The spatial autocorrelation of individual tree characteristics in loblolly pine stands
Mathematical methods of assessing the spatial autocorrelation associated with individual tree characteristics in forest stands were identified. These measures were used to investigate the spatial autocorrelation of discrete tree characteristics including the species, product, and defect classifications. With the exception of the species classification, none of the discrete tree characteristics examined showed any evidence of significant (α = 0.05) levels of spatial autocorrelation in loblolly pine stands. The significant autocorrelation of the species classification was probably due to past stand history or microsite variability rather than overall stand conditions such as age, density, or percent pine.
The relationship between the level of spatial autocorrelation associated with basal area and several descriptive stand characteristics was also examined. No strong relationships were identified but trends were noticed between the autocorrelation measures and measures of stand competition such as basal area and crown competition factor. The measures of spatial association indicate positive autocorrelation between the characteristics of neighboring trees at very low levels of competition with the autocorrelation becoming increasingly negative as competition increases. At extremely high levels of competition, the spatial autocorrelation measures become positive again, reflecting the stagnated condition of the stand.
Methods were developed, using the measures of spatial autocorrelation, to assign characteristics to individual trees in computer generated stands. These methods, applicable for discrete or continuous characteristics; assign the characteristics to individual trees depending on the spatial location of the individual tree and the locations and characteristics of its neighbors.