Correlation among spatial and non-spatial variables describing a cut-to-length thinning site in the Pacific Northwest, USA.

摘要

Variables describing the pre- and post-logging conditions of a thinning site in 47-yr-old naturally regenerated stand of second-growth Douglas fir (Pseudotsuga menziesii) and western hemlock (Tsuga heterophylla) in western Oregon, were examined forcorrelation, and subsequently used to develop models to estimate residual stand damage levels. A cut-to-length harvesting system was utilized to perform the thinning operation, which used a single-grip harvester and a forwarder, and marked logging trails. Several of the variables were measured in an intensive field survey; other variables were developed using geographic information system (GIS) processes. An analysis of correlations among the site variables showed several obvious, and a few interesting,results that describe the operation. Most of the variables provided negative, or inconclusive, assistance in describing the variation in stand damage levels. Only one variable, the number of original trees/hectare, was significantly correlated with residual stand damage levels, and was represented in the models that were developed to estimate residual stand damage levels. The resulting models are of limited practical value, however, since they explain little of the variability in damage levels. Most ofthe variation in residual stand damage levels may well be explained by random chance, operator error, other unmeasured operational variables associated with this harvesting system, or interactions among variables. The main conclusion from the study is that although both spatial and non-spatial data were utilized in describing the logging operation and in developing models to estimate stand damage levels, the importance of using spatial data was inconclusive.