Documentary Records of Past Climate and Tropical Cyclones from the Southeastern United States

Climatic changes at interannual and decadal timescales are of important concern to residents of the Southeastern United States. This region has a low water-storage capacity and is extraordinarily susceptible to prolonged drought and fire hazards. Superimposed within interannual climatic variations are high-magnitude flood, snowstorm, and hurricane hazards that occur at daily to weekly timescales. Winter polar outbreaks and killing frosts also occasionally occur, severely impacting the fruit tree industry (Stahle, 1990). Modern climate studies indicate that teleconnections, such as ENSO, the PNA, and the NAO, substantially influence southeastern U.S. climate. A longer perspective is critical to fully investigate the temporal and spatial stability of these important modes of climatic variability. Documentary weather records are available for several areas of the Southeastern United States for the past several hundred years prior to the establishment of the first organized weather bureau in 1871. These records include hundreds of plantation diaries, numerous newspapers, and early instrumental records. The first systematic meteorological observations conducted in the United States began in Charleston, South Carolina in 1838 by Dr. John Lining (Aldredge, 1940). The abundant climatic details of plantation manuscripts for the Southeast are well-known to historians, with many daily records containing weather observations. Yet, despite their huge potential, manuscripts in the Southeast have only been recently utilized for climate reconstruction. Clearly, the reliability of historical reports cannot be taken at face value, with some materials actually being inappropriate for climate reconstruction. Original sources were used if available, as typescript copies may contain erroneous edited information. Written descriptions of severe storms and climate may also reflect individual biases from diarists. Daily historical data were collected whenever possible in order to properly assess data quality, as these data reveal important aspects such as diurnal temperature ranges and the timing of precipitation events. Discontinuities were searched carefully in time series of greater than ten years, and any suspect data were not included in our reconstructions. A major goal of our research was to calibrate and present historical climate data in a format that is closely compatible with the modern record. Two examples of historical reconstructions for Charleston, South Carolina are presented in the following: 1) June-September (warm-season) precipitation frequency and 2) tropical cyclone frequency. Precipitation frequencies from both early instrumental and documentary evidence are among the most reliable and plentiful climatic indicators for the historical climatologist, as these reconstructions can be directly compared with modern climatic data (Mock, 1991). Precipitation frequencies also more accurately reflect homogeneous signals that respond to changes in synoptic-scale atmospheric circulation (Woodhouse and Meko, 1997). With plentiful documentary data within the city limits, Charleston presents an ideal location to reconstruct frequencies of measurable precipitation, defined as days on which precipitation was at least 0.0254 cm (0.01 inches). The Charleston tropical cyclone frequency record is the longest of its kind that is reconstructed for a location in the United States. Its record is continuous since 1778, augmenting the Atlantic Basin historical hurricane database (Fernández-Partagás and Diaz, 1996). A complete reconstruction of June-September precipitation frequencies for Charleston extends back to 1823 (Fig. 1). The twentieth century record reveals both interannual and decadal variability, with somewhat greater precipitation during most of the first half of the twentieth century. A distinctive period of lower precipitation is evident for much of the period from 1845-1870, and this activity is unprecedented relative to the twentieth century record. Such a persistent period of drought most likely had dramatic climate impacts on the antebellum plantation agriculture and lifestyle. Although underestimation of precipitation is a common problem in historical climatic reconstructions, the author is confident that this is not a problem for the 1845-1870 period, as the quality of the documentary data is clearly superior and more plentiful than the relatively wetter pre-1845 period. Results of the tropical cyclone frequency reconstruction for Charleston are displayed for each year and by a 5-year centered moving average (Fig. 2). The number of storms per