Community Structure And Diversity Pattern Of Herbaceous Plants In A Sub-tropical Evergreen Broadleaved Forest Of Gutianshan

Biodiversity patterns and the mechanisms for species coexistence are the most fundamental questions in ecology. Nowadays, forest dynamic plot (FDP) has been a effective way to understand the causes for forest biodiversity patterns and various progresses. Species Inventory in.the FDPs is the basis of the other work, however, most study are based on the woody plants and the understory herbaceous community is rarely understood. The subtropical forest of China is the broadest in the worldwide, in this case, systematical study of the herbaceous plants in this area will help improving the understanding of local biodiversity and species coexistence and promoting the comparation of community features among climate zones.In this study we censused the understory vascular herbaceous plants in a24hm2plot in Gutianshan, described the herbaceous community structure, distribution pattern and biodiversity characteristics, and partitioned contributions of different types of environmental factors and spatial factors in explaining the variations of community composition. As an important component of forests, herbaceous community study contribute to the biodiversity forest plots and provide intact basis for development of species coexistence.81vascular herbaceous species were surveyed in3005×5m2quadrats, belonging to31families and55genera. In floristic composition types, the families of world distribution contribute the biggest part of the total families, the genera and species of tropical distribution make up the most part and world distribution take the second place and a finite proportion of temperate distribution type. The most important5species of the community are Diplopterygium glaucum, Dicranopteris pedata, Carex tristachya, Woodwardia japonica, Carex filicina. D. glaucum often appear with huge members, their module component accounted for65.22%of the total number of community individuals, with relative frequency11.54%and relative coverage81.90%respectively, indicating its extensive distribution range and the dominant position. Species with top15importance values take account for92.19%of the importance value sums. There are44species that distributed in no more than five quadrats, which are54.32%of the species richness, with their important values only have a proportion of2.18%of the total, reflecting the existence of large number of rare species in the community. Shannon-Weiner diversity index and Pielou index are1.39and0.32.In one single quadrat Shannon-Weiner index are shifting between0.00to2.50. According to Ward’s minimum variance clustering analysis, similar plots are spatially continuous and the different components ones show mosaic distribution. A Multivariate Partition Tree (MRT) analysis based on topographic factors divided the community into4classes. The species-abundance distribution pattern is fitted by5models and the Zipf model has the least AIC, implying that the community was structured by both external environmental conditions and species interactions.Linear correlation analysis of environmental factors showing that the a diversity index of herbs were negative correlation with convexity and woody plants basal areas while positively correlated with woody plants diversity. Canonical correspondence analysis of species distribution against6environmental factors showing that the constained axes have13.10%component of the total varience, implying that the community structure is constrained by enviroments.Indicative species test were also used based on herbs,8representative indicator species of particular habitat types were chosen. Torous-translations analysis was used to test the species-habitat associations of38species that appeared in at least5quadrats, including all the species that appeared in no more than5quadrats were combined as1rare species, there are26species out of the38(68.42%) show significant positive or negative association with at least one of the5habitat types of the24hm2plot by using abundance data, while using abundance data it comes28species. Many species prefer low valley and show negative correlation with ridge, and their distribution patterns indicate avoidance to the D. glaucum.Principal coordinates of neighbor matrices (PCNM) variables were selected as spatial factors, they also have important effect on the community spatial distribution pattern, varpatition were used to disentangle the contributions of pure special structure from environment factors in explaining community variations, the adjusted R2of12selected PCNMs and6selected environmental factors are11.69%and10.32%respectively, with a certain degree of overlap.Taken together, we concluded that habitat filtering, biotic interactions and spatial factors will influence the distribution and regeneration patterns of herbaceous vascular plant and consequently, will affect the diversity of the whole community. We suggest dynamic monitoring of herbs should be carried out and it is necessary to evaluate the interactions between herbs and woody plants in forests.