High-Resolution Image

Rainfall over the SE U.S. (region in box at top of figure) appears to vary with the day of the week during the summertime, presumably due to the well-documented weekly variation in pollution levels in the U.S. and the way pollution affects the development of storms. For the eleventh summer in a row, afternoon rainfall, as estimated by satellite, tends to be larger Tuesday to Friday than on weekends.

The two "clock plots" at the bottom of the figure show how strong the weekly cycle of rainfall, as measured by a sinusoidal fit to the data, is when compared with the "noise" level in such fits--an estimate of how large such fits might be when the natural variability of rainfall in the region is taken into account. The results for each summer are shown as colored disks, with the year notated in each disk. The position of the disk on the plot indicates by its distance from the center how strong the weekly-cycle signal is relative to the noise levels for that summer, and by its direction the day of the week on which the sinusoidal fit peaks. For the years 1998 to 2007 the weekly cycle maximums tended to avoid the "weekend" sector (left plot), though the signal strengths were not very large in some years. Data for 2008 just became available. The weekly cycle this summer peaked on Friday (right plot), but the signal was weak compared with signal strengths earlier in the decade.

A detailed discussion of what has been learned about the weekly cycle in this region may be found in an article by Bell et al. (2008). The rainfall data were obtained from the Giovanni web site, and are based on multi-satellite rainfall estimates from the Tropical Rainfall Measuring Mission (Product 3B42, or TMPA).

(Submitted by T. L. Bell)