Dendroclimatic Studies of White Spruce in the Yukon Territory, Canada

An extensive network of 111 white spruce tree-ring chronologies (2983 trees) from treeline sites was developed across the Yukon Territory and adjacent areas of Alaska and British Columbia. Ring-width series from 73 chronologies with adequate signal strength back to 1800 were analysed using correlation and Principal Component analyses. Although 50 chronologies showed a strong common growth pattern over the 1900-1950 period (45.6% of the variance in PC1), PC1 over the 1950-2000 period included only 22 (27.1% of the variance). Correlation with temperature data from the central-north Yukon indicated that 1900-1950 PC1 chronologies showed significant positive relationships to summer (JJA) minimum temperatures and strong negative relationships with prior summer maximum temperatures. Only four of these chronologies retained the positive summer signal for the 1950-2000 period and approximately one third exhibited significant negative responses to spring/summer minimum temperatures during the 1950-2000 period. The loss of positive temperature sensitivity indicates a divergent temperature response in ring width for most sites throughout the north and central Yukon, inhibiting the proposed temperature reconstruction from these data. Analyses of 12 maximum latewood density (MXD) chronologies indicated that nine chronologies have significant relationships with summer maximum or mean temperatures prior to 1950 and six sites, in the central and southern Yukon, retained a slightly weaker but positive summer signal post-1950. Calibration against a regional temperature record (1938-2002) from the southern Yukon indicates that a regional MXD chronology from these six sites captures ca. 39% of the variance of summer (May-August) maximum temperatures. The first, MXD-based, summer maximum temperature reconstruction (1623-2002) was developed for the Yukon Territory. Most of the reconstruction is characterized by high frequency fluctuations with warmer and cooler intervals lasting rarely more than a decade, although the early portion (1630s-1750s) shows a more extended cooler period. This reconstruction showed similarities with the adjacent St. Elias-Wrangell Mountain reconstruction of July-September mean temperatures from Alaska particularly during the 17th and 19th centuries. These results indicate that MXD data are less influenced by divergence and could form the basis for a long temperature reconstruction in the Yukon