Constructing models for direct prediction of 5-year crown recession in southwestern Oregon Douglas-fir

Abstract

Crown recession rates were estimated by branch mortality dating on 357 sectioned Douglas-fir (Pseudotsugamenziesii (Mirb.) Franco) stems from temporary plots. Numerous nonlinear, logarithmic, and gamma-theory generalized linear models were developed for predicting 5-year crown recession across a range in tree, stand, and site conditions. Residual analyses and indices of fit demonstrated that a multiplicative model with lognormal errors was the most appropriate model form. The recommended logarithmic model predicts crown recession from current crown ratio, total height, breast height age, height growth, and crown competition factor. Data from southwestern Oregon indicate that within a given stand, trees with midsized crown ratios experience the most rapid crown recession.