A TEST OF THE REALITY OF RAINFALL SINGULARITIES

Abstract

The tendency for the recurrence of weather events on or near the same calendar date has been the subject of controversy for many years. The statistical evidence has been far from conclusive, and no generally acceptable physical explanation of such phenomena, if real, is available. Dr. E. G. Bowen has advanced the meteoritic-dust hypothesis to explain singularities in the number of ice nuclei and in world rainfall amounts. This paper describes a statistical test comparing three independent series of daily rainfall: (1) world rainfall amounts determined from several hundred stations assembled by Bowen for the period 1880 to 1950; (2) average daily precipitation amounts during the period 1952 to 1957 for a network of approximately 150 stations distributed over the United States; and (3) average daily precipitation amounts for the same United States network during the year 1958. A non-parametric test, made possible through, use of an electronic computer, shows a highly significant association among these series. These results lead to the conclusion that there has been a strong tendency for precipitation anomalies (both high and low) to occur on specific calendar dates.