IDENTIFYING OLD TREES TO INFORM ECOLOGICAL RESTORATION IN MONTANE FORESTS OF THE CENTRAL ROCKY MOUNTAINS, USA

摘要

Old trees (defined here as >= 150 years old) can be rare in many forests because of past timber harvest, uncharacteristically severe wildfires, and-increasingly-climate change. Old trees provide unique structural, ecological, scientific, and aesthetic values missing in forests containing only younger trees. Here we compile crossdated ages from over 10,000 living and dead trees sampled in montane forests of the central Rocky Mountains in Colorado and southern Wyoming, USA, to examine changes in age structure of the oldest trees since Euro-American settlement and to provide guidelines to aid in identification of old trees for retention during ecological restoration treatments. Eroded stumps (containing only heartwood) were found in over 93% of 179 randomly sampled plots. Number of stumps found in each plot was proportional to reconstructed historical (1860 C. E.) stand basal area. The regional median date of maximum plot tree recruitment was over 150 years older when including stumps versus only living trees, suggesting that if all those harvested trees had survived to the present, the ages of oldest trees would be substantially greater than it is today. However, the regional median age of oldest trees in 1860 before harvesting was not different from the median age of oldest living trees in the current forest (246 vs. 248 years), which alternatively suggests that the regional population of oldest trees has recovered to near historical levels in the time since early Euro-American harvests. Each living tree at the time of sampling was assigned to one of three potential age classes based on a subjective assessment of tree morphology: old (likely >= 150 years old), young (likely <150 years old), or transitional (containing a mixture of young and old tree characteristics). Trees assigned to the old and young morphology categories were classified correctly 88% to 96% of the time depending on species as confirmed by their crossdated ages. Regression tree analysis revealed that tree diameter at breast height was not as reliable a predictor of tree age as were morphological characteristics. A measure of site productivity was a significant variable to use to separate transitional morphology trees into old and young age classes, but classification accuracy was not high because of large variability in ages of these trees. Our results suggest that residual live old trees in the current forest, although perhaps not rare compared to historical age distributions, should be retained during restoration treatments, and that using simple morphological and environmental criteria to identify old trees is more reliable than tree size alone.