| description abstract | The basis of many statistical tests is the assumption that the data are drawn from a normal population. The change in pressure over one to ten days at 3000 dynamic meters, at 50°N and 70°W, is tested, and the circumstances under which such pressure changes form a normal distribution are described. Correlograms of the pressure change at this location are calculated, by season, for the period 1934 through 1940. From this, a population of near-zero autocorrelation is defined. For this limited population, it is demonstrated that the mean and skewness are zero, and that the kurtosis is three. Next, the distribution is shown to be normal by the ?2 test. Finally, an estimate is given of the size of sample necessary to eliminate the synoptic effect. | |
