukurunduense and Corylus mandshurica preferred higher elevations, steeper slopes and north-facing aspects. This result shows obvious habitat niche specialization and confirms niche theory. It is concluded that habitat niche specialization and mass effect are important factors in structuring distributions of understorey tree species in a broad-leaved Korean pine forest community. Based the plot, Another objective was to study mechanisms of the formation of dominant species (P. krainsis and T. amuransis) spatial distribution patterns at different layers(overstory, midstory, understory) and to explore how environmental gradients influence these patterns in a broad-leaved Korean pine mixed forest. The results indicate that P. krainsis and T. amuransis show a clumped spatial distribution at small scales and random at large scales. Higher layers tended to be randomly or regularly distributed. While lower layers tended to be aggregated at smaller scales. There was generally a positive correlation at small scales between the two species and no significant correlation at large scales. But the association varied among different layers, P. koraiensis stems showed negative correlation between overstory and understory. At last, the Pearson correlation coefficient between species abundance in a quadrat at every layer and microtopographic attributes of the quadrat was calculated to test the effects of microtopography (elevation, slope, aspect, convexity) on species distribution. Two species showed that low layers were associated with microtopography stronger than high layers. So that habitat had an important effort on distribution of P. krainsis and T. amuransis. Many tree species prefer similar habitats which indicates similar resource usage and niche overlapping. Other tree species prefer specific habitat indicating repellency between different species, which may be the result of long term adaptation to different habitats, different resource usage, and which may explain the niche separation.