Detection of high-wind events using tree-ring data

Windstorms are common events in midlatitude west coast climates yet little is known about their long-term history and influence on regional forests. In this paper, we present a procedure that detects the timing and frequency of these high-wind events (HWEs) along the Pacific Northwest coast of North America using crossdated tree-ring measurements and detrended tree-ring chronologies derived from windsnapped trees. Our results show that abrupt changes in ring width patterns closely match dates of known HWEs and can serve as a nonclimatic basis for crossdating. Experimentation with different growth suppression and release criteria revealed that either a 50% decrease or a 50% increase in radial growth relative to the mean index value provided the best criterion for detecting windstorm-related growth anomalies. Comparing these results with the known occurrence of windstorms during the past century successfully identified all known major wind events during the study period. Low correlations between tree growth and climatic variables further imply that HWEs are the principal cause of interannual growth variations. Accordingly, we discuss the application of our procedure toward the development of a multicentury reconstruction of HWEs, a long-term analysis of decadal climate variability and HWEs, and the ecological role of HWEs in Pacific Northwest and other west coast forests.