ABSTRACT
Meteorological drought and its impact on agriculture in the grain belt of the interior U.S. has spurred much research into the possibility of a regular or predictable pattern in drought recurrence there. The earliest recorded tree-ring study in the Plains, by Jacob Keuchler in central Texas in 1859 (Campbell 1949) pre-dated the development of dendrochronology as a quantitative science. Tree-ring data has since played a major role in providing statistical evidence for a relationship between solar variations and drought. The first widespread tree-ring collections over western North America analyzed for cycles were those of Schulman (1956). Using an optical method of cycle analysis developed by Douglass (1936) Schulman found cycles near the 20-year wavelength in regionally averaged tree-ring series from the Banff area, British Columbia, to the Rio Grande basin. The cycles represented average wavelengths, however, with wide variations in time from peak to peak. Moreover, changes in amplitude, waveform, and phase were common. Schulman concluded that “in no extensive data were cyclic tendencies found of such regularity and strength as to suggest physical reality”. LaMarche and Fritts (1972) found a periodicity near the 22-year double sunspot period in spectral analysis of the scores of the first eigenvector of tree-growth over the western U.S., but found no phase linkage with the sunspot series itself. A statistically significant phase linkage was later reported, however, between the double sunspot series and a 300-year tree-ring reconstruction of an index of drought area in the western U.S. (Mitchell et al. 1979). When the reconstructed drought-area index (DAI) and sunspot series were analyzed with a harmonic dial, a method of examining phase relationships developed by Brier (1961) drought-area was found to peak about two years after alternate minima in the sunspot series. Currie (1981) and Bell (1981) have independently examined the same DAI series and found rhythmic features near the 20-year wavelength, but argue for importance of a component near the 18.6-year lunar nodal cycle.
