Files
Abstract
Two data sets were used in this study. Data set one came from a large permanent plot in a natural mixed temperate forest in the southeaster United States; data set two was from Continuous Forest Inventory (CFI) plots of bottomland mixed species hardwood stands in the lower reaches of the Mississippi River.|First, an objective approach (Principal Component and Cluster Analysis procedure) to group tree species according to their characteristics, such as tree growth, size structure, mortality, recruitment and light demand of species, was developed and successfully applied to these two data sets.|For the completely stem-mapped tree data set 1, the tree spatial patterns of within-species- group and the spatial associations among species groups were analyzed with Ripleys K(d) function and Monte Carlo simulation. All spatial patterns of the live trees of within-species-group were significant clustering distributions over a wide range of scales and the clustering intensities and the pattern scales changed as stand developed. Some of the spatial associations among species groups were significantly attractive, some were significantly repulsive, and some were independent. Distance-dependent individual tree growth models were developed and used to analyze the interactions of trees among and within species groups in detail.|For data set 2, a multi-species density-dependent matrix growth model and distance-independent individual tree basal area and mortality models were developed and these two modeling approaches were compared.