Comparing the impact of live-tree versus historic-timber data on palaeoenvironmental inferences in tree-ring science, eastern North America

Abstract

Dendroarchaeological data from historic structures and artefacts have the potential to extend tree-ring chronologies spatially and temporally, especially where old-growth forests have been extensively modified or harvested. While these data may contribute to an improved understanding of past climate and ecology, critical differences in the properties of live-tree and historic-timber data might affect results and interpretations of large-scale studies, such as those relying on large datasets from public databases like the International Tree-Ring Data Bank (ITRDB). The objective of this work was to compare summary measures of live-tree versus historic-timber datasets likely to affect outcomes and inferences of typical paleoenvironmental applications. We used 99 live-tree (LT) and 41 historic-timber (HT) datasets collected in the Appalachian region of the eastern United States and compared common analytical measures for understanding past climate and ecology, including temporal coverage, species composition, recruitment patterns, segment length, series coherence/mean interseries correlation (as Rbar), expressed population signal (EPS), subsample signal strength (SSS) and response to drought and extreme climate events. We found that tree-ring data from historic timbers record some ecological events similarly to live trees and are sensitive to some climate conditions, with important caveats related to the influence of site and tree selection on analytical measures. In some cases, these caveats can be overcome through improved collection of metadata and additional analyses. In all cases, potential differences in LT and HT data should be considered by those who perform large-scale analyses using public tree ring databases, especially as more scientists contribute historic-timber datasets.