An empirical evaluation of a three-dimensional crown model for predicting volume growth

Abstract

A three-dimensional geometric model of crown competition was developed which can be used for empirical distance-dependent individual-tree models or biologically based process models. The crown model calculates the shape of a subject tree's crown based on the positions of its competitors' crowns and the rules of competition. It is necessary to use a crown profile equation to use the crown model. The crown profile equation was estimated from the crown width of open-grown trees at the base of the live crown. Ten replicates of a loblolly pine ( Pinus taeda L.) progeny study planted in a variable density Nelder's wheel layout re-measured annually from Ages 6 to 10 and at Age 12 were used for the analysis. The use of two different rules of competition and two different methods of calculating crown volume, surface area, and exposed surface area resulted in eight different estimates related to the geometry of each crown. The relationships of the natural logarithm of periodic annual wood volume growth of different time lengths were fit to geometric estimates generated from the crown model. These relationships were compared with crown length and wood volume relationships to evaluate the utility of the three-dimensional crown model. Exposed surface area of the crown was the best predictor of volume growth produced by the crown model. Wood volume was the single best predictor of growth until Age 9, when exposed surface area became superior. The best model of all possible models was wood volume, exposed surface area, and surface area. Wood volume, exposed surface area, and total surface area length accounted for 1–11% more of the variability than crown length and wood volume. This indicated that the crown model offered benefits over the use of the raw data. The relative performance of the crown model compared with wood volume and crown length improved as competition increased. The crown model appears to offer potential for growth and yield modeling; it must be further evaluated, however, for use in mature stands.