In developing a water supply for any purpose, the first step is to determine the approximate re lation between the quantity needed and the vari ous supplies that are or appear to be available. Where the source is to be some large river or lake, or say an artesian well from a basin of known capabilities. and in general where the re quired supply is obviously only a small part of the minimum yield of the source, no prelimi nary investigation as to quantity available will be requited. But in the majority of cases this point demands at least some consideration. The cardinal prineiple in estimating the quantitative fitness of a source of water supply is to compare the maximum demand with the minimum sup ply. averages are also of great importance, but the foregoing is the crucial test. There should be determined not only the actual or estimated yield of the source of supply during the driest month on record, but also the corresponding yield for the driest two or three years in succession. This will enable those concerned to calculate what storage will he necessary for use during droughts, what supplementary steam plants may be required in connection with water supply de velopments, and what delays to navigation may be expected in ease it is not feasible to provide all the storage indicated as necessary to main tain the levels of navigable canals or rivers. Although the yield, of both surface and under ground supplies depend primarily upon the rain fall, other factors vary in the two cases. Streams and lakes are supplied partly from the water that flows into them directly from the surface of the ground. hence the name surface water, and partly from water that soaks into the ground, then percolates through the soil to brooks. rivers, and lakes. The underground supplies are re plenished from time to time from that portion of the percolating water that does not find its way to the surface supplies, or else is intercepted by artificial means. The drainage areas of sur face supplies are well defined and the flow of streams may be readily measured.
The drainage areas of underground sources are not SO easily determined. particularly deep seated waters, and the determination of their volume is often extremely difficult and generally only an approximation. Instead of a stream flowing in a well-defined channel, as is true of surface supplies. underground waters are hidden in the earth and follow a tortuous course through thousands of tiny channels in the voids between the separate grains that compose their water bed. The size of these channels, the general slope of the underground water surface, the depth of the water-bearing stratum, or absence of the pressure that gives rise to artesian eon ditimis. all play their part in the possible yield of wells and springs. Some or all of these fac tors are but poorly known and their ascertain ment may be surrounded with grave difficulties.
The safest and often the only practicable means of determining the yield of an under ground source of supply is to measure it, or a portion of it. In the case of springs this may done by means of a weir (q.v.). Where wells are proposed one or more may be sunk and tested experimentally. but unless the test is for a long period and the deductions are made in the light of the hest engineering and geological knowledge the results arc quite certain to lie misleading and the future yield exag gerated. if a relatively small supply is all that
i. reyuirCd one or more test wells may settle the question of capacity with suffieieut definite ness at the outset, with the understanding that NVhea the supply, as finally developed. proves inadequate, the plant will be enlarged. ff. how ever. a quantity close to the probable safe yield of the underground leisin he desired, the lest wells may be so located as to determine approx imately the slope of Ii e water level and the porosity of the water-bearing material. From these facts, combined with other valuable data, an estimate of the available supply may be made. The effect of a well i, to lower the water level for a greater or less distance around it as a centre, so the new water level in this zone as sumes the shape of a flat inverted cone, which has had its surface curved somewhat in the vicinity of the well. The more the water is low ered the flatter becomes the cone and the greater the area of the zone influence, until the water practically fails through exhaustion, or through diminution due to increased friction. Obviously, if wells are placed too close together their zunes of influence will overlap and one will rob the other.
In determining the possible yield of surface supplies the first step is to ascertain what gang ings, if any. of the actual flow are available. and particularly whether such g.angings cover a series of years of minimum rainfall. Where no gaug ings are to be had it is desirable that a gaug ing station or stations should be established. If this cannot lie clone all the available rainfall records in and near the drainage area should be gathered and studied. These figures may be compared with the rainfall and corresponding yield. or run-off, of other drainage areas, as nearly similar as possible. Deductions may then be drawn as to the average, minimum, and maxi mum yields. The latter must be known to make possible the provision of adequate means for the passage of flood waters without damage to any of the structures connected with the supply works.
Storage reservoirs may often be provided to make good the deficiency of run-off or stream flow in dry periods. The extent to which this is feasible will depend partly upon the character of the available reservoir sites and largely upon the economic value of the water thus conserved. Absence of proper reservoir sites frequently turns the scales in favor of some other source of sup ply. The yield of underground water supplies may he supplemented storage, also. but to a. very small extent, as compared with the storage of surface supplies. since most underground waters would have to be pumped as well as stored, and since the conditions for water stor age on a large scale are rarely favorable in the vicinity of underground sources. Moreover, underground sources are chiefly used for publie water supplies, and ground water in storage for more than brief periods is liable to.deteriorate unless protected from the light. The expense of covering large reservoirs would be prohibitive. Relatively small supplies of well water are some times stored for irrigation (q.v.).