Community Structure And Spatial Pattern In The Valley Spruce-fir Forest In Liangshui Of Xiao Xing’an Mountains, China

Community assembly process can be regarded as a selection process by which species from a regional pool elected and colonized to form local communities through multilayer habitat filtering and biological interactions. Driven by two basic factors of habitat filtering and biological interactions, local community structure and various spatial patterns formed, reflecting the species population dynamics and maintaining the community biodiversity. As climax community in valley area of Xiaoxing’an Mountains, spruce-fir valley forest is one of the important typical zonal vegetations in temperate forest northeast of China. Based on data from spruce-fir valley forest 9.12 hm2 forest dynamic plot in Liangshui National Natural Reserve, we analyzed species composition, DBH size class distribution, and spatial pattern of species distribution using method of point pattern of different DBH size classes for 16 major species, the roles of density dependent effects in different life history stages of dominant species. We also examined four hypotheses about neighborhood effect at the same time, removing the effect of habitat heterogeneity using the method of point pattern analysis. The main results are as follows:(1) We documented 44 woody species with 35,656 individuals, belonging to 14 families and 29 genera, including 20 rare species and 11 occasional species. The DBH distribution of individuals of canopy and understory and two dominant species generally showed reversed "J" type distribution, and shrub showed "L" type distribution.(2) Using the pair-correlation function based on the method of point pattern analysis to calculate the aggregating intensity of three life history stages. Saplings (1 cm< DBH<5 cm) of major species presented clumped distribution across the scale of 1-50 m, and the aggregation intensity decreased as scale increased. The aggregation of juveniles (5 cm< DBH<10 cm) shifted from clumped distribution to random distribution as scale increased. Adults (DBH>10 cm) showed random or uniform distribution approximately.(3) Using case-control design and random-labelling null model belonging to the point pattern analysis to estimate density dependence in dominant species. The results showed that saplings and juveniles of Abies nephrolepis exploit the habitat differently from adults. For Picea koraiensis, saplings followed different ways from adults to exploit the spatially available habitat, and juveniles followed the same ways from adults to exploit the spatially available habitat. Saplings and juveniles of Abies nephrolepis and Picea koraiensis displayed additional clustering compared with adults, suggesting the occurrence of density-dependent thinning in the stage of both sapling-to-juvenile and juvenile-to-adult.(4) Testing the neighborhood interactions, constructing test statistics based on the point pattern analysis, improving the random-labelling null model to consider the effect of habitat heterogeneity. Our study provides evidence for crucial roles of both neighbour size and species identity for determining local competition. And the large conspecific neighbours generally exert a stronger negative effect on smaller trees than heterospecific neighbours across distances, confirming the Janzen-Connell hypothesis and partly supporting the spatial segregation hypothesis and the size asymmetry hypothesis. There was no general pattern of conspecific and heterospecific interaction distances within and between size classes, and thus, there was no evidence for heteromyopia. The results indicate that the competition of conspecific larger neighbour is the key factor to form the pattern of dominant species in Liangshui valley spruce-fir forest.