Comparing Juniperus virginiana tree-ring chronologies from forest edge vs. forest interior positions in the Cedars Natural Area Preserve in Virginia, USA

We created two eastern red cedar master chronologies, one from trees growing along the edge of grassy openings and another from trees growing within the intact forest canopy. Correlation coefficients were calculated between ring width indices from two time periods (1895–1949 and 1950–2001) of the residual chronology and temperature, precipitation, and Palmer Drought Severity Indices (a standardized measure of dryness). The two time periods represented younger and older eastern red cedar. The younger, interior eastern red cedar had significant, positive correlations between ring width index (RWI) and June and November precipitation, March temperature, and Palmer Drought Severity Index (PDSI) from May through December. There were significant, negative correlations between RWI and May, June, and December temperature and May through July PDSI from the previous year. For the older, interior eastern red cedar significant, positive correlations existed between RWI and precipitation from the previous June, October, and December, and PDSI from July. Significant, negative correlations existed between RWI and January precipitation and May temperature. For the edge eastern red cedar within the earlier time segment (1895–1949) there was only one significant dendroclimatic correlation and this was a negative correlation with December temperature. For the edge eastern red cedar within the later time segment (1950–2001) there were significant, positive correlations between RWI and precipitation from June, September, October, and December of the previous year, March precipitation from the current growing season, and PDSI from July. There were significant, negative correlations between RWI and precipitation from January and November, temperature from the previous June, temperature from May and December, and PDSI from June of the previous year. Thus, eastern red cedar from the interior had more significant correlations to climate than trees growing along the edge. This result does not match other studies that have found edge trees to be more responsive to climate than interior trees. Perhaps this difference can be explained by some of the variations in the significant dendroclimatic correlations between the earlier and later time periods. The differences between the two time periods (within a single site) imply that the environmental conditions of trees changed over time. These differences may be a result of tree encroachment into the forest openings which creates a constantly changing environment for the eastern red cedar and results in some of the high variability of dendroclimatic relationships identified in this study. These results imply that trees growing in communities without stable edges, i.e. where the environment around the trees is in a constant state of flux, would be unsuitable for climatic reconstruction because they do not conform to the uniformitarian principle.