38

34

DURING THE 29-YEAR PERIOD, 1888-1916.

50

1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916

46

42

38

30

the average monthly rainfall for each state. By taking as weights the area of each state in square miles, it was possible to compute the average depth of rainfall for a given month over any desired section of the United States by adding the weighted means in inch-miles for all the states within such section, and dividing the sum by the aggregate area of the section in square miles. The total rainfall each year for the section was then obtained by adding the averages, so computed, for the twelve calendar months.

Figures 107 and 108 show that the variations in annual rain all for the larger area are approximately two-thirds as great as those for the smaller area. The average yearly departure from the mean annual rainfall for the two areas studied, and also for Dayton and for Ohio, for the period considered, was found to be as follows:

This indicates, as seems reasonable, that the larger the area considered the smaller is the variation in total annual precipitation over that area.

Figures 107 and 108 possess many features in common. Comparing them year by year, the two diagrams agree in 72 per cent of the cases as to annual totals being greater or less than the mean annual rainfall. The great period of drouth which occurred in 1894 and 1895 is represented in figure 107 by deficiencies of 5 and 4 inches, and in figure 108 by deficiencies of 8 and 7 inches, respectively.

area.

In a few instances there are marked discrepancies between the two diagrams, which must be attributed to decided differences in precipitation over large areas. Thus, for the year 1900, the rainfall over the eastern half of the United States averaged 3 inches in excess of the 29-year mean, as against a deficiency of 24 inches for the smaller In 1909 the smaller area averaged an excess of 5 inches while the rainfall over the larger area averaged normal. The greatest excess shown in figure 108 is 9 inches for 1890, as against only 2 inches for the same year in figure 107. In making these comparisons it should, of course, be borne in mind that the fluctuations, for the larger area tend to be less than for the smaller area, the former having not only a smaller mean annual rainfall, but also the greater tendency for compensating extreme local variations.

A striking feature of both diagrams is the absence of any evidence as to cyclic or progressive variation. Although it may be argued that the 29-year period is too short to furnish conclusive evidence

regarding this matter, it seems reasonable to suppose that if any important cyclic variation does exist some indications, at least, would be present in one or both of these diagrams.

In order to show graphically the extent of the departure of the maximum rainfall from the mean for different continuous periods, four diagrams were constructed, figures 109 to 112. The first two figures apply to the larger area; the latter two, to the smaller area. Figure 109 shows the maximum departure in inches from the monthly means for continuous periods varying in length from 1 to 6 months. Thus the greatest excess of precipitation for any month during the 29 years over the mean for that month, was in February, 1903, when the average over the 37 states was 4.30 inches, an excess of 1.82 inches. The maximum deficiency for any one month was in March, 1910, when the average precipitation over 37 states was 1.99 inches below the mean. In the upper portion of the diagram, the monthly precipitation for the 6 months, January to June, 1911, shows an average monthly deficiency of only 0.68 inch below the 29-year means for those months. Similarly, the 6-month period, September, 1902, to February, 1903, has an average monthly excess of 0.81 inch over the means for the period.

An inspection of the chart shows that the principal excesses and deficiencies generally occur during the spring and early summer months, illustrating a tendency for the depth of rainfall to be more erratic during these months than at other times of the year.

The second diagram, figure 110, shows the maximum variation for continuous periods of 1, 2, 3, 4, 8, 12, 16, 20, 24, and 28 years, respectively, from the 29-year mean. The departures are expressed as percentages of the 29-year mean.

Figures 111 and 112 are similar to figures 109 and 110. Figure 111 shows that the maximum departures from the normal rainfall also occur during the spring and early summer months in the case of the smaller area. In practically all cases the actual values of the departures are seen to be considerably greater for the smaller area than for the larger area, as would be expected.

EFFECT OF STORMS AND EXCESSIVE PRECIPITATION

It is conceivable that the variations in the number of great storms occurring each year might be responsible for a large part of the fluctuations in the total annual rainfall. With this in mind table 10 was prepared, showing the chronological distribution of the 160 storms previously discussed. Table 2 is analogous, showing the chronological distribution of rainfalls in which there were one or more records of excessive precipitation. A comparison of these two tables with the