G

span a 700-km zone from the Kenai Peninsula southeastward to Boranof Island, were derived from either mountain hemlock (Tsuga mertensiana Bong. Carr.) or Sitka spruce (Picea sitchensis Bong. Carr.). The seven chronologies are well replicated over the interval 15891990. For the common period of 402 years, the mean correlation between the Alaska ring-width series is r = 0.43 (range 0.22-0.72). These correlations, all significant at the 99% confidence level, suggest that trees are responding in a coherent manner to climatic variations across the region.

We compared variations in tree-ring widths with a regional record of temperature for three coastal Alaska stations: Kodiak, Seward, and Sitka. Tree growth is positively correlated with temperature variations over most of the year. The first principal component (PC) scores of the seven tree-ring records are significantly correlated with temperature from January to June and from August to October (Fig. 2b). These results are consistent with those of previous studies showing that coastal Alaska tree rings strongly reflect spring-summer temperature variations in the northeastern Pacific (Briffa et al., 1992; Wiles et al., 1996, 1998).

Both the Patagonia and Alaska tree-ring records clearly show the documented 1976 transition from cold to warm conditions over the tropical Pacific, reflecting a comparable sensitivity to SST changes in the Pacific Ocean (Fig. 3). If decadal-scale climatic variations forced by the tropical Pacific had affected past temper-

Grsmng.itaioii

Grsmng.itaioii

FIGURE 2 Correlation functions, calculated over the interval 1949-88, showing the relationships between regional temperature variations and tree growth in (a) northern Patagonia and (b) coastal Alaska. For northern Patagonia, the leading scores resulting from a principal components (PCs) analysis of three upper elevation chronologies were used for comparison with temperature. For Alaska, the tree-ring record used for comparison is the first PCs resulting from seven chronologies located along the Gulf of Alaska. Positive correlation indicates that above-average tree growth is associated with above-average values of temperature. The meteorological stations included in the temperature regional records are Bariloche, Mascardi, and Esquel in northern Patagonia; and Kodiak, Seward, and Sitka in Alaska. Correlation coefficients greater than 0.27 and 0.35 are significant at the 95 and 99% confidence level, respectively.

FIGURE 3 Comparison of annual (April-March) temperature fluctuations in (a) northern Patagonia and (c) Alaska with the amplitudes from the leading components resulting from principal components (PCs) analyses of the (b) three northern Patagonian and (d) seven coastal Alaska chronologies listed in Table 1. The simultaneous rise in temperature and tree growth starting in both regions during the mid-1970s is indicated by a vertical dotted line.

FIGURE 3 Comparison of annual (April-March) temperature fluctuations in (a) northern Patagonia and (c) Alaska with the amplitudes from the leading components resulting from principal components (PCs) analyses of the (b) three northern Patagonian and (d) seven coastal Alaska chronologies listed in Table 1. The simultaneous rise in temperature and tree growth starting in both regions during the mid-1970s is indicated by a vertical dotted line.

ature changes along the extratropical coasts of North and South America, tree-ring records for Alaska and northern Patagonia should present similar oscillatory modes. Indeed, for the common interval 1591-1988, the amplitudes from the first PC of the Alaska chronologies are significantly correlated with those of Patagonia (r = 0.35, p <0.0001; Fig. 4), which may be considered a first indication of a common climatic forcing mechanism affecting past temperature variations in both regions.

0 0

Post a comment