CHAPTER XV.

=155. Waste of Water, Particularly in the City of New York.=—The quantity of water involved in designing a water-supply for cities and towns is much larger than that which is actually needed. The experience of civil engineers in many cities, both in this country and in Europe, shows conclusively that the portion of water actually wasted or running away without serving any purpose will usually run from 30 to 50 per cent of the total amount brought to the distributing system. In the city of New York there is strong reason to believe that the wastage is not less than two thirds of the total quantity supplied. It is frequently assumed that both the quantities supplied and the quantities uselessly wasted in New York are larger than in other places. As a matter of fact those quantities are actually smaller than in some other large cities. While the supply per inhabitant in New York City is much larger than should be required, the use of water by its citizens is not extravagant when gauged by the criterion of use in other large cities. This question was most carefully and exhaustively investigated in 1899 and the early part of 1900 by Mr. John R. Freeman of Boston, acting for the comptroller of the city of New York.

The usual wastes of a water-supply system may be distributed under six principal heads. First, leaky house-plumbing; second, and “possibly first in order of magnitude,” leaky service-pipes connecting the house pipe system with street-mains; third, leaving water-cocks open unnecessarily; fourth, leaky joints in street-mains or pipes; fifth, possibly pervious beds and banks of distributing-reservoirs; sixth, stealing or “unlawful diversion” of water through surreptitious connections.

The sixth item is probably an extremely small one in New York, although instances of that kind of waste have been found. It is an old wastage known as far back in time as the ancient Roman water-supply. The second and third items probably constitute the bulk of the wastage in this city.

=156. Division of Daily Consumption in the City of New York.=—In the course of his search for the various sources of consumption, Mr. Freeman concluded from his examinations and from the use of the various means placed at his command for measuring the daily consumption between December 2nd and December 5th, 1899, and December 8th and December 15th, 1899, that the average daily consumption could be divided as follows:

Gallons per Inhabitant per Day.

Probable average amount really used 40 Assumed incurable waste 10} Curable waste, probably 65} 75 ---- Daily uniform rate of delivery by Croton Aqueduct 115

In his investigations Mr. Freeman had the elevation of water in the Central Park reservoir carefully observed every six minutes throughout the twenty-four hours. At the same time the uniform flow through the new Croton Aqueduct was known as accurately as the flow through such a conduit can be gauged at the present time. Knowing, therefore, the concurrent variation of volume in the Central Park reservoir supplied by the new Croton Aqueduct and the rate of flow in that aqueduct, the consumption of water per twenty-four hours would be known with the same degree of accuracy with which the flow in the aqueduct is measured. It was found by these means that the actual consumption between the hours of 2 and 4 A.M. was at the rate of 94 gallons per inhabitant per day, although the actual use at that time was as near zero as it is possible to approach during the whole twenty-four hours. Nearly all of that rate of consumption represents waste.

Summing up the whole matter in the light of his investigations, Mr. Freeman made the following as his nearest estimate to the actual consumption of the daily supply of water of New York City:

Gallons per Inhabitant ACTUAL USE: per Day. Domestic (average) 12 - 20 Manufacturing and commercial 20 - 30 City buildings, etc. 2 - 4 Fires, street flushing and sprinkling 0.4 - 0.7 ----------- Total 34 - 55

INCURABLE WASTE (probabilities): Leaks in mains 1 - 2 Leaks in old and abandoned service-pipes 1 - 2 Poor plumbing, all taps metered and closely inspected 2 - 3 Careless and wilful wastes 1 - 2 Under-registry of meters 1 - 1 ------- Total incurable waste and under-registry 6 -10 ======== Minimum use and waste 40 -65

NEEDLESS WASTE: Leaks in street-mains (a guess) 15- 10 Leaks in service-pipes between houses and street-mains 15- 10 (a guess) Defective plumbing (a guess) 25- 15 Careless and wilful opening of cocks (a guess) 17- 14 To prevent freezing in winter and for cooling in summer 3 - 1 -------- Total needless waste 75- 50 ======== Total consumption 115-115

=157. Daily Domestic Consumption.=—The quantity assigned in the preceding statement to domestic use is confirmed by the abundant experience in other cities where services are carefully metered, as in Fall River, Lawrence, and Worcester, Mass., and in Woonsocket, R. I., where measurements by meters show that the domestic consumption has varied from 11.2 to 16.3 gallons per inhabitant per day. Furthermore, annual reports of the former Department of Public Works and the present Department of Water-supply for the City of New York show that during the years 1890 to 1898 such meters as have been used in the territory supplied by the Croton and the Bronx aqueducts indicate a daily consumption varying from 13.8 to 24.2 gallons per inhabitant per day. The same character of confirmatory evidence can be applied to the quantities assigned to manufacturing and commercial uses, city buildings and fires, street flushing and sprinkling.

=158. Incurable and Curable Wastes.=—The items composing incurable waste, unfortunately, cannot be so definitely treated. It is perfectly well known, however, among civil engineers, that a large amount of leakage takes place from corporation cocks, which are those inserted in the street-mains to form the connection between the latter and the house service-pipes. Again, many of these service-pipes are abandoned and insufficiently closed, or not closed at all, leaving constantly running streams whose continuous subsurface discharges escape detection and frequently find their way into sewers. Water-pipes which have been laid many years frequently become so deeply corroded as to afford many leaks and sometimes cracks. Doubtless there are many portions of a great distributing system, like that in New York City, which need replacing and afford many large leaks, but undiscoverable from the surface. Many lead joints of street-mains also become leaky with age, while others are leaky when first laid in spite of inspection during construction. Just how much these items of waste would aggregate it is impossible accurately to state, but from careful observations made in other places 5 to 10 gallons per day per head of population seems reasonable. A three-year-old cast-iron fire-protection pipe 5.57 miles long and mainly 16 inches in diameter, under an average pressure of 114 pounds per square inch, was tested in Providence in 1900 and showed a leakage at lead joints only of 446 gallons per mile per twenty-four hours, which was equivalent to .22 gallon per foot of lead joint per twenty-four hours. Further tests in 1900 of seven lines of new pipe laid by the Metropolitan Board of Boston, and tested under pressures varying from 50 to 150 pounds per square inch by Mr. F. P. Stearns, chief engineer, and Mr. Dexter Brackett, engineer of Distribution Department, and having an aggregate length of 51.4 miles with diameters ranging from 16 to 48 inches, gave an average leakage per lineal foot of pipe in gallons per twenty-four hours ranging from .6 to 3.7 gallons (average 2.47 gallons), equivalent to an average leakage of 3 gallons per twenty-four hours per lineal foot of lead joint. The possible rates of leakage from street-mains are to be applied to a total length of pipe-lines of 833 miles for the boroughs of Manhattan and the Bronx. The borough of Brooklyn has somewhat over 600 miles of street-mains, but they are not to be considered in connection with the Croton and Bronx water-supply.

All these considerations either confirm or make reasonable the estimates of the various items of actual use and waste set forth by Mr. Freeman.

=159. Needless and Incurable Waste in City of New York.=—Concisely summing up his conclusions, it may be stated that in the year 1899 the average consumption per inhabitant of the boroughs of Manhattan and the Bronx was 115 gallons; of these 115 gallons the needless average waste may be 65 gallons, while the incurable or necessary waste may probably be taken at 10 gallons per inhabitant per day. It is further probable that the total underground leakage in New York City is to be placed somewhere between 20 to 35 gallons per inhabitant per day.

=160. Increase in Population.=—The total volume of daily supply to any community is determined by the population; but the population is as a whole constantly increasing. It becomes necessary, therefore, to estimate the capacity of a water-supply system in view of the future population of the city to be supplied. No definite rule can be set as to the future period for which the capacity of any desired system is to be estimated. It may be stated that no shorter period of time than probably ten years should be considered, indeed it is frequently prudent to provide for a period of not less than twenty years, and it may sometimes be necessary or advisable to consider a possible source of supply for even fifty years. Provision must be made not only for the present population, but for the increase during those periods of time, or at least for the possible development that may be needed.

The increase in population of cities will obviously vary for different locations with the character of the occupations followed and with the development of such important factors of industrial life as railroad connections, facilities for marine commerce, the capacity for development of the surrounding country, and other influences which aid in the increase of commerce and industrial activity and the growth of population. It has been observed, as a matter of experience, that large cities generally reach a point where their subsequent increase of population is represented by a practically constant percentage, the value of that percentage depending upon local considerations. In 1893, when it was desired to estimate the future population of London for as much as forty years, it was found that the increase for the ten years from 1881 to 1891 was 18.2 per cent, with an average of about 20 per cent for several previous decades. It could, therefore, be reasonably estimated for the city of London that its population at the end of any ten-year period would be 18.2 per cent greater than its population at the beginning of that period. In Appendix 1 of the report of the Massachusetts State Board of Health upon the Metropolitan water-supply for the city of Boston made in 1895, the increases for the two ten-year periods 1870-1880 and 1880-1890 were 6 per cent and 9.6 per cent respectively for the city proper, but for the population within a ten-mile radius from the centre of the city they were 28.7 per cent and 33.7 per cent respectively. The corresponding percentages for the cities of New York, Philadelphia, and Chicago for the same periods are as shown in the following tabular statement:

------------+-----------------------------+------------------------ | Population. |Percentages of Increase. ------------+---------+---------+---------+---------+-------------- | 1870. | 1880. | 1890. |1870-80. | 1880-90. ------------+---------+---------+---------+---------+-------------- New York |1,626,119|2,131,051|2,821,802| 31 | 32 Philadelphia| 726,247| 921,458|1,162,577| 27 | 26 Chicago | 310,996| 550,618|1,075,158| 77 | 95[6] ------------+---------+---------+---------+---------+--------------

[6] Includes added territory.

Obviously every estimate of this kind must be made upon the merits of the case under consideration. The probable increase of population for any particular city is sometimes estimated by considering the circumstances of growth of some other city of practically the same size, and if possible with the same commercial industries or residential environment, or making suitable allowances for variations in these respects. Since it is imperative to secure as accurate estimates as practicable, both methods or other suitable methods should be employed, in order that the results may be confirmed or modified by comparison. In every case the supply system should be designed to meet reasonable estimated requirements for the longest practicable future period, preferably not less than twenty years.

=161. Sources of Public Water-supplies.=—One of the most important features of a proposed water-supply is its source, since not only the potable qualities are largely affected by it, but frequently the amount also. The two general classes into which potable waters are divided in respect to their sources are surface-waters and ground-waters. Surface-waters include rain-water collected as it falls, water from rivers or smaller streams, and water from natural lakes; they are collected in reservoirs and lakes or impounding reservoirs. Ground-waters are those collected from springs, from ordinary or shallow wells, from deep or artesian wells, and from horizontal galleries, like those sometimes constructed near and parallel to subsurface streams or in subsurface bodies of water, affording opportunity for filtration from the latter through sand or other open materials to them.

The quality of water will obviously be affected by the kind of material through which it percolates or flows. Surface-waters, flowing over the surface of the ground or percolating but a short distance below the surface, naturally have contact with vegetable matter, unless they are collected in a country where the soil is sandy and where the vegetation is scarce. If such waters flow through swamps or over beds of peat or other similar vegetable mould or soil, they may become so impregnated with organic matter or so deeply colored by it that they are not available for potable purposes. Ground-waters, on the other hand, possess the advantage of having flowed through comparatively great depths of sand or other earthy material essentially free of organic matter. They may, however, in some locations, carry prejudicial amounts of objectionable salts in solution, rendering them unfit for use. As a rule, ground-waters are apt to be of better quality than surface-waters, but they do not generally stand storage in reservoirs as well as surface potable waters. It is advisable to store them in covered reservoirs from which the light is excluded, rather than in open reservoirs. They are sometimes impregnated with salts of iron to such an extent as to make it necessary to resort to suitable processes for their removal, and they are also occasionally found so hard as to require the employment of methods of softening them.

Both sources of supply are much used in the United States. Table IV shows the percentages of the various classes of supplies as found in this country during the year 1897; the total number of supplies having been at that time nearly 4000.

TABLE IV.

WATER-SUPPLIES OF THE UNITED STATES.

Source. Per Cent of Total.

{Rivers 25 Surface-waters: {Lakes 7 {Impounding reservoirs 6 {Combinations .5 ----- 38.5

{Shallow wells 26 {Artesian wells 10 Ground-waters: {Spring 15 {Galleries and tunnels 1 {Combinations 2 ----- 54

Surface- and {Rivers and ground-waters 6 ground-waters: {Lakes and ground-waters 1 {Impounding reservoirs and ground-waters .5 ---- 7.5 ------- Total 100.0

It will be observed that a little more than one half of the supplies are from ground-waters. The practice in connection with European public water-supplies is different in that a considerably larger percentage of the total is taken from ground-waters.

The original source of essentially all the water available for public water-supplies is the rainfall. It becomes of the greatest necessity, therefore, to secure all possible information regarding rainfall wherever it may be necessary to construct a public water-supply. Civil engineers and other observers have for many years maintained continuous records of rainfall observations at various points throughout the country, but it is within only a comparatively short time that the number of those points has been large. Through the extension of the work of the Weather Bureau, points of rainfall observation are now scattered quite generally throughout all States of the Union. The oldest observations are naturally found in connection with stations located in the Eastern States, where the rainfall is more uniformly distributed than in many other portions of the country. Obviously rainfall records become of the greatest importance in those localities like the semiarid regions of the far West where long periods of no rain occur.

=162. Rain-gauges and their Records.=—The instrument used for the collection of rain in order to determine the amount falling in a given time is the rain-gauge, which may be fitted with such appliances as to give a continuous record of the rate of rainfall. It has been found that the location of the rain-gauge has a very important influence upon the amount of rain which it collects. It should be placed where wind currents around high structures in its vicinity cannot affect its record. The top of a large flat-roofed building is a good location in a city, although the elevation above the surface of the ground, as is well known, affects the quantity of water collected by the gauge. The collection will be greater at a low elevation than at a high one, in consequence of the greater wind currents at the higher point, it being well known that less rain will be collected where there is the most wind, other things being equal.

[Illustration: Ordinary Rain-gauge.]

=163. Elements of Annual and Monthly Rainfall.=—In consequence of the great variations in the rate of rainfall, not only for different portions of the country, but at different times during the same storm, it becomes necessary to determine various quantities such as the maximum, minimum, and mean annual rainfall, the actual monthly rainfall for different months of the year, and the maximum and minimum monthly rainfall for as long a period as possible. The minimum monthly rainfall and the minimum annual rainfall are of special importance in connection with public water-supply and water-power questions, since those minima will, in connection with the area of a given watershed, determine the greatest amount of water which can be made available for use. In entering upon the consideration of such questions, therefore, civil engineers must inform themselves with the greatest detail as to the characteristics of the monthly and the annual rainfall of the locality in which their works are to be located.

[Illustration: MONTHLY VARIATIONS IN RAINFALL.

FIG. 1.]

TABLE V.

GENERAL RAINFALL STATISTICS FOR THE UNITED STATES.

----------------------------+---------+---------+--------- | | Per Cent| | | of | Per Cent | Mean | Summer | Driest Station. | Yearly | and | Year to |Rainfall.| Autumn | Mean | | Rain to| Year. | | Mean | | | Yearly. | ----------------------------+---------+---------+--------- North Atlantic: | | | Boston | 45.4 | 50 | 60 New York | 44.7 | 52 | 62 Philadelphia | 42.3 | 52 | 70 Washington | 42.9 | 51 | 69 South Atlantic: | | | Wilmington | 53.7 | 61 | 75 Charleston | 49.1 | 61 | 48 Augusta | 48.0 | 50 | 81 Jacksonville | 54.1 | 65 | 74 Key West | 38.2 | 70 | 54 Gulf and Lower Mississippi: | | | Montgomery | 52.5 | 42 | 76 Mobile | 62.6 | 51 | 68 New Orleans | 60.3 | 52 | 64 Galveston | 47.7 | 58 | 50 Nashville | 50.2 | 46 | 67 Ohio Valley: | | | Pittsburg | 36.6 | 53 | 70 Cincinnati | 42.1 | 50 | 60 Indianapolis | 42.2 | 51 | 59 Cairo | 42.6 | 47 | 62 Louisville | 47.2 | 48 | 74 Lake Region: | | | Detroit | 32.5 | 56 | 65 Cleveland | 36.6 | 54 | 71 Duluth | 30.7 | 63 | 65 Upper Mississippi Valley: | | | St. Louis | 40.8 | 52 | 55 Chicago | 34.0 | 54 | 66 Milwaukee | 31.0 | 55 | 66 Madison | 33.2 | 58 | 39 The Plains: | | | Omaha | 31.4 | 63 | 57 North Platte | 18.1 | 61 | 56 Denver | 14.3 | 48 | 59 Cheyenne | 12.7 | 55 | 39 The Plateau: | | | Tucson | 11.7 | 65 | 44 Santa Fé | 14.6 | 69 | 53 Salt Lake City | 18.8 | 39 | 55 Walla Walla | 15.4 | 38 | 46 Pacific Coast: | | | Astoria | 77.0 | 33 | 64 Portland | 46.2 | 31 | 67 Sacramento | 19.9 | 16 | 42 San Francisco | 23.4 | 17 | 51 Los Angeles | 17.2 | 15 | 33 San Diego | 9.7 | 18 | 30 ----------------------------+---------+---------+--------- | | | | | | | Per Cent| Per Cent| Number Station. | Two | Three | of | Driest | Driest | Years’ | Years. | Years. | Record. | | | | | | ----------------------------+---------+---------+--------- North Atlantic: | | | Boston | 70 | 80 | 79 New York | 77 | 80 | 61 Philadelphia | 75 | 80 | 72 Washington | 71 | 74 | 45 South Atlantic: | | | Wilmington | 80 | 81 | 26 Charleston | 55 | 62 | 89 Augusta | 88 | 87 | 27 Jacksonville | 77 | 83 | 27 Key West | 61 | 73 | 49 Gulf and Lower Mississippi: | | | Montgomery | 80 | 83 | 24 Mobile | 75 | 78 | 26 New Orleans | 75 | 77 | 26 Galveston | 65 | 72 | 26 Nashville | 73 | 83 | 32 Ohio Valley: | | | Pittsburg | 78 | 85 | 54 Cincinnati | 72 | 71 | 62 Indianapolis | 76 | 82 | 27 Cairo | 75 | 81 | 25 Louisville | 81 | 85 | 25 Lake Region: | | | Detroit | 72 | 79 | 46 Cleveland | 74 | 81 | 41 Duluth | 81 | 88 | 26 Upper Mississippi Valley: | | | St. Louis | 65 | 75 | 60 Chicago | 80 | 86 | 30 Milwaukee | 74 | 73 | 53 Madison | 58 | 68 | 28 The Plains: | | | Omaha | 63 | 70 | 27 North Platte | 67 | 72 | 22 Denver | 71 | 77 | 27 Cheyenne | 62 | 75 | 27 The Plateau: | | | Tucson | 79 | 80 | 19 Santa Fé | 63 | 66 | 37 Salt Lake City | 64 | 74 | 29 Walla Walla | 81 | 86 | 27 Pacific Coast: | | | Astoria | 68 | 77 | 34 Portland | 76 | 79 | 27 Sacramento | 67 | 84 | 47 San Francisco | 73 | 78 | 47 Los Angeles | 48 | 59 | 24 San Diego | 54 | 61 | 47 ----------------------------+---------+---------+---------

The diagram Fig. 1 and Table V are constructed from data given in the bulletins of the Weather Bureau and exhibit some of the general features of the rainfall for different points throughout this country. The heavy lines of the diagram show the average monthly precipitation at the points indicated, for periods of a considerable number of years, as shown in the table. It will be observed that the rainfall is comparatively uniform in the North Atlantic States but quite variable on the Pacific coast, as well as in the Mississippi and Missouri valleys and west of those valleys.

=164. Hourly or Less Rates of Rainfall.=—Although not often of great importance in connection with public water-supply systems, it is sometimes necessary to possess data regarding maximum hourly (or less) rates of precipitation in connection with sewer or drainage work. The earlier records give exaggerated reports of maximum rates of rainfall, although that rate varies rapidly with the time. Throughout a rain-storm the rate of precipitation is constantly varying and the maximum rate seldom if ever extends over a period equal to a half-hour; usually it lasts but a few minutes only. In this country an average rate of 1 inch per hour, extending throughout one hour, is phenomenal, although that rare amount is sometimes exceeded. A maximum rate of about 4 inches per hour, lasting 15 to 30 minutes, is, roughly speaking, about as high as any precipitation of which we have reliable records. The waste ways or other provisions for the discharge of surplus or flood-waters of the Metropolitan Water-supply of Boston are designed to afford relief for a total precipitation of 6 inches in twenty-four hours. It is safe to state that an excess of that accommodation will probably never be required.

=165. Extent of Heavy Rain-storms.=—In all engineering questions necessitating the consideration of these great rain-storms it is necessary to remember that their extent is frequently much greater than the areas of watersheds usually contemplated in connection with water-supply work. The late Mr. James B. Francis found in the great storm of October, 1869, which had its maximum intensity in Connecticut, that the area over which 6 inches or more of rain fell exceeded 24,000 square miles, and that the area over which a depth of 10 inches or more fell was 519 miles. Again, in the New England storm of February, 1886, 6 inches or more of rain fell over an area of at least 3000 square miles. Storm records show that as much as 8 or 10 inches in depth have fallen over areas ranging from 1800 to 500 square miles, respectively, in a single storm.

=166. Provision for Low Rainfall Years.=—The capacity of any public water-supply must evidently be sufficient to meet not only the general exigencies of the year of lowest rainfall, but also the conditions resulting from the driest periods of that year. It is customary among civil engineers to consider months as the smaller units of a dry year. It is necessary, therefore, to examine not only the annual rainfalls but the monthly rates of precipitation during critical years, i.e., usually during dry years.

It is impossible to determine absolutely the year of least rainfall which may be expected, but evidently the longer the period over which observations have extended the nearer that end will be attained. It is sometimes assumed that the lowest annual rainfall likely to be expected in a long period of years is 80 per cent of the average annual rainfall for the same period. Or, it is sometimes assumed that the average rainfall for the lowest two or three consecutive years will be 80 per cent of the average for the entire period, and that the year of minimum rainfall may be expected to yield two thirds of the annual average precipitation. Such features will necessarily vary with the location of the district considered. Conclusions which may be true for the New England or northern Atlantic States probably will not hold for the south Atlantic and Gulf States. Data for such conclusions must be obtained from the rainfall of the locality considered. Table VI exhibits the comparative monthly rainfall which J. T. Fanning suggests may be used approximately for the average Atlantic coast districts.

TABLE VI.

+-----------+-----------+------------+-----------+ | | Mean | | Probable | | | Monthly | Respective | Depth in | | | Rainfall, | Ratios. | Inches of | | | Inches. | | Actual | | | | | Rainfall. | +-----------+-----------+------------+-----------+ | January | 4 × 1.65 = 6.6 | | February | 4 × 1.50 = 6.0 | | March | 4 × 1.65 = 6.6 | | April | 4 × 1.45 = 5.8 | | May | 4 × .85 = 3.4 | | June | 4 × .75 = 3.0 | | July | 4 × .35 = 1.4 | | August | 4 × .25 = 1.0 | | September | 4 × .30 = 1.2 | | October | 4 × .45 = 1.8 | | November | 4 × 1.20 = 4.8 | | December | 4 × 1.60 = 6.4 | +-----------+------------------------------------+

If the average monthly rainfall throughout the year were one inch, the values of the ratios would show the actual monthly precipitation. In general the table would be used by dividing the total yearly rainfall by 12, and then multiplying that monthly average by the proper ratio taken from the table opposite the month required. Such tables should only be used for approximate purposes and when actual rainfall records are not available for the district considered.

=167. Available Portion of Rainfall or Run-off of Watersheds.=—If the public water-supply is to be drawn from a stream where the desired rainfall records exist, it is necessary to know what portion of the rainfall, either in the driest or in other years, may be available. This is one of the departments of the hydraulics of streams for which much data yet remain to be secured. The watersheds or areas drained by some streams, like the Sudbury River of the Boston, and the Croton of the New York water-supply, have, however, been studied with sufficient care to give reliable data. The amount of water flowing in a stream from any watershed for a given period, as a year, is called the annual “run-off” of the watershed, and it is usually expressed as a certain percentage of the total rainfall on the area drained. For certain purposes it is sometimes more convenient to express the run-off from the watershed as the number of cubic feet of water per second per square mile of area. Table VII, taken from Turneaure and Russell, exhibits run-off data for a considerable number of streams in connection with both average and minimum rainfalls.

TABLE VII.

STATISTICS OF THE FLOW OF STREAMS.

-----------------+---------+-------+----------------+----------------- | Area | | Average Yearly.|Year of Minimum Stream. | Drained,| | | Flow. | Square | Years.+-----+-----+----+-----+------+---- | Miles. | |Rain,|Flow,|Per |Rain,|Flow,|Per | | |Inch.|Inch.|Cent|Inch.|Inch.|Cent -----------------+---------+-------+-----+-----+----+-----+-----+---- Sudbury | 75.2 |1875-97|45.77|22.22|48.6|32.78|11.19|34.1 Cochituate | 18.87|1863-96|47.08|20.33|43.2|31.20| 9.76|31.3 Mystic | 26.9 |1878-96|43.79|19.96|45.6|31.22| 9.32|29.8 Connecticut |10,234 |1871-85|44.69|25.25|56.5|40.02|18.25|45.6 Croton | 338 |1870-94|48.38|24.57|50.8|38.52|14.54|37.8 Upper Hudson | 4,500 |1888-96|39.70|23.36|59.0|33.49|17.46|52.2 Genesee | 1,060 |1894-96|39.82|12.95|32.5|31.00| 6.67|21.5 Passaic | 822 |1877-93|47.08|25.44|54.0|35.64|15.23|42.7 Upper Mississippi| 3,265 |1885-99|26.57|4.90 |18.4|22.86| 1.62| 7.1 =================+=========+=======+=====+=====+====+=====+=====+==== | | | Average for |Average for June | Area | |December to May.| to November. Stream. | Drained,| | | | Square | Years.+-----+-----+----+-----+-----+---- | Miles. | |Rain,|Flow,|Per |Rain,|Flow,| Per | | |Inch.|Inch.|Cent|Inch.|Inch.|Cent -----------------+---------+-------+-----+-----+----+-----+-----+---- Sudbury | 75.2 |1875-97|22.98|17.52|76.0|22.61| 4.70|20.8 Cochituate | 18.87|1863-96|22.97|14.87|64.7|24.10| 5.46|22.7 Mystic | 26.9 |1878-96|22.11|15.12|68.4|21.66| 4.84|22.4 Connecticut |10,234 |1871-85|20.13|17.95|89.1|24.56| 7.30|29.7 Croton | 338 |1870-94|23.39|17.81|76.1|24.99| 6.76|27.0 Upper Hudson | 4,500 |1888-96|18.20|16.23|89.0|21.50| 7.13|33.0 Genesee | 1,060 |1894-96|19.58|10.20|52.2|20.24| 2.75|13.6 Passaic | 822 |1877-93|22.47|18.22|81.1|24.39| 7.19|29.5 -----------------+---------+-------+-----+-----+----+-----+-----+----

The information to be drawn from this table is sufficient to give clear and general relations between the recorded precipitation and run-off. The percentage of run-off is seen to vary quite widely, but as a rule it is materially less for the year of minimum flow than for the average year. That feature of the table is an expression of the general law, other things being equal, that the smaller the precipitation the less will be the percentage of run-off. A number of influences act to produce that result. During a year of great precipitation the earth is more nearly saturated the greater part of the time, and hence when rain falls less of it will percolate into the ground and more of it will run off. Again, if the ground is absolutely dry, a certain amount of rain would have to fall before any run-off would take place. The area and shape of a watershed will also affect to some extent the flow of the stream which drains it. A larger run-off would reasonably be expected from a long narrow watershed than from one more nearly circular in outline. The greater the massing of the watershed, so to speak, the more opportunity there is for the water to be held by the ground and the less would be the run-off.

TABLE VIII.

AVERAGE YIELD OF SUDBURY WATERSHED, 1875-1899, INCLUSIVE, VARIOUSLY EXPRESSED.

(Area of watershed, 75.2 square miles.) ---------+--------------------+------------------------------- | Per Square Mile. | Rainfall. Month. +--------------------+----------+----------+--------- |Cubic Feet |Million |Collected,|Per Cent | Total, |per Second.|Gallons |Inches. |Collected.| Inches. | |per Day.| | | ---------+-----------+--------+----------+----------+--------- January | 1.937| 1.252| 2.233| 51.6| 4.33 February | 2.904| 1.877| 3.050| 71.7| 4.26 March | 4.489| 2.901| 5.175| 117.4| 4.41 April | 3.124| 2.019| 3.485| 107.5| 3.24 May | 1.680| 1.086| 1.936| 58.1| 3.33 June | .735| .475| .821| 28.0| 2.93 July | .305| .197| .352| 9.3| 3.77 August | .478| .309| .551| 13.3| 4.16 September| .376| .243| .419| 13.0| 3.23 October | .829| .536| .956| 21.9| 4.37 November | 1.474| .953| 1.645| 39.0| 4.22 December | 1.612| 1.042| 1.859| 51.9| 3.58 +-----------+--------+----------+----------+-------- Year | 1.655| 1.070| 22.482| 49.1| 45.83 ---------+-----------+--------+----------+----------+--------

=168. Run-off of Sudbury Watershed.=—Table VIII has been given by Mr. Charles W. Sherman, as representing the average yield of the Sudbury watershed for the period 1875 to 1899, inclusive, expressed in several different ways. The average rainfall was 45.83 inches, and the percentage which represents the run-off is 49.1 per cent of the total. The average monthly run-off varies from .305 cubic foot (for July) to 4.489 cubic feet (for March) per second per square mile. As a general rule it may be stated that the average run-off from the drainage areas of New England streams amounts very closely to 1,000,000 gallons per square mile per day. The area of the Sudbury watershed is 75.2 square miles, with 6.5 per cent of that total area occupied by the surface of lakes or reservoirs. As will presently be seen, the amount of exposed water surface in any watershed has an appreciable influence upon its run-off.

=169. Run-off of Croton Watershed.=—The total area of the Croton watershed, from which New York City draws its supply, i.e., the area up-stream from the new Croton Dam, is 360.4 square miles, of which 16.1 square miles, or 4.47 per cent, of its total area is water surface. Mr. John R. Freeman found in the investigations covered by his report to the comptroller of the city of New York in 1900 that the average annual rainfall on that area for the thirty-two years beginning 1868 and ending 1899 was 48.07 inches, and that the average run-off for the same period was 47.7 per cent of the total average rainfall, equivalent to a depth of 22.93 inches.

[Illustration: Aqueducts near Jerome Park Reservoir, New York City.]

Table IX gives the main elements of the rainfall and run-off for the Croton watershed during the thirty-two year period, for the averages just given.

TABLE IX.

RAINFALL ON CROTON WATERSHED IN TOTAL INCHES—1868-1898. NATURAL FLOW OF CROTON RIVER AT OLD CROTON DAM, IN EQUIVALENT INCHES. PERCENTAGE OF RUN-OFF TO RAINFALL FOR EACH YEAR.

-----------+---------+---------+----------- Year. | Total | Total | |Rainfall.| Run-off.| Per Cent. -----------+---------+---------+----------- 1868 | 50.33 | 33.33 | 66.22 1869 | 48.36 | 23.61 | 48.82 1870 | 44.63 | 19.20 | 43.02 1871 | 48.94 | 19.46 | 39.76 1872 | 40.74 | 16.92 | 41.53 1873 | 43.87 | 25.02 | 57.03 1874 | 42.37 | 25.10 | 59.24 1875 | 43.66 | 24.77 | 56.73 1876 | 40.68 | 21.09 | 51.84 1877 | 48.23 | 20.22 | 41.92 1878 | 55.70 | 27.17 | 48.78 1879 | 47.04 | 19.65 | 41.77 1880 | 36.92 | 12.63 | 34.21 1881 | 46.69 | 19.25 | 41.23 1882 | 52.35 | 24.28 | 46.38 1883 | 42.70 | 13.33 | 31.22 1884 | 51.28 | 24.08 | 46.96 1885 | 43.67 | 17.71 | 40.55 1886 | 47.74 | 20.10 | 42.10 1887 | 57.29 | 26.61 | 46.45 1888 | 60.69 | 35.27 | 58.12 1889 | 55.70 | 31.39 | 56.36 1890 | 54.05 | 25.95 | 48.01 1891 | 47.20 | 23.48 | 49.75 1892 | 44.28 | 17.68 | 39.93 1893 | 54.87 | 29.05 | 52.94 1894 | 47.33 | 20.56 | 43.44 1895 | 40.58 | 15.95 | 39.31 1896 | 45.85 | 23.26 | 50.73 1897 | 53.12 | 25.59 | 48.17 1898 | 57.40 | 29.72 | 51.77 1899 | 44.67 | 22.28 | 49.88 Average for| 48.07 | 22.93 | 47.70 32 years. | | | -----------+---------+---------+-----------

The table shows that the least annual rainfall was 36.92 inches for 1880, and that the run-off represented a depth of 12.63 inches only, or 34.21 per cent of the total annual precipitation. As a rule the same feature of a low percentage of run-off will be found belonging to the years of low rainfall, although there are many irregularities in the results. On the other hand, the high percentages of run-off are for the years 1868, 1888, and 1889, and they will generally be found belonging to years of relatively great precipitation. A low percentage of run-off will also be lower if the year to which it belongs follows a dry year or a dry cycle of two or three years. Similarly the high percentages of run-off will, as a rule, be higher if they follow years of high precipitation; that is, if they belong to a cycle of relatively great rainfall.

=170. Evaporation from Reservoirs.=—If it is contemplated to build reservoirs on a watershed the capacity of which is being estimated on the basis of either the driest year or the driest two- or three year cycle, it is necessary to make a deduction from the rainfall for the evaporation which will take place from the surface of the proposed reservoir. In order that that deduction may be made as a proper allowance for added water surface in a drainage area, it is necessary that the amount of evaporation be determined for the district considered. The rate of evaporation is dependent upon the area of water surface, upon the wind, and upon the temperature both of the water and air above it. Numerous evaporation observations have been made both in this and other countries, and extensive evaporation tables have been prepared by the Weather Bureau, from which a reasonable estimate of the monthly evaporation for all months in the year may be made for almost any point in the United States. Particularly available observations have been made by Mr. Desmond Fitzgerald of Boston on the Chestnut Hill reservoirs of the Boston Water-supply, and by Mr. Emil Kuichling, engineer of the Rochester Water-works, on the Mount Hope reservoir of the Rochester supply. Table X exhibits the results of the observations of both these civil engineers.

[Illustration: Aqueduct Division Wall of Jerome Park Reservoir, New York City.]

As would be anticipated, the period from May to September, both inclusive, shows by far the greatest evaporation of the whole year, while December, January, and February are the months of least evaporation. The total annual evaporation at Boston was 39.2 inches and 34.54 inches at Rochester.

TABLE X.

MEAN MONTHLY EVAPORATIONS. ----------------+-------------------------+------------------------- | | | Chestnut Hill Reservoir,| Mount Hope Reservoir, Month. | Boston, Mass. | Rochester, N. Y. | | ----------------+------------+------------+------------+------------ | | Per Cent | | Per Cent |Evaporation,| of Yearly |Evaporation,| of Yearly | Inches. |Evaporation.| Inches. |Evaporation. ----------------+------------+------------+------------+------------ January | 0.96 | 2.4 | 0.52 | 1.5 February | 1.05 | 2.7 | 0.54 | 1.6 March | 1.70 | 4.3 | 1.33 | 3.9 April | 2.97 | 7.6 | 2.62 | 7.6 May | 4.46 | 11.4 | 3.93 | 11.4 June | 5.54 | 14.2 | 4.94 | 14.3 July | 5.98 | 15.2 | 5.47 | 15.8 August | 5.50 | 14.0 | 5.30 | 15.4 September | 4.12 | 10.4 | 4.15 | 12.0 October | 3.16 | 8.1 | 3.16 | 9.1 November | 2.25 | 5.7 | 1.45 | 4.2 December | 1.51 | 3.9 | 1.13 | 3.2 +------------+------------+------------+------------ Total for year | 39.20 | | 34.54 | +------------+------------+------------+------------ Mean temperature| 48°.6 | 47°.8 ----------------+-------------------------+-------------------------

A reference to data of the Weather Bureau will show that annual evaporation as high as 100 inches, or even more, may be expected on the plateaux of Arizona and New Mexico. Other portions of the arid country in the western part of the United States will indicate annual evaporations running anywhere from 50 to 90 inches per year, while on the north Pacific coast it will fall as low as 18 to 40 inches.

=171. Evaporation from the Earth’s Surface.=—Data are lacking for anything like a reasonably accurate estimate of evaporation from the earth’s surface. It is well known that the loss of water from that source is considerable in soils like those of swamps, particularly when exposed to the warm sun, but no reliable estimate can be obtained for the exact amount. Nor is this necessary for the usual water-supply problems, since it is included in the difference between the total rainfall of any district and the observed run-off in the streams. Indeed evaporation from reservoirs is similarly included for reservoirs existing when the run-off observations are made.