APPENDIX F. RUN-OFF OF GILA RIVER AT SAN CARLOS, ARIZ. 1. The run-off information available to the board is that published in the Water Supply Papers of the Geological Survey; that furnished by the Chief Hydrographer of the Geological Survey, as the result of a recomputation made in January, 1913, of the data from observations taken at San Carlos in 1899, 1900, 1901, 1902, 1903, 1904, and 1905; and that furnished for the years 1910, 1911, and 1912 by the San Francisco office of the Geological Survey. Specific references to the Water Supply Papers will be found in the following discussion. 2. Prior to July, 1899, the flow of the Gila River, as recorded on pages 25 to 30 of Water Supply Paper No. 33, was determined at the Buttes below the mouth of the San Pedro River for the seasonal year from September 1, 1889, to September 1, 1890; the fractional year from August 1, 1895, to December 31, 1895; the full year 1896; the fractional year from January 1, 1897, to October 3, 1897; the year 1898; and the fractional year from January 1, 1899, to June 30, 1899. 3. It has been generally believed that the volume of water passing San Carlos is 90 per cent of that passing the Buttes. The reasons for this assumption are discussed on pages 22 and 23 of Water Supply Paper No. 33, where it is stated that for the period May to August, 1890, 207,936 acre-feet passed the Buttes, of which only 19,488 acre-feet, or 9.4 per cent, was contributed by the San Pedro; and that the synchronous observations at the Buttes and San Carlos extending from July 19 to September 23, 1899, indicated that during that period 74 per cent of the water passing the Buttes was available at San Carlos. It is now possible to extend the latter comparison to cover a somewhat longer period, for revised data furnished in blue-print sheets to the board in 1913 by the Chief Hydrographer of the United States Geological Survey, give San Carlos discharge from July 11 to September 30, 1899, to be 36,896 plus 12,654 plus 12,487 ÷ 82 = mean flow of 757 second-feet; while Water Supply Paper No. 33, page 27, gives for the Buttes between the same dates 36,830 plus 20,835 plus 21,996÷82 = mean flow of 971 second-feet, San Carlos here being 757÷971=78 per cent of the Buttes. Were September excluded the percentage would be 86. Regarding the comparison for August, it is explained on page 22 of Water Supply Paper No. 33 that though the rainfall over the State for that month was very deficient, the precipitation in the headwaters of the San Pedro was abnormally large, thus indicating that the above 86 per cent, which included August, is below the normal ratio. 4. Now, in 1913, still further comparisons are possible, for there have been made available for the use of the board, provisional readings for the years 1911 and 1912 at Kelvin, Guthrie, and Clifton, indicating discharges for the two years, of 876,911 acre-feet at Kelvin as against 690,670 acre-feet for Guthrie and Clifton combined, the latter being about 21 per cent the smaller. Considering the intermediate position of San Carlos, with the San Simon and San Carlos Rivers above and the San Pedro River below, its water supply would have the intermediate value of about 10 per cent less than that at Kelvin or at the Buttes. 5. The factor 90 per cent for reducing Buttes determinations to San Carlos is therefore accepted with increased confidence, and is applied to the determinations at The Buttes as recorded on pages 28-30 of Water Supply Paper No. 33 for last four months of 1889; first 8 months of 1890; last 5 months of 1895; all 12 months of 1896; first 9 months of 1897; all 12 months of 1898; and first 6 months of 1899. The next 3 months of 1899, namely, July, August, and September, are taken from San Carlos measurements as given on blue-print record sheets thereof. 6. It may be noted that the 73,300 acre-feet entered for July, 1899, was for only the last 21 days of the month, but Water Supply Paper No. 33, on page 27, indicates that at The Buttes the river was practically dry during the first 10 days of the month. 7. The missing months in the above record, as well as in subsequent years, entered in parentheses in the table found at the end of this appendix, are not copied from any record, but are supplied from the rainfall for such months at the 14 stations distributed over the watershed above San Carlos; using the proportion, that as the average rainfall for a given month for a long period of years is to the rainfall in the same month of the year under consideration, so is the mean water supply of that month for all the years in which it has been determined, to the probable supply of that missing month. In some instances the results of the proportion were somewhat increased or decreased, dependent on whether the preceding month was wetter or drier than common, for the known figures, as well as reason, indicate that the preceding month has its influence. 8. The last 9 months of 1900 and the 12 months of 1901 are taken from the revised blue-print record sheets furnished the board in 1913. 9. For the year 1902 the figures given in Water Supply Paper No. 85, on page 35, are used instead of those given on the later blue prints, for the latter show themselves too small whether comparison is made with the rainfall or with the Roosevelt Reservoir supply for that year. Even the figures used from Paper 85 are probably somewhat too small; and they seem to have been so regarded in that paper, for in long series of dates without gauge heights it was not recorded that the river was dry but that there was no water at the gage, and the measured flows were not totaled to give the year's supply as they are elsewhere when the monthly records are complete. That "no water at the gauge" did not necessarily mean no flow in the river is indicated on blue-print record of July 2, 1905, where under like conditions of "dry at the gauge" there was a flow some distance away which was estimated to be 40 cubic feet per second. 10. It may be noted that the principal cause of the difference between the earlier record in Water Supply Paper No. 85 and the present revised one on the blue-print sheet, results from having changed a discharge measurement, taken August 26, after three days of flood, while the gauge still read 3.4 feet, from 3,118 acre-feet to 1,350 acre feet, which change had a marked effect on the rating curve for that part of the year. Later in the year, December 14, as much as 2,414 acre-feet was used in the revised blue-print record as corresponding to the same gauge height of 3.4 feet. 11. It is stated in one or more of the water-supply papers that the gaugings at San Carlos are subject to inaccuracies on account of width and shallowness and shifting of stream. It is also stated that the bed where the river was gauged is shifting quicksand. It is known that such river bed may be much lowered temporarily while a flood is passing. All of which makes the first used figure of 3,118, though apparently out of harmony with its neighbors, as trustworthy as the latter substituted lower value of 1,350. Furthermore, a like rise in 1913 was estimated to flow at least 3,000 second-feet. 12. All of the year 1903 and all of 1904 are taken from the revised blue-print record sheets furnished the board in 1913. 13. In the great flood year of 1905 the San Carlos gauge was destroyed two or more times, with the result that none of the big floods were recorded there. But gaugings were made the same year on the upper Gila at Cliff, and on its tributary, the San Francisco, at Alma. 14. Selecting so much of the year's records as are common to the three stations, namely: From May 22 to November 26, 1905, as given for San Carlos on the blue-print sheet furnished the board in 1913, and for Cliff and Alma in Water Supply Paper No. 175 on pages 161-162 and 169-170, respectively, the following comparative table results: 15. Here the Cliff total is 24 times that of Alma; and the San Carlos total is a third greater than the sum of Cliff and Alma. It will be observed in the table that in times of low water San Carlos shows loss, due probably to irrigation and evaporation above; but that in higher stages of the river, which can not yet be called floods, the excess at San Carlos is more than one-third part. It is noteworthy that during the period tabulated above nothing that can be called a large flood occurred at any of the three stations-the greatest flow at San Carlos being 3,240, at Cliff 985, and at Alma 1,575 second-feet. The above period stopped on the eve of a real flood, which reached 13,640 second-feet at Cliff on November 27, and was not measured at Alma, and which destroyed the San Carlos gauge on November 28. Between San Carlos below and Cliff and Alma above, there is a large area drained in to the Gila by the San Simon River and numerous washes, and by the San Carlos River, which is an important tributary. The above considerations alone would seem to warrant increasing the combined discharge at Cliff and Alma as much as 33 per cent to give corresponding discharge at San Carlos. The comparison for 1911 and 1912 between Kelvin, Guthrie, and Clifton, as given in paragraph 4 of this paper, indicates that an increase of 10 per cent or a little more, should be applied to Guthrie and Clifton to represent San Carlos, and it would be expected that a considerable larger increase would need to be applied to Cliff and Alma, which are farther up the rivers. Such expectation is verified to a degree by synchronous readings at Alma and Clifton in 1910, hereinafter given in paragraph 26. 16. In view of all the above it seems certain that the Cliff and Alma discharges should be increased as much as 20 per cent, and very probably considerably more, to give the water supply at San Carlos. Hence 20 per cent increase is herein used for 1905, 1906, 1907, and 1909, with every confidence that the results will be conservative. 17. Proceeding now to greater detail. Water Supply Paper No. 175, on pages 169 and 170, gives for Alma 263,309 acre-feet discharge for 1905, excluding the time from November 26 to December 3, during which a large flood occurred as stated on page 3 of this paper. The same Water Supply Paper, No. 175, on pages 161-2, gives for Cliff 190,090 acre-feet for the period extending only from May 22 to the close of the year. But the comparisons shown in table on page 3 of this paper indicate that the flow at Cliff may be 2 times that at Alma; and a still better comparison is furnished by the following year, 1906, for which the records are complete as given in Water Supply Paper No. 211, giving on page 128 Alma discharge of 111,812 acre-feet, and on page 123 Cliff discharge of 243,760 acre-feet, which is 23 times as large. 18. Accepting the smaller of the two factors, namely, 24 instead of 24, the missing months of Cliff, 1905, record are so supplied as to make Cliff's total for the year two and one-fifth times Alma's total, or 579,280 acre-feet. This result is obtained by making Cliff: January, 30,000; February, 75,000; March, 140,000; April, 125,000; and May, 29,410. It should be remarked in passing that these monthly amounts, though bringing the year up to two and one-fifth times Alma, fall considerably short of being two and one-fifth times the corresponding monthly amounts of Alma, being in fact only about one and three-fourths times as much. But it is more conservative to apply the ratio of 2 to the year rather than to obtain a larger amount by applying it to the missing months only. 19. The results now are as follows: Alma, excluding flood between Nov. 26 and Dec. 3.. Total...... Increase, 20 per cent. Estimate for San Carlos for 1905....... 263, 309 579, 280 842, 589 168, 518 1, 011, 107 20. The flood, excluded above, not measured at Alma, amounted to 79,868 acre-feet at Cliff. Were this divided by 24 to give estimate of same flood at Alma and this increased 20 per cent, the 1,011,107 adopted for San Carlos would have been 43,560 higher. But to apply a ratio to a time so short-only six days-in which a flood can be on one branch of a river and not on the other, would not be justified, especially when conservatism is the guide. 21. For 1906 the records are, as stated above, complete for both Cliff and Alma, as given in Water-Supply Paper No. 211, on pages 123 and 128, and since there are no better data for San Carlos, or in fact any at all, the totals of these two stations are simply increased 20 per cent. 22. For 1907 Water-Supply Paper No. 249 gives on page 180 a discharge for Alma of 147,000 acre-feet, excluding the last 12 days of January. Supplying this omission by using the discharge measurements of January 22 and January 24, as given on page 178, and using gauge heights for January 27 to January 31, given on page 179, and interpolating for January 20, 21, 23, 25, and 26, the January flow and the total are each increased 15,900 acre-feet, making the latter 162,900 acre-feet. 23. The 1907 discharge for Cliff is not computed in the record at hand, but on page 176 are gauge readings for the whole year and a few discharge measurements. Using these measurements for a rating curve and the rating of the preceding year for the higher gage heights and from them computing the discharge for each fifth day in every month, or 72 in the year, there results a mean discharge of 512 secondfeet, which is about 50 per cent greater than the 337 second-feet of the preceding year; and the monthly discharges in round numbers become 116,000, 56,000, 38,000, 27,000, 13,000, 9,000, 8,000, 27,000, 35,000, 11,000, 17,000, and 10,000 acre-feet, or 367,000 acre-feet for the year. The ratio of this total to Alma's total, namely, 367,000÷ 162,900-21, compares well with the ratios of 2 for the fractional year 1905 and the 2 for the whole of 1906, all as given on pages 3 and 4 of this paper. Hence the results are accepted with confidence and increased 20 per cent to give the flow at San Carlos. 24. During 1908 there were not sufficient gagings anywhere that would indicate the probable flow at San Carlos, the only known gaging being at Redrock from November 16 to December 31; but it being desirable to have figures to show the most probable water supply and to complete certain diagrams, they are supplied from the known rainfall over the drainage area, using the proportion given on page 2 of this paper. 25. For 1909 records are complete for Redrock, which is some miles below Cliff; also at Alma, all as given on pages 221 and 224 of WaterSupply Paper No. 269. The Redrock total, 153,520, and the Alma total, 74,300, give a ratio of about 2.1, which compares fairly well with the Cliff-Alma ratios of 21, 24, and 23, as given above. Hence, as with Cliff and Alma amounts, these amounts for Redrock and Alma are increased 20 per cent to represent San Carlos. 26. For 1910, Water Supply Paper No. 289, page 202, gives for Redrock 53,960 acre-feet, and on page 207 it gives 11,865 acre-feet for Alma. But the Alma record seems erroneous in gauge height or gauge reading or in rating. Only four times during the year was the discharge measured-one time giving a discharge of 28 second-feet for a gauge height of 0.96 feet, and another giving 7.8 second-feet for a gauge height of 1.95 feet, or a foot higher than before. That the Alma discharge for 1910 must be erroneous is further evidenced by synchronous observations farther down the river at Clifton from October 23 to December 31, of which the gauge readings and discharge measurements are recorded on page 208 of Paper 289. The discharge at |