Final Report: Retrospective studies of the effects of green tree retention on conifer production and biodiversity on the Willamette National Forest

Practices developed from ecosystem science, such as greentree retention, are becoming integrated into forest management asworking hypotheses, rather than as rigorously established facts.Although in the long run manipulative experiments may provide themost definitive information on the consequences of new practices,in the short term other sources of information are needed. Thisretrospective study of natural, two-aged forest stands on theWillamette National Forest is intended to provide suchinformation.
Eighteen stands throughout the Forest were sampled, 14 inthe western hemlock zone and four in the Pacific silver fir zone.In each stand, adjacent, paired plots were measured, one withresidual trees over a younger cohort and one with only theyounger cohort. Younger cohort stand ages ranged from 66 to 131years for plots in the western hemlock zone, and from 82 to 236years for plots in the Pacific silver fir zone. Characteristicsof both the residual trees and the younger cohort were recorded,to describe these stands as natural models for green treeretention, and to estimate quantitative effects of the residualtrees on the younger cohort and on the total stand.
The natural, two-aged stands in this study in the westernhemlock zone provide a precedent for management with green treeretention. For the majority of the plots with residual trees,live residual tree density was within the range of prescriptions of recent National Forest plans for western Oregon (i.e. 5 to 25trees per ha, or 2 to 10/ac). Broken and/or dead tops werecommon among the live residual trees. Although Douglas-fir wasthe dominant species in the younger cohort, western hemlock waspresent in most of the plots in the western hemlock zone, whetheror not residual trees were present.
For the western hemlock zone stands, basal area, volume,mean annual increment, current volume growth rate of the youngercohort, and basal area growth of the younger cohort plus residualtrees all declined with increasing residual tree basal area.These relationships were best described by curvilinear models;the greatest effect per unit residual basal area occurred at lowresidual tree levels. The relationship between younger cohortmean annual increment (MAI) and estimated basal area of residualtrees at the time of establishment of the younger cohort was thestrongest pairwise relationship between measures of the youngercohort and the residual stands. For 10 m2/ha residual tree basalarea (44 ft2/ac; equivalent to about 12 average-sized residualtrees per ha for these stands or 5/ac), the regression model forthis relationship predicted a decline in younger cohort MAI of26% (95% confidence interval -30 to -22%). Total live standingvolume (younger cohort plus residual) was not significantlycorrelated with any measure of the residual component of thestands.
The stands in this study were unlike typical Douglas-fir plantations in important respects. Density of trees was muchhigher than under a management regime including thinning. Theperiod of time over which the younger cohort became establishedwas much longer than the period allowed under law forcontemporary post-harvest stands. These differences complicateapplication of these results to managed stands, and may indicateecologically important distinctions between natural and managedstands. Because of the nonlinear relationship between residualtree basal area and younger cohort volume and basal area,extrapolation larger spatial scales such as cutting units isnot straightforward. Extrapolation to larger spatial scales willrequire a better understanding of the effects of low levels ofresidual trees on growth of the younger cohort.