Dendroclimatic estimates of a drought index for northern Virginia

Abstract

A 230-year record of the Palmer drought-severity index (PDSI) was estimated for northern Virginia from variations in widths of tree rings. Increment cores were extracted from eastern hemlock, Tsuga canadensis (L.) Carr., at three locations in northern Virginia. Measurements of annual growth increments were made and converted to standardized indices of growth. A response function was derived for hemlock to determine the growth-climate relationship. Growth was positively correlated with precipitation and negatively correlated with temperature during the May-July growing season. Combined standardized indices of growth were calibrated with the July PDSI. Growth accounted for 20-30 percent of the PDSI variance. Further regressions using factor scores of combined tree growth indices resulted in a small but significant improvement. Greatest improvement was made by using factor scores of growth indices of individual trees, thereby accounting for 64 percent of the July PDSI variance in the regression. Comparison of the results with a 241-year reconstruction from New York showed good agreement between low-frequency climatic trends. Analysis of the estimated Central Mountain climatic division of Virginia PDSI record indicated that, relative to the longterm record (1746-1975), dry years have occurred in disproportionally larger numbers during the last half of the 19th century and the mid-20th century. This trend appears reversed for the last half of the 18th century and the first half of the 19th century. Although these results are considered first-generation products, they are encouraging, suggesting that once additional tree-ring chronologies are constructed and techniques are refined, it will be possible to obtain more accurate estimates of prior climatic conditions in the mid-Atlantic region. INTRODUCTION AND OBJECTIVES Extreme climatic variation, drought in particular, has a serious effect on man and his society. Historically, crop failures and famine have frequently followed or accompanied widespread or long term droughts. Today, with our capacity to store great volumes of water we might expect droughts to have little effect. However, as our population continues to grow, the water supplies of many communities are all too often unable to meet the demand placed upon them. This becomes critical during periods of moderate to extreme drought. To avoid crises, advance planning and resource development is a necessity. The object is, through analysis of long climatic records, to predict the probability of droughts and thus be able to plan for them. One of the main stumbling blocks is the lack of continuous, long-term climatic 2 A DROUGHT INDEX FOR NORTHERN VIRGINIA records. In areas where a climatic record of sufficient length is not available, determination of the probability of drought may be aided by the development of a substitute or proxy record. In both North Carolina and New York, recent research based on the relationship of tree growth to climate has demonstrated the feasibility of dendroclimatic reconstructions in humid regions. Using multivariate techniques developed for the semiarid Southwest, proxy records of precipitation and drought severity have been reconstructed (Cleaveland, 1975; Cook and Jacoby, 1977). It is the objective of this study to use similar dendroclimatic techniques to reconstruct a proxy record of drought for the Northern Virgina and the Central Mountain climatic divisions of Virginia.