Optimal Sampling Of Vegetation Classification In Evergreen Broad-leaved Forests

Vegetation classification is a fundamental research in vegetation ecology,and it is one of the most complex issues for ecologists.Furthermore,there are the complex forest structures and ecotone in the subtropical evergreen broad-leaved forest.Therefore,this type of forest is more difficult to classify.In addition,method of selection is the major step in vegetation classification.Particularly,plot size and quantitative characteristic selection can influence the result of vegetation classification significantly and determine the sampling method and efficiency of field investigation.Due to the development of quantitative ecology rapidly in recent decades,the method of vegetation classification can be further optimized.Therefore,there is a possible opportunity for studying optimal selection method in vegetation There are large-scale forest community data for each individuals in various habitats in 20 ha Tiantong dynamic monitoring forest plot(hereinafter abbreviated to Tiantong plot).These data can support the study of optimal selection method in vegetation classification.Therefore,we selected Tiantong plot as research object,and classified the matrix of random combination of plot size and quantitative characteristic by partition around medoids.Then,we evaluated the vegetation classification results by different random combination,and selected the optimal selection method.The main results in following:1)According to Mantel test,Silhouette width,Linear discriminant analysis,and Canonical correspondence analysis,the results show the optimal plot size for vegetation classification in Tiantong ploy is 900 m~2,and followed by 400 m~2,625 m~2,1225 m~2,1600 m~2,100 m~2 and 225 m~2.2)According same analysis in above,the results show the optimal quantitative characteristics for vegetation classification in Tiantong plot is abundance,and followed by important values,coverage and presence/absence of data.In addition,transformed data is worse than original data for vegetation classification.3)According to distance-based Redundancy analysis and stability analysis,the results show both plot size and quantitative characteristic have significant effects on the results of the vegetation classification.However,plot size is more important than quantitative characteristic.The optimal combination is 900 m~2(plot size)and abundance(quantitative characteristic).Synthesis.Species composition is fundamental to vegetation quantitative classification.The optimal plot size of vegetation classification of Tiantong subtropical evergreen broad-leaved forest is 900 m~2.The evaluation level of vegetation classification will decrease by selecting larger plot size than 900 m~2 with more disturbance area.The optimal quantitative characteristic is abundance,and important values followed,because there are many shrub species in subtropical evergreen broad-leaved forest.Ultimately,fitted plot size and representative quantitative characteristic are most important for vegetation quantitative classification.We should pay more attention in plot size and quantitative characteristic selection and take into account vegetation types in vegetation quantitative classification.