Spatial Patterns of Tree Growth Anomalies in the Pacific Northwest

Publisher Website
Google Scholar
Abstract
Regional patterns of ring width anomalies in Washington, Oregon, and Idaho, USA forest are examined using eigenvector (principal component) techniques. The first two eigenvectors, accounting for nearly 50% of the total variance, represent large—scale spatial patterns. Eigenvector I represents a pattern in which growth anomalies are positively correlated among all sites, and Eigenvector II one in which growth anomalies are negatively correlated between sites located on opposite sides of the Cascade Mountain crest. These patterns most likely result from two types of tree growth responses to climate variations that extend uniformly across the study area. Common responses to spring—summer rainfall probably cause the positive, region—wide correlations identified by Eigenvector I, and opposite east—west responses to summer temperature and winter precipitation may account for the negative correlations identified by Eigenvector II. No evidence supports the idea that the tree growth eigenvectors directly reflect two distinctive patterns of climate anomalies. The spatial patterns of these tree growth anomalies have remained essentially constant during the past 300 (possibly 400) yr.