FLOOD STAGES, 1912-COMPARATIVE DATA.

Table V shows the highest stages reached at various places during the flood, together with the dates thereof; also the highest recorded stages and dates at the same places previously to the flood of 1912, and the departures of the latter from the same. By flood stage is meant the stage above which damaging overflow ordinarily begins, and the data are repeated for the sake of convenience. On March 1, 1912, the flood stage at Memphis was increased from 33 to 35 feet, but for purposes of comparison the old flood stage of 33 feet has been used throughout this report.

An inspection of the above table shows that new high-water marks were established from Cairo to New Orleans, except in the vicinity of Vicksburg, where a higher stage was prevented by the crevasses at Panther Forest, Ark., and at Salem, near Lake Providence, La. Had not these crevasses occurred, or had they occurred several days later, the crest stage at Vicksburg would certainly have been at least 1 foot above the high-water mark of April 16, 1897, when the stage was 52.5 feet. The crest stage of 22 feet at New Orleans represented the maximum effect of wind and water, and was as high as it could have been under any combination of existing circumstances.

COMPARISON WITH THE FLOODS OF 1882, 1897, AND 1903.

Precipitation. During the month of January, 1912, the precipitation was generally deficient over the Mississippi Basin except southeastern Louisiana, with the greatest deficiency, something over 2 inches, over the Cumberland and the Tennessee watersheds, and between 1 and 2 inches over the remaining areas east of Kansas City, except along the lower Ohio River, where there was a slight excess. In February there was almost a similar deficiency east of the ninety-fifth meridian, and a slight excess to the westward, while in March there was a general excess except over the Mississippi Valley above Keokuk, Iowa, and over the extreme Northwest. The excess ranged from 2 to 4 inches, and was greatest over northwestern Georgia, northern Alabama, Mississippi, southern Arkansas, and northern Louisiana, comprising princi

pally the watersheds of the Tennessee, Yazoo, lower Mississippi, lower Arkansas, and Ouachita Rivers. For the period from January 1 to April 2, inclusive, the latter being the date of the last storm directly concerned in the flood causation, there was an excess of from 2 to 2.5 inches over the Missouri Basin east of Kansas City, and somewhat less to the northwestward; between 1 and 2 inches over the lower Ohio Basin, and generally from 2 to 2.5 inches over the remainder of the Mississippi Basin below the mouth of the Missouri River. There was, however, a deficiency, due to light January and February rains, of over 2 inches in the district extending from the extreme lower Yazoo Valley southwestward through central Louisiana, and an enormous excess over southeastern Louisiana, amounting to 5 inches at New Orleans and 19 inches at Donaldsonville, La. This excess was due to the fact that the January rains, as well as those of March, were unusually heavy.

Charts Nos. 12 to 19, inclusive (appendix), show the total and normal precipitation by months (the data for the first two days of April being incorporated with those for March), the total and normal precipitation for the entire period from January 1 to April 2, inclusive, and the departures from the normals for each month and for the entire period.

Preceding and during the flood of 1882 there was an excess of from 1 to 8 inches of precipitation above Cairo, except in the Missouri Valley, and from 8 to 12 inches below. There was a great excess (3 to 11 inches) below Cairo in January, and a comparatively small one above, while in February there was a general excess east of Kansas City, ranging from 1 to 7 inches, with the maximum over southern Illinois, southeastern Missouri, and Arkansas. During March the precipitation was nearly normal above Cairo, and from 2 to 4 inches in excess below. The flood of 1882, however, was a February flood, and the precipitation of March was not of material consequence.

Charts Nos. 20 to 26, inclusive (appendix), show the total and normal precipitation for January, February, and March, 1882, the total and normal precipitation for January and February combined, the departures from the normals for January and February, and for the two months combined.

In 1897 the precipitation was much less than in 1882. There was an excess in January over the lower Missouri and the Arkansas Basins and in the Central Valley, which brought on the normal winter rise. During February the precipitation was moderate, and really deficient, except over the upper Ohio watershed, but much of that over the Ohio Basin occurred within a short period of time, and started the flood. During March there was a general excess east of Kansas City, ranging, as a rule, from 2 to 4 inches above Cairo, and from 4 to 6 inches below. Roughly speaking, the excess in 1897 was only about one-half that of 1882, with the rapidity of fall as the balancing factor.

Charts Nos. 27 to 32, inclusive (appendix), show the normal precipitation and the actual conditions that occurred.

The precipitation that caused the flood of 1903 did not differ greatly in amount from that of 1897, although there was great difference in its distribution. In 1903 the excess above Cairo was only about one-half as large as in 1897, while below Cairo it was nearly double, the heaviest rains falling during February below Cairo, and during March above.

The charts Nos. 33 to 35, inclusive (appendix), show the actual and normal precipitation for January, February, and March, 1903.

A comparison of the precipitation charts discloses the fact that the precipitation preceding and during the flood of 1882 was much greater than that of the three other floods. That of 1897 was somewhat, although not decidedly, greater than that of 1903, and also somewhat greater than that of 1912, while that of 1903 was the least of all. So that measured by the amount of precipitation, the relative order of the four great floods was as follows: 1882, 1897, 1912, and 1903. These data, while of some interest, are not of great importance, as the rate of fall and the interval between rainstorms, as well as the character of the seasons, are the natural controlling factors.

Comparative stages. The table immediately following shows the highest stages reached at various stations during the floods of 1882, 1897, 1903, and 1912, the highest stage at each station during the four floods being in italics. No reference is made to dates, and the table is inserted merely to display comparative data in convenient form.

Diagrams Nos. I and II (appendix), also show the same data in another form, but for a limited number of stations only.

TABLE NO. VI.-Highest river stages at various places during the floods of 1882, 1897, 1903, 1912, and 1913.1

An inspection of the above table shows clearly that there has been a steady increase in the flood heights below Cairo without a corresponding increase in the quantity of precipitation. While the highest stages occurred in 1912, the greatest precipitation occurred in 1882, and while there were no marked differences in the quantity of precipitation in 1897 and 1903, yet the stages in the latter year were considerably higher, except in certain localities where levee crevasses in 1903 were followed by some depression in the flood plane. This absence of the natural relation between volume and stage is, of course, due to some cause other than differences in the quantity of precipitation, and that cause is the influence of the levees which have gradually been extended until the entire river has been practically canalized between Cairo and the Passes. It should be stated, however, that the flood of 1912 occurred later in the season than usual, permitting the northern and western flood waters to meet and combine with those from the eastern tributaries, and thereby to increase somewhat the stages from Cairo southward.

Just what effect the upper Mississippi stages had upon the stage at Cairo can not now be stated with precision. However, the crest stage of 49.9 feet at Paducah, Ky., in 1912 indicated some disturbance of the normal gage relations between that place and Cairo, the stage of 54 feet at the latter place having been somewhat too high. In 1882 with about the same general conditions over the Ohio watershed above Cairo (Cincinnati about 5 feet higher than in 1912, but with the Cumberland and Tennessee in moderate flood only) the crest stage at Paducah was 50 feet, almost exactly the same as in 1912. In 1882 the crest stage at Cairo was 51.9 feet, with 28.2 feet at St. Louis, whereas in 1912 the crest stage at Cairo was 54 feet (and would have been at least 55 feet had the levees held), while the stage at St. Louis was 30.8 feet, or 2.6 feet higher than

in 1882. How much of this 2.1 or possibly 3.1 feet of excess at Cairo came from the excess of 2.6 feet at St. Louis, and how much was due to the extension of the levee system, the writer is unable to state.

Table VII shows in another form the progressive influence of levee construction upon the gage relations between Cairo and Helena.

The effects of levee crevasses were less noticeable at Memphis than at New Madrid or Helena, and the figures show that with an open river, as in 1882, the difference between the Cairo and Memphis stages was nearly 17 feet, whereas at the present time with a closed river it is only about 9 feet, indicating a rise of 8 feet at Memphis due to the building of levees. As the levee system is now practically complete, so far as linear extent is concerned, this difference of about 8 feet will probably stand for the future, it being assumed that the losses through crevasse effect during the flood of 1912 were about the same at Cairo as at Memphis.

Comparative duration.-Table No. VIII shows the number of days the rivers at various places were at or above the flood stage and the number of days they were 5 feet or more above the flood stage during the floods of 1882, 1897, 1903, and 1912.

It is rather difficult to decide from the above figures as to the relative importance of these four floods, but it appears that, measured by volume of water, the flood of 1882 was the greatest. It is true that the confinement of the waters between the levees in 1897, 1903, and 1912 resulted in a greater velocity of stream flow and consequent shortening of the flood period, but this condition was probably modified considerably by the return of the overflow and crevasse water which would operate to prolong the flood period. It is noted, however, that the extreme flood heights were more prolonged in 1912 than in 1882, 1897, or 1903, due both to levee effect and to the slow return of the overflow and crevasse water. The general superiority of the flood of 1897 over that of 1903 is also indicated, as well as the inferiority of the former to those of 1882 and 1912.

Relative importance of the four floods-Final conclusions. If the estimates of the relative importance of the floods of 1882, 1897, 1903, and 1912, as stated in the foregoing, are accepted as correct, the general conclusions may be summarized as follows:

The maximum stage is of course the principal factor, the one that governs in any discussion of the problem. It is the basis of comparison for the past and of estimates for the future, and all measures that may be devised for absolute protection from future floods must be predicated upon probable gage heights, and be measured by them. To the flood of 1912, therefore, must be assigned first place in the flood history of the lower Mississippi Valley.

DAMAGE AND LOSSES.

In the statement regarding damage and losses no attention will be paid to the impairment of the levees and the amount of money necessary to restore them. The estimates given will be limited to damage to and losses of property and crops and to the losses occasioned by the enforced suspension of business. It must be remembered that it is at all times extremely difficult to arrive at any accurate conclusion as to losses from floods, and the figures given here will, of course, be estimates only. They were, however, based upon careful observation and in many instances upon actual reports from those directly interested, and they are believed to be as accurate as it was possible to make them under the conditions existing at the time. Many losses were of such a character that the money equivalent could not be estimated and many others were unreported, so that to the total losses reported it would probably be reasonable and safe to add at least 10 per cent for others regarding which data were not available. The losses and damage will be classified into four groups as follows:

1. Property losses and damage, exclusive of crops.

2. Crop loss and damage.

3. Damage to farm lands.

4. Losses occasioned through enforced suspension of business.

In the interest of further detail the data will also be localized by reference to existing river districts as maintained by the United States Weather Bureau, it being understood that each district includes the tributary watersheds from the headquarters of the district to the headquarters of the one immedately above.