Estimating forest crown area removed by selection cutting: a linked regression-GIS approach based on stump diameters

摘要

The purpose of this research was to develop a model that could be used to provide a spatial representation of uneven-aged silvicultural treatments on forest crown area. We began by developing species-specific linear regression equations relating tree DBH to crown area for eight bottomland tree species at White River National Wildlife refuge, Arkansas, USA. The relationships were highly significant for all species, with coefficients of determination (r~2) ranging from 0.37 for Ulmus crassifolia to nearly 0.80 for Quercus nuttallii and Taxodium distichum. We next located and measured the diameters of more than 4000 stumps from a single tree-group selection timber harvest. Stump locations were recorded with respect to an established grid point system and entered into a geographic Information System (ARC/INFO). The area occupied by the crown of each loggd individual was then estimated by using the stump dimensions (adjusted to DBHs) and the regression equations relating tree DBH to crown area. Our model projected that the selection cuts removed roughly 300 m~2 of basal area from the logged sites resulting in the loss of approx= 55 000 m~2 of crown area. The model developed in this research represents a tool that can be used in conjunction with remote sensing applications to assist in forest inventory and management, at well as to estimate the impacts of selective timber harvest on wildlife.