Date of Award

1-1-2009

Document Type

Campus Access Thesis

Department

Geography

First Advisor

Gregory J. Carbone

Abstract

There are apparent multiyear periods of breakdown in the relationship between growth indices and local spring precipitation, stream flow and drought data for the Baldcypress chronologies on the Georgia/South Carolina (GA/SC) coastal plain. The correlations of Baldcypress tree ring growth with local growth season precipitation-related indicators exhibit a sequence of high and low periods between 1870 (when early precipitation records start) and 1985 (the year of the latest tree indices). The length of a high/low cycle is approximately 60 years. The highs and lows are most evident in correlations of tree indices with March through June (MAMJ) season local precipitation, in spite of the Baldcypress stem growth season having been generally accepted as being May through August (MJJA).

The change in correlation over time of tree growth indices with MAMJ season precipitation is probably climate driven. The signal is evident in all four of the chronologies located across 200 miles of the GA/SC plain and the signal is a sequence of correlation highs and lows which are contemporaneous across chronologies and coincident with climate signatures. However, the signal is not evident in Baldcypress chronologies farther north in North Carolina and Virginia and farther south in the Florida Panhandle.

The hypothesis for the cause of this temporal change in correlation is the temporal change of precipitation spatial homogeneity on the GA/SC coastal plain. There is no actual breakdown of the precipitation/growth relationship during the correlation lows. The apparent breakdowns are the result of mismatches between the precipitation experienced by a chronology and the precipitation measured at local weather stations because of the spatially heterogeneous precipitation associated with a predominance of airmass thunderstorms. The correlation highs are caused by a predominance of spatially homogeneous precipitation associated with frontal systems.

Baldcypress tree stem growth in Georgia and South Carolina does not correlate well with the local precipitation of the stem growth season (MJJA). Even though Baldcypress growth is responsive to rainfall (rather than ground moisture or standing water), it correlates best with the local precipitation of the MAMJ season. The hypothesis to explain this apparent conundrum is a combination of drivers. The main driver is that the spatial homogeneity of precipitation tends to be greater in the MAMJ season than in the MJJA season. The secondary drivers are: extreme temperature episodes disrupting the precipitation/stem growth relationship in July/August much more than in the spring; the precipitation/height growth relationship in March/April contributing to determining a year's subsequent stem growth.

Rights

© 2009, Ian William Millar

Share

COinS