Pillsbury, A.F. 1981. The salinity of rivers. Scientific American. 245(1):54-65.
Rivers normally wash into the ocean the salts dissolved out of rock. When the flow is held up by irrigation, however, evapotranspiration concentrates the salt in the soil, creating problems for agriculture
Many ancient civilizations rose by diverting rivers and irrigating arid lands to grow crops. For such projects to succeed human beings had to learn to work cooperatively toward a common objective. The most fruitful of the ancient systems was created at the southeastern end of the Fertile Crescent, the broad valley formed by the Tigris and the Euphrates in what is now Iraq. From there civilization spread eastward through present-day Iran, Afghanistan, Pakistan, India and thence into China, where ever rivers disgorged through valleys of recently deposited alluvial soil. At its peak of productivity each irrigated region probably supported well over a million people. All these civilizations ultimately collapsed, and for the same reason: the land became so salty that crops could no longer be grown on it. The salts that were washed out of the soil at higher elevations became concentrated in the irrigated fields as the water evaporated from the surface and transpired through the leaves of the growing crops. Although floods, plagues and wars took their toll, in the end the civilizations based on irrigation faded away because of salination.
There was a notable exception to this common fate: the valley of the Nile, at the western end of the Fertile Crescent. The explanation is that until very recent times the Nile valley was not really irrigated. Instead the annual flood of the Nile inundated a broad stretch of land extending the length of the valley, depositing a new layer of fertile soil year after year. The flooding prevented the salts from accumulating because as the water drained into the soil it leached the salts out of it and carried them downward into the ground water, which eventually drained into the bed of the river itself. As a result salt balance was achieved more or less automatically; the salts were borne into the Mediterranean century after century as they had been for millenniums before the dawn of agriculture. With construction of the Aswan Dam at the headwaters of the Nile and the introduction of conventional irrigation Egypt is now faced with the universal problem of keeping salts from accumulating in the irrigated fields.
Irrignted agriculture has many advantages over humid-region agriculture in spite of the substantial costs of water and its application. Farmers can grow two or more crops per year of certain quicker-growing plants, such as grains and vegetable Moreover, an arid climate can be expected to decrease the costs of cultivation, planting; harvesting and disease control and to increase yields and quality. Water can be supplied whenever it is needed to ensure optimum growth. Above all, the technology now exists to solve the salination problem and to make irrigated agriculture truly permanent. Here I shall describe that technology in terms of its application in the western U.S., where intensive irrigation has been practiced since the turn of the century.
FORTUITOUS JUXTAPOSITION of the Imperial Valley and the Salton Sea in southern California near the border between the U.S. and Mexico makes it possible to export the brackish drainage waters from irrigated fields to a natural sink 370 square miles in area that is roughly
as salty as the ocean. This photograph, taken from a U-2 aircraft operated by the National Aeronautics and Space Administration from a height of 65,000 feet, shows only the extreme sortheastern tip of the nearly 500,000 acres that make the Imperial Valley the largest single expanse of irrigated agriculture in the Western Hemisphere. Growing vegetation, a strong reflection of infrared radiation, appears bright red in the Aerochrome infrared film used for the photograph. The water for irrigation is carried 80 miles from the Colorado River by the All American Canal (see map at top of next page). The All American Carnal feeds the East High-line Canal, visible immediately to the fields along the Salon Sea, and the Coachella Canal, slightly to the east, which crosses the picture from upper left to lower right. The Coachella Canal carries Colorado River water another 123 miles to Irrigated area north of Salton Sea.
All natural waters, including those described as fresh, contain salts. A virgin stream emerging from a mountain watershed may contain as little as 50 parts per miIlion p.p.m.) of "salt," or total dissolved solids. Ocean water averages about 35,000 p.p.m., or about 3.5 percent, of dissolved solids. I am speaking, of course, not simply of salt such as one finds on the dinner table: sodium chloride. Table salt does happen to be the most soluble of all the common salts, making up nearly a third of all the salt found in seawater but in natural waters there are many other salts as well. The salts dissolved in such waters are usually dominated by the carbonates, chlorides and sulfates of calcium, magnesium and sodium.
Accepting the commonly quoted estimate that the ocean basins contain 317 million cubic miles of salt water at an average concentration of 35,000 p.p.m. of dissolved solids, there are about 3.2 x 1018 tons of salt in the world's oceans. Additional large amounts of salt, largely in crystalline form, are sequestered in inland sinks and buried ancient sinks. Such salt deposits are one result of the geological processes summed up in the word weathering.
Weathering takes place under conditions where there is ample opportunity for the mineral crystals that constitute rock to oxidize. Although weathering embraces physical, chemical and biological processes, the physical processes are pervasive and central. Mechanical action fractures rock, exposing a far greater surface area to weathering agents. For example, the alternate freezing and thawing of water in the crevices of the rock exerts forces of compression and expansion that can break down the strongest material. Flowing water, wind and the grinding action of rocks in the bed of streams and the bottom of glaciers all contribute to physical weathering. Weathering manufactures both salts and the particles of rock that are borne from the uplands to the lowlands, where they are the principal constituents of soil.
The physical forces act most strongly at the highest elevations, where their extremes are greatest. As a result weathering and the consequent salt production are greatest at the highest elevations. An essential component of the weathering process is the water of precipitation, which serves to dissolve the salts. In areas of high precipitation, of course, the salts tend to be dilute.
When rain or snow falls, most of the water generally percolates into and through the ground; the excess beyond the capacity of the ground to retain water usually forms a water table at some depth below the surface. The ground water ultimately seeps into streams or reappears at the surface in the form of springs. It is this flow that makes streams and springs far more persistent than one would expect from the intermittent nature of the precipitation.
Much of the water retained in the soil and in the capillary fringe above the water table (the zone in which water can be drawn upward from the water table by capillary action) is available for the roots of plants to take up through the process of osmosis. Most of the water entering the roots is transpired through the leaves of the plant and passes as a vapor into the atmosphere. Water also evaporates directly from the surface of the soil. When the barometric pressure changes or when transient changes in pressure are induced by wind shear, the soil "breathes." Since the air in the soil is almost always essentially saturated with moisture, breathing represents additional evaporation. It is therefore not practical to determine how much of the moisture in the soil is lost by evaporation and how much is lost by transpiration, and so the total loss from land surfaces with vegetation is called evapotranspiration.
Water also evaporates directly from inland water surfaces such as those of creeks, rivers, lakes, marshes, canals and reservoirs. The result of all such evaporative processes is to distill pure water from the liquid phase, leaving all the salts behind in the water that remains. Since most of the watersheds in the western U.S. are quite thickly covered with vegetation, the natural concentration of salts in fresh waters is a widespread and significant process.
One of the principal natural resources of the western U.S. is its rivers, whose waters can be stored behind dams or in natural underground aquifers until they are needed. Another major resource is the alluvial soil of arid and semiarid valleys suitable for irrigated agriculture in a variety of excellent climates for growing crops. These resources and climatic conditions have been exploited to provide the nation and foreign markets with a diversified selection of foods and fibers that are largely complementary to the agricultural products of the humid regions.
Irrigated agriculture is expensive. Dams must be built to trap water and canals must be excavated to carry the water to where it is needed. If the water is in underground aquifers, energy must be expended to pump it to the surface. In California alone untold billions of public and private dollars have been invested in water development, chiefly for agriculture. The state is crisscrossed with thousands of miles of major irrigation canals and concrete pipelines that supply water to nearly nine million acres, a fifth of all the irrigated farmland in the nation. In 1980 the value of California's crops was $9.2 billion, representing more than 13 percent of the total value of all U.S. crops. California's farms provide some 40 percent of the nation's fresh food and vegetables.
The amount of water that must be applied annually to irrigated land is equivalent to between one foot and five feet of water covering the area cultivated, depending on the crop and the climate. Generally about three-fourths of the applied water is lost to evapotranspiration. The rest, which contains all the originally dissolved salts (except for the tiny amounts incorporated in the crop it-self), percolates downward and laterally through the soil. It may enter an underground aquifer or it may reappear somewhere downstream as seepage into a river or into a natural sink, either directly or by way of a drainage ditch.
If, as seems reasonable, the average annual amount of water applied in irrigation in the western U.S. is equivalent to three feet covering the area cultivated, about 120 million acre-feet of water is applied annually to some 40 million acres of land. Roughly 90 million acre-feet of the total volume is lost by evapotranspiration. The remaining 30 million acre-feet hold: essentially all the original salts: a fourfold concentration. (Water of good quality can be used at least twice, directly or indirectly, before it becomes too brackish for further use.) As a consequence of intensive irrigation the western states, particularly California, which applies more than 40 million acre-feet of water to its crops, face a major problem in isolating and disposing of brackish water.
In addition to the concentration of salts by evapotranspiration another kind of concentration takes place in the storage and transport of water before it ever arrives at the point of use. Evaporation occurs at the large storage reservoirs behind dams, along the extensive transmission channels and at the numerous regulating reservoirs. If the new water surfaces created are in mountainous areas where precipitation is high and where the land was formerly covered by forests and meadows, the increase in evaporation is probably of little significance; the reason is that the former evapotranspiration of natural vegetation is merely replaced by a roughly equal degree of simple evaporation. Most existing large reservoirs, however, are in desert areas or in environments formerly covered only sparsely by brush, grass-woodland or scrub timber, with the result that evaporation greatly exceeds the former evapotranspiration. Typical examples are the reservoirs formed by Hoover Dam on the Colorado River, Elephant Butte Dam on the Rio Grande and Grand Coulee Dam on the Columbia River.
The great rivers of the western U.S. originate in mountains where there is generally high annual precipitation. Characteristically in or near the mountains the concentration of total dissolved solid: is low, commonly in the range of 50 p.p.m. for most rivers. Since a number of the western rivers flow through arid or semiarid regions for great distances, particularly in their lower reaches, the concentration of salts through evaporation rises steadily with distance downstream. If part of the flow is diverted for irrigation, the salt concentration is amplified by evapotranspiration. Increasingly water serves for cooling purposes, as in the cooling systems of fossil-fuel and nuclear power plants, which therefore also "consume" water through evaporation and thereby concentrate salts.
Some years ago I thought it would be interesting to analyze several major river systems to see how the quantity of salts carried by a river at a given point was related to the quantity of water that had entered the river above that point. To this end I calculated the annual salt production in tons per square mile of watershed above a given point on the river and the annual water production in acre-feet per square mile of the same watershed. The salt content and the flow of major rivers at various points are recorded by a network of stations run by the U.S. Geological Survey. I selected 10 years as the period over which to average the records because that was generally the maximum length of record available. For the survey I selected the following rivers: the Missouri, the Arkansas, the Rio Grande, the Pecos, the Colorado, the Gila, the Salt, the Agua Fria, the Sevier, the Humboldt, the San Joaquin, the Mokelumne, the American, the Columbia and the Willamette. These 15 rivers carry the bulk of the precipitation that falls on the 11 westernmost states of the U.S.
When salt production is plotted against water production on a logarithmic scale, the relation not only proves to be remarkably linear but also refutes the common assumption that rivers flowing through the more arid region: carry the most salt (see bottom illustration on opposite page). Actually the higher the water production per square mile, the higher the production of salts. The relation is all the more remarkable when one considers the great variation in the type and age of the rocks in the several river basins, the relative shortness of the 10 years of records available and the effects of diversions across watersheds, which transport substantial volumes of water of low salt content from one river basin to another.
ALL-AMERICAN CANAL originates at a reservoir behind Imperial Dam on the Colorado River. The 80-mile can carries two-thirds of the 5.3 million acre-feet per year of water that California has recently been drawing from the Colorado. Ultimately the state will be limited to 4.4 million acre-feet per year. Some three million acre-feet per year is now applied to farmland in the Imperial Valley. The rectangle outlines the area shown in the U-2 photograph on page 54. The water as it is drawn from the Colorado carries about 800 parts per million (p.p.m.) of dissolved solids ("salts"). After about three-fourths of the water applied to crops has evaporated or transpired through the leaves of plants, the water drained from the fields contains about 3,200 p.p.m. of salts. Most of the drainage water is carried to the Salton Sea by the New River and the Alamo River. In past millenniums the Colorado emptied about half the time into the Gulf of California and half the time into the Salton Sea, forming the ancient Lake Cahuilla. When the Colorado was first seen by European explorers, it was flowing into the Gulf of California. It continued to do so until 1905, when it was again diverted into the sink for two years before it could be rediverted to its former channel. Surface of the Salton Sea is now 230 feet below sea level, giving the water a maximum depth of 40 feet. The Salton Sea is the largest body of water in California.
DRAINAGE OF IRRIGATED FIELDS is often required to maintain salt balance in the root zone of plants. The irrigation water that percolates downward through the soil is enriched in salts because of evapotranspiration, the combination of direct evaporation and plant transpiration. To collect this brackish water the farmer installs pipes in parallel lines, usually about eight feet deep and 250 to 300 feet apart. The pipes, which are either loosely fitted or perforated, were once commonly made of tile or concrete but now are usually made of plastic. The drains form open channels through which ground water will flow when it reaches a level high enough to harm the roots. The brackish water usually flows to a drain but sometimes empties into sumps and is pumped into larger channels called collector drains.
The most productive river in both salts and water is the Willamette in Oregon, which flows through a region of very high precipitation. The least productive in both salts and water is the Gila below Gillespie Dam in southwestern Arizona, a desert region. Even today it is asserted that rivers such as the Arkansas, the Pecos and the Colorado are unusually salty because they are slowly leaching away ancient buried salt beds. The evidence does not support such an assertion. The saltiness of rivers is simply a matter of the relative amount of water that is turned into water vapor through consumption, whether it is natural or the result of human activity.
The key to maintaining a salt balance in irrigated fields is adequate drainage. Whether it is natural or artificial, drainage refers to the removal of water from a place where it is not wanted to some other place, through a pipe or channel that can be on, above or below the land surface. The term agricultural drainage refers specifically to measures intended to lower the depth of a water table that is too close to the surface to allow the successful growing of crops. In humid regions the water table may have to be lowered in order to provide aerated soil around plant roots and to increase the firmness of the soil for tillage and other farm operations. For this purpose in such regions a network of ditches or tile drains is normally laid out at a depth of three to five feet below the surface.
In arid regions drainage must serve the additional function of maintaining a satisfactory salt balance in the vicinity of plant roots. When the weather is rainy or when irrigation water is being applied, water and salts both percolate downward. In dry weather and between irrigations the water and salts percolate upward through capillary action. In a humid area the amount of salt entering the soil is low to begin with, and so there is far less to travel upward through capillary action in dry periods. Moreover, the dry periods are usually short In arid climates the drainage ditches or tile drains must be deeper than they are in humid regions in order to prevent a net upward movement of salts. To be effective the drainage channels must usually be at least six feet below the surface. The effluent from the drainage network must be discharged in such a way that the salt imported with the irrigation water is exported without harming the interests of water users downstream.
A remarkable and fortuitous facility for the disposal of brackish drain water exists in southern California adjacent to the Imperial Valley, the largest single expanse of irrigated agricultural land in the Western Hemisphere. The water needed to irrigate the valley's more than 500,000 acres is carried a distance of 80 miles from the Colorado River by the All-American Canal. The All-American Canal also supplies the Coachella Canal, which carries Colorado River water an additional 123 miles to another rich agricultural area of some 65,000 acres. (The Colorado supplies a little more than half of the water used in southern California including the municipal water of state's two largest cities, Los Angeles and San Diego.) Brackish irrigation water that drains from the Imperial Valley and Coachella agricultural district is channeled into the Salton Sea, which at present is a little saltier than the ocean. Some 90 percent of the surface inflow to the Salton Sea is waste water from t Imperial Valley, Coachella and Mexicali districts.
Since 1955 the Imperial Valley Irrigation district has been a net exporter of salts, draining out about about 15 percent more salt than the All-American Canal carries into the district from the Colorado.
IRRIGATION OF AGRICULTURAL LANDS represents the biggest consumptive use of water in the U.S. Water applied to crops is termed consumptive because three-fourths of it is dissipated into the atmosphere through evapotranspiration. It is estimated that aoubt 120 million acre-feet per year of water is applied to some 40 million acreas of land in the western states, ore about three feet of water to every irrigated acre over the growing season. Assuming that three-fourths of the water, or 90 million acre-feet, is lost to evapotranspiration, the salts in the original volume, represented here by shades of color, are concentrated in the remaining 30 million acre-feet of water. This water, often containing more than 2,000 p.p.m. of salts, must be drained from the fields and then disposed of to prevent a buildup of salts.
The water drained into the Salton Sea contains about 3,500 p.p.m. of salts and serves to retard the rate of increase in the Salton Sea's overall salinity level. The Salton Sea itself, which lies 230 feet below sea level, was a dry, salt-encrusted depression until 1905, when a flood on the Colorado broke through natural levees. The water of the Colorado poured into the sink for two years be-fore it was rediverted into its former channel. With an area of 370 square miles, the Salton Sea is California's largest lake and a major recreational area.
Where no Salton Sea or its equivalent exists to accept the drainage from irrigated fields the problem of achieving salt balance is more complex. People cherish the notion that the water of a river not only is fresh but also should be kept fresh right down to the river's mouth or to its entry into an estuary. In the humid regions of the world the departure from this ideal is seldom great, but the ideal is unrealistic in the more arid regions, where many of the rivers have been developed for irrigated agriculture. Before man began harnessing the rivers the seasonal floods were highly effective in carrying salts to the ocean and keeping the river basin in reasonably good salt balance. Today, with river flows being regulated by storage systems, and with high consumptive use of the released water, there is not enough waste flow left to achieve anything approaching balance. The salt is being stored, in one way or another, within the river basins.
Not only are salts getting bogged down somewhere in the system but also various measures are being taken that deliberately impede the flow of salts to the sea. In the U.S. it is the law of the land, reflecting the demands of both environmentalists and water users, that rivers remain, if not forever "wild and scenic," at least fresh for their entire length. The measures being planned and effectuated to accomplish this ideal are dangerous for the future. The general concept is to divert saline flows, where they can be found, into evaporation basins. There water will evaporate from the surface, leaving behind layer on layer of crystalline salts. It is proposed that the evaporation basins be situated either where the underlying ground is already saline or where the soil is relatively nonporous. Where neither is the case the ponds are to be lined with a presumably impervious material. Such schemes, designed to Store the salts in the river basins themselves, may work for a few years or decades but are bound to be disastrous in the long run.
RIVER AND LOCATION
CONDUCTIVITY (MILLIMHOS PER CENTIMETER) COLUMBIA AT
WENATCHEE, WASH. 78 .15 .2 SACRAMENTO AT
TISDALE, CALIF. 180 .16 .6 MISSOURI AT
WILLISTON, N.D. 574 .84 2.0 COLORADO AT YUMA,
ARIZ. 740 1.06 2.2 RIO GRANDE AT EL
PASO,TEX. 754 1.16 3.6 ARKANSAS AT LA lUNTA,
RIVER AND LOCATION
ELECTRICAL CONDUCTIVITY (MILLIMHOS PER CENTIMETER)
COLUMBIA AT WENATCHEE, WASH.
SACRAMENTO AT TISDALE, CALIF.
MISSOURI AT WILLISTON, N.D.
COLORADO AT YUMA, ARIZ.
RIO GRANDE AT EL PASO,TEX.
ARKANSAS AT LA lUNTA, COLO.
QUALITY OF IRRIGATION WATERS in the western U.S. varies greatly. The wide diversity is reflected in these samples taken from six rivers. The Colombia in the extreme northwest, 1,240 miles long and with a total discharge of 455 million acre-feet per year, is second in total flow only to the Mississippi. The Arkansas, which is even longer than the Columbia (1,460 miles), discharges only a sixth as much water and ranks 13th among U.S. rivers. The discharges of the other four rivers are considerably smaller. The term "total dissolved solids" is still widely used as a measure of salinity, but for the farmer the other two characteristics in the table are even more significant. Electrical conductivity has a greater influence on plant growth than salinity alone. Low values are preferable. Sodium absorption ratio is calculated from the abundances of sodium, calcium and magnesium ions, expressed in milliequivalents of each ion per liter. Broadly speaking, the ratio expresses the excess of sodium, or the deficiency of calcium, that adversely affects the permeability of water in the soil. Values below 10 are satisfactory.
LEACHING OF SALTS FROM THE GROUND Increases directly with the amount of precipitation. The higher the runoff of water per unit of watershed area is, the more salt is carried into the river that drains the watershed. To quantify the relation, the author analyzed U.S. Geological Survey records for a 10-year period showing the salt content and river-flow volume at selected sampling points for 15 rivers in the western U.S. The salt content and flow rates were then related to the areas of the watersheds above the sampling points in order to derive values for salt production in tons and water production in acre-feet per square mile per year. A logrithmic plot of the two values falls remarkably close to a straight line, The Willamette River, which drains a watershed with the highest precipitation per square mile, also yields the most salt per square mile of watershed. The least productive river in both water and salts is the Gila River in extremely arid Arizona. The four sampling points (colored dots) on the Colorado River and the watersheds to which they correspond are shown in the map on page 62. Colored crosses are values for the Rio Grande, the only other river for which four sampling points are Included.
Why? The schemes will fail for any of several reasons. Although the ground waters under the evaporation basin may well be brackish or saline, every groundwater basin with a flow gradient must have an outlet somewhere near its lower end. The saline water in the evaporation basin will serve to increase the "head," or hydraulic pressure, on the saline waters below and will thereby increase the rate of discharge at the natural outlet, wreaking havoc in downstream ground waters and downstream lands. If the evaporation basin is situated above soil shown to be impermeable to fresh water, it will be found that the soil will gradually become more permeable when the waters are saline. This fact is well established. Many types of materials have been proposed for making evaporation basins impervious: rubber and plastic sheeting, asphaltic mixtures and special types of concrete. Conceivably some linings will be effective for as long as 50 years, but ultimately one must expect them to fail. In all probability their lifetime when they are exposed to saline water will be much shorter than their lifetime is when they are exposed to fresh water, for which they are normally intended.
Another prime effort today, designed to keep waters fresh in the lower reaches of river-basin systems, is the construction of "brine lines." These are lined canals or pipelines through which brackish waters are conveyed to the ocean or some other sink. The lines must be elevated above nearby rivers or the adjacent systems for distributing irrigation water to ensure that the saline drain water does not recontaminate the fresh water. This means that the effluent from field-drainage systems, along with any brackish or saline water from wells, must be pumped into the brine lines at a considerable cost in energy. Even if the energy cost is accepted, brine lines alone cannot establish a salt balance because there will still be ground-water flow below the drainage lines or above the saline aquifers that might be pumped.
The U.S. Water and Power Resource Service (formerly the U.S. Bureau of Reclamation) has recently completed 82 miles of an open channel brine line, roughly parallel to the San Joaquin River in California, that is designed to drain up to 30,000 acres of prime agricultural land in the state's San Joaquin Valley. Probably the longest brine line yet built, it now discharges into the Kesterson Reservoir south of Modesto.From there the saline waters gradually seep into the San Joaquin River as it approaches "the Delta," the estuarial area at the head of San Francisco Bay formed by the confluence of the San Joaquin and Sacramento rivers. The section of the drain that will carry the brine directly into the Delta has yet to be built. The existing canal is the first segment of a proposed "Master Drain, 290 miles long, to be financed with Federal and state funds. The estimated cost is more than $1.2 billion. By the year 2005 the proposed drain will serve 500,000 acres and will have the capacity to remove more than three million tons of salt per year.
DESALTING TEST FACILITY is one of two installations at Yuma, Ariz., recently built for the U.S. Water and Power Resources Service as the first stage in a program to treat brackish water from the Welton-Mohawk Irrigation and Drainage District that is too salty to be returned to the Colorado River. At present the brackish water is carried by a 51-mile bypass drain to the Gulf of California. This test facility, built by the Fluid Systems Division of UOP, is designed to extract some 325,000 gallons per day of low-salinity water (about 250 p.p.m. of solids) from a brackish input stream of 470,000 gallons per day with a salinity of about 2,800 p,p,m. Desalting is accomplished by a process of reverse osmosis. Saline water is forced against a plastic membrane at a Pressure 300 to 400 pounds per square inch. Water passes through and salt is left behind. The full-scale plant using two different but similar processes, will have a design capacity of some 72 million gallons of desalted water per day. The product stream will be blended with untreated drain water to yield about 92,000 acre-feet per year of water (with a quality higher than that required by treaty) that can be added to the Colorado for export to Mexico. An equivalent volume of water will be made available for irrigation upstream in the U.S.
The difficulty of achieving salt balance in river basins where there is high consumptive use of water can be better appreciated if one reflects on the basins' complex hydrological history. Valley and basin lands consist primarily of soils that have been deposited by floods. The river channel that winds through the valley and basin today has a look of permanence that is deceptive. In a flood heavy debris is tumbled along the deepest part of the channel and the river overflows its banks. As the water moves laterally there is a marked decrease in the velocity and depth of its flow. As a result sand and other small particles settle out, creating natural levees. The finest soil particles settled out at a considerable distance from the central channel and at a much lower rate. The soil that settles in the valley lands is thus of medium texture, chiefly loams.
Where the velocity of flow falls
almost to zero in the broad basin lands farther downstream
the fine-textured clays settle out. The channel itself,
including the natural levees, tends gradually to rise above
the surrounding land. Later, in some exceptional flood, the
river will overflow its banks again and create a new channel
where the slope below it is steeper. The new channel will
capture the old channel as it erodes upstream. In time, as
deposition from flood after flood raises the land, an
interlacing network of buried channels is covered over.Both
shallow and deep ground water tend to follow such
interlacing channels, with the deeper water appearing in
what are called finger aquifers. Particularly in the basin
lands the shallow ground water commonly seeps into the
present river channel.Such diffused flow is deeper than the
ditches or tile drains constructed for agricultural
drainage. Tile drains are most effective, as a
rule, in the basin and basin-rim lands and in certain
stratified soils where the "semiperched" water table rises
closer to the surface than six or seven feet. The water
table is said to be semiperched when the variable layering
of the river-deposited soil tends to isolate the water near
the surface from the main body of ground water lying at
deeper levels. Under those conditions there is little
opportunity for the irrigation water, enriched in salts, to
percolate downward and degrade the deep ground water, which
remains available for irrigation or other uses. Farther
upriver in the valley lands the subterranean structure is
such that near-surface water cannot be isolated from deeper
water, with the result that ditch or tile-drain systems are
powerless to preserve the quality of the ground
water. Before the advent of intensive
irrigation the ground waters of the western valleys and
basins were almost uniformly of high quality. The
underground aquifers were largely recharged at the upper end
of the valleys where the rivers disgorged onto the valley
lands. The ground waters subsequently discharged into the
basin lands and for the most part into the rivers themselves
in the form of diffused flow. TO CONSERVE WATER formerly lost
through seepage 49 miles of the 123-mile Coachella Canal
have recently been rebuilt and lined with concrete. The new
section, a part of which is shown near completion here, is
expected to save 132,000 acre-feet of water per year,
reducing the amount drawn from the Colorado River via the
All-American Canal from 498,000 acre-feet per year to
Where the velocity of flow falls almost to zero in the broad basin lands farther downstream the fine-textured clays settle out. The channel itself, including the natural levees, tends gradually to rise above the surrounding land. Later, in some exceptional flood, the river will overflow its banks again and create a new channel where the slope below it is steeper. The new channel will capture the old channel as it erodes upstream. In time, as deposition from flood after flood raises the land, an interlacing network of buried channels is covered over.Both shallow and deep ground water tend to follow such interlacing channels, with the deeper water appearing in what are called finger aquifers. Particularly in the basin lands the shallow ground water commonly seeps into the present river channel.Such diffused flow is deeper than the ditches or tile drains constructed for agricultural drainage.
Tile drains are most effective, as a rule, in the basin and basin-rim lands and in certain stratified soils where the "semiperched" water table rises closer to the surface than six or seven feet. The water table is said to be semiperched when the variable layering of the river-deposited soil tends to isolate the water near the surface from the main body of ground water lying at deeper levels. Under those conditions there is little opportunity for the irrigation water, enriched in salts, to percolate downward and degrade the deep ground water, which remains available for irrigation or other uses. Farther upriver in the valley lands the subterranean structure is such that near-surface water cannot be isolated from deeper water, with the result that ditch or tile-drain systems are powerless to preserve the quality of the ground water.
Before the advent of intensive irrigation the ground waters of the western valleys and basins were almost uniformly of high quality. The underground aquifers were largely recharged at the upper end of the valleys where the rivers disgorged onto the valley lands. The ground waters subsequently discharged into the basin lands and for the most part into the rivers themselves in the form of diffused flow.
TO CONSERVE WATER formerly lost through seepage 49 miles of the 123-mile Coachella Canal have recently been rebuilt and lined with concrete. The new section, a part of which is shown near completion here, is expected to save 132,000 acre-feet of water per year, reducing the amount drawn from the Colorado River via the All-American Canal from 498,000 acre-feet per year to 366,000 acre-feet.
In most regions the water pumped from aquifers supplies both irrigation and urban needs. The urban wastes collected as sewage generally show an in-crease in total dissolved solids of 300 to 350 p.p.m. Where such wastes are not discharged directly into the ocean all the salts coming into the basin remain trapped and build up within it.
When the aquifers being pumped for agricultural and urban needs are near the coast, the water table has often been pumped below sea level, with a consequent intrusion of seawater into the ground-water basin. The usual way to stop the intrusion has been to drill a series of injection wells parallel to the coast. The water pumped into the injection wells can be somewhat brackish and in some cases is treated urban sewage. The technique has been successful in creating "mounds" of water that repel the seawater. The objection to such schemes is that they totally block the export of salts that would otherwise be carried to the sea by ground waters. s Obviously corrective measures must eventually be taken.
The only effective way to keep ground-water basins fresh is to pump from wells near the lower end of each basin, where the salinity is highest, and to hurry the effluent on its way to the ocean or some other sink. At the same time it will probably be essential to augment the recharge near the upper end of each basin. Unless these steps are taken one can foresee the day when the aquifers will be destroyed by salinity.
COLORADO RIVER BASIN drains an area of 242,000 square miles in seven states. Nine major reservoirs on the river have a total storage capacity of about 65 million acre-feet, or roughly 4.5 times the annual flow of 14 million acre-feet measured at Lee Ferry. The largest reservoir, with a capacity of 26 million acre-feet, is Lake Mead, made by Hoover Dam. Under the terms of the Colorado River Compact of 1922 the four states in the upper basin above Lee Ferry are allotted 6.5 million acre-feet annually and the three states in the lower basin (Arizona, Nevada and California ) are allotted 7.5 million acre-feet. When the Central Arizona Project, designed to carry 1.2 million acre-feet per year, is completed in 1985, California will be limited to 4.4 million acre-feet, leaving 1.5 million acre-feet for Mexico. In the drought year of 1977 California drew on the Colorado River for 5.6 million acre-feet of water. The four sampling points selected by the author in his study of salt and water production in the Colorado basin, shown in the bottom illustration on page 59, are Glenwood Springs(1), Lee Ferry(2), Grand Canyon (3) and a point below Hoover Dam (4). The watersheds corresponding to these areas are identified by the light-colored broken outline. Whereas the Glenwood Springs watershed produces 350 acre-feet of water and 135 tons of salt per square mile per year, the Hoover Dam watershed yields only 60 acre-feet of water and 50 tons of salt per square miles per year.
The custom traditionally followed in the U.S. in developing water resources has been to expect the river itself to carry supplies of fresh water to points of diversion almost down to the river's mouth. Such a design is generally the cheapest and has the advantage of capturing floodwaters that upstream storage would miss. This ignores the basic principle, essential for the long term, of going upstream for supply and allowing the lower rivers to become brackish. Let me illustrate with three examples: the Rio Grande system, the Colorado River system and the Delta east of San Francisco Bay.
The Rio Grande, which rises in southwestern Colorado and empties into the Gulf of Mexico, is nearly 1,900 miles long. It is the third-longest river in the U.S., yet like the Colorado it does not qualify for inclusion in the Geological Survey list of 33 rivers with the highest discharge. Over the final 800 miles of its length it is the principal boundary between the U.S. and Mexico, and the allocation of its water has long been a matter of contention between the two countries. Essentially the entire flow of the upper Rio Grande, except during floods, is stored and utilized upstream from El Paso, at the extreme western end of the U.S.-Mexico border. There is almost no waste that would make it possible even to approach salt balance. Severe salt problems are gradually developing in southern New Mexico and western Texas. Along the lower river, between El Paso and the mouth in the Gulf of Mexico, there are three major international dams: Amestad, for storage; Falcon, for more storage and hydropower generation, and Anzalduas, for diversion, chiefly irrigation. (Mexico calls the lower river from El Paso to the Gulf the Rio Bravo del Norte; most of the water entering it is runoff from mountains in Mexico.) The present plan of development eliminates any chance of achieving salt balance either above El Paso or below it.
The Colorado River, 1,450 miles long, supplies more water for consumptive use than any other river in the nation. Its well-known Hoover Dam created Lake Mead, a storage reservoir and recreational area of some 250 square miles. Below Hoover Dam and its 1,345-megawatt hydroelectric power plant are seven more dams, two of which also serve to generate power. The dam farthest downstream is Morelos Dam, which stores water for irrigation the Mexicali Valley in Mexico. Irrigated regions near the river send back their drainage water enriched fourfold in salt. About 70 percent of the total flow below Hoover Dam, containing about 700 p.p.m of salt, is exported to California, chiefly through the Colorado River Aqueduct, which supplies Los Angeles and San Diego, and the All-American Canal, which supplies the Imperial Valley and Coachella agricultural regions.
Primarily because so much water (and salt) is exported, the lower reaches of the Colorado are in reasonable salt balance. The problem is that the salt content of the lower river is high: more than 800 p.p.m. In 1974, at the behest of the Environmental Protection Agency, the seven states of the Colorado River basin agreed on a program to maintain the salinity in the lower basin at or be-low the level measured in 1972: 723 p.p.m. below Hoover Dam, 747 p.p.m. below Parker Dam and 879 p.p.m. at Imperial Dam.
The same year, in partial satisfaction of the 1944 Mexican Water Treaty, the U.S. agreed that the salinity of the Colorado River water delivered to Mexico at Morelos Dam should not be allowed to exceed the average salinity of water arriving at imperial Dam by more than 115 (±30) p.p.m. The treaty with Mexico provides that the U.S. shall deliver to Mexico 1.36 million acre-feet of Colorado River water and another 140,000 acre-feet from well fields adjacent to the Colorado at the U.S.-Mexico border, for a total of 1.5 million acre-feet. In order to ensure the required river flow at the agreed salinity, the Water and Power Resources Service has undertaken to build a desalting plant at Yuma, Ariz., that will process a large volume of brackish drain water from the Wellton -Mohawk Irrigation and Drainage District in Arizona. The brackish water was formerly discharged into the Colorado below Morelos Dam; it is now exported to the Gulf of California through a bypass drain 51 miles long, mostly in Mexico, that was completed in 1977 at a cost of $27 million, paid by the U.S.
The desalting plant, when it is completed in the mid- 1980's at an estimated cost of $216 million, will be the largest of its kind. It will take in about 107,000 acre-feet per year of water with an average salinity of 2,800 p.p.m. and will yield a cleansed stream of 73,000 acre-feet (65 million gallons a day) with salinity of only 255 p.p.m. and a brine stream of 34,000 acre-feet with a salinity of 8,200 p.p.m. The brine stream will continue to be sent on to the Gulf of California. If low-salinity makeup watt is needed to meet treaty obligations, the cleansed stream can be returned undiluted to the Colorado at Yuma. Normally, however, the cleansed stream will be blended with untreated drainage water to yield up to 92,000 acre-feet of water with a salinity of less than 800 p.p.m. The reclaimed water will cost about $250 per acre-foot, more than 31 times the cost of irrigation water in the Imperial Valley. The desalination will be accomplished by plastic membranes that remove salts by means of reverse osmosis.
To further ensure the flow to Mexico and to reduce the volume of water taken by California, 49 miles of the Coachella Canal have recently been rebuilt and lined with concrete at a total cost of $45 million in an effort to save 132,000 acre-feet per year of water previously lost by seepage into the surrounding desert. It isl expected that the amount drawn from the Colorado River via the All-American Canal can be reduced from 498,000 acre-feet per year to 366,000 acre-feet. The lining of the canal will also help California to live within its ultimate alocation of 4.4 million acre-feet of Colorado water. (It has recently been drawing 5.3 million acre-feet per year, and it drew 5.6 million acre-feet in the drought year of 1977.) Although these various costly measures should enable the U.S. to meet its treaty obligations to Mexico, it is clear that the Colorado can yield no additional water for the expansion of agriculture.
The Delta is an estuarial area of some 1,200 square miles formed by the confluence of the Sacramento, the San Joaquin and several smaller rivers near the middle of the rich Central Valley agricultural region. This large area was originally bulrush-covered marshland at or near mean sea level. Beginning in a small way sometime after the passage of the California Swamp and Overflow Act of 1850 but not before the rare great flood of 1861-62, the land was gradually reclaimed by the building of levees around numerous small islands outlined by the rivers and sloughs. When better pumps became available around the turn of the century, the process was greatly accelerated. The reclamation continued into the 1920's. The material for the levees came from dredging the rivers and sloughs. As with the polders of the Netherlands, the reclamation requires that drainage water and rainwater be continually pumped out into an adjacent waterway. Irrigation water is obtained simply by siphoning water over the levees. There are some 550,000 acres of cropped land in the Delta complex, and the labyrinth of waterways is heavily exploited for recreation.
The land was originally all peat, of-ten to depths of 35 feet or more. As happens whenever peat soil is drained, there has been widespread subsidence of the ground level, amounting to about three inches per year. Some land is now as much as 21 feet below mean sea level. Not surprisingly there have been many levee failures. Over the past two years four large islands were inundated when their levees broke. State and Federal agencies are now working hard at public expense to repair and strengthen the levees involved. It is reasonable to question whether the expense is justified, given the fact that the subsidence problem is clearly a long-term one.
An alternative approach to the Delta problem, proposed some years ago, would be to provide the stablest farmlands in the Delta with an overland supply of fresh water from the projected Peripheral Canal. The canal would also serve to carry water from the Feather River, north of the Delta, around the eastern end of the Delta to the San Joaquin Valley on the south. Although the California legislature authorized construction of the Peripheral Canal last year, the project has run into heavy opposition from environmentalists, among others, who are demanding that the issue be submitted to a statewide referendum. In addition to the construction of the Peripheral Canal, at a currently estimated cost ranging from $700 million to $1.3 billion, the legislature authorized the expenditure of $4.1 billion for construction over the next quarter century of dams and other water-development facilities in the northern part of the state.
At present the water exported south and west must move through the Delta sloughs before it reaches the pumps that send it to the ultimate users. Periodically fresh waters that normally flow down-stream through the Delta reverse their direction and allow salt waters from San Francisco Bay to travel upstream toward the export pumps. As a result the Delta waters and the water destined for irrigation are both seriously degraded in quality.
The Peripheral Canal has been opposed
by many in northern California who evidently do not
recognize that the water contractors of the State Water
Project have long been paying in full on the Oroville Dam
complex on the Feather River that is designed to create the
"new" regulated water destined for the canal. The export
pumps on the Delta cannot possibly provide the amounts of
water the state has contracted to deliver unless the
Peripheral Canal is built. Opposition now comes from those
who advance the concept that the water belongs in the "area
of origin" and that those who are paying for it must now
surrender their right to it.
ELABORATE SCHEME for
diverting water from abundant sources in Alaska and
Canada to regions that are already growing short of
water was originally proposed by the Ralph M,
Parsons Company of Pasadena, CA. in 1964. Only the
western section of the total plan, known as the
North American Water and Power Alliance (NAWAPA),
is depicted schematically here. An eastward
extension of the plan would divert water from
British Columbia to the Delta of the Mississippi
River, the Great Lakes and eastern Canada. It is
estimated that the entire project would cast at
least $200 billion and take 30 years to complete.
NAWAPA would yield about 16 million acre-feet of
water per year for industrial, municipal and
agricultural uses. Roughly half would he allocated
to the US. and the balance to Canada and Mexico. A
series of hydroeletric plants would generate all
the power needed for pumping and provide a surplus
capacity of at least 70,000 megawatts. Water
collection would begin at the headwaters of the
Yukon and Tanana rivers in Alaska (1) and
would be supplemented by utilizing a series of
streams extending down to the Pecos River (2).
The collected water would flow into a 5OO mile
reservoir, the Rocky Mountain Trench, and by
damming the upper reaches of the Columbia, Fraser
and Kootenay rivers. More water would enter the
system from the Clark, Snake and other rivers in
the northwestern U.S. (3). Water for
irrigation and other purposes would he sent to
southwestern states and northwestern Mexico
(4), A branch would supply New Mexico,
Texas, Colorado, Kansas, Nebraska, Oklahoma and
other parts of Mexico (5).
ELABORATE SCHEME for diverting water from abundant sources in Alaska and Canada to regions that are already growing short of water was originally proposed by the Ralph M, Parsons Company of Pasadena, CA. in 1964. Only the western section of the total plan, known as the North American Water and Power Alliance (NAWAPA), is depicted schematically here. An eastward extension of the plan would divert water from British Columbia to the Delta of the Mississippi River, the Great Lakes and eastern Canada. It is estimated that the entire project would cast at least $200 billion and take 30 years to complete. NAWAPA would yield about 16 million acre-feet of water per year for industrial, municipal and agricultural uses. Roughly half would he allocated to the US. and the balance to Canada and Mexico. A series of hydroeletric plants would generate all the power needed for pumping and provide a surplus capacity of at least 70,000 megawatts. Water collection would begin at the headwaters of the Yukon and Tanana rivers in Alaska (1) and would be supplemented by utilizing a series of streams extending down to the Pecos River (2). The collected water would flow into a 5OO mile reservoir, the Rocky Mountain Trench, and by damming the upper reaches of the Columbia, Fraser and Kootenay rivers. More water would enter the system from the Clark, Snake and other rivers in the northwestern U.S. (3). Water for irrigation and other purposes would he sent to southwestern states and northwestern Mexico (4), A branch would supply New Mexico, Texas, Colorado, Kansas, Nebraska, Oklahoma and other parts of Mexico (5).
Over the years a number of visionary schemes have been proposed for diverting water from rivers in the humid northwestern U.S. to more arid regions to the south. One plan. conceived by a private engineer, William G. Dunn, proposes to transfer 2.4 million acre-feet per year from the Snake River in Idaho to a point below Hoover Dam on the Colorado River, a distance of some 600 miles. The Snake, which discharges about 30 million acre-feet per year into the Columbia River, is the nation's 12th-largest river. The present flow of the Colorado below Hoover Dam is about 14 million acre-feet per year, of which California will be entitled to take 4.4 million after completion of the Central Arizona Project, beginning in about 1985. A more ambitious plan, put forward by Frank Z. Pirkey, another experienced consulting engineer, proposes the transfer of 15 million acre-feet per year from the Columbia River to Lake Mead, behind Hoover Dam. In Pirkey's plan the water would first be pumped 4,900 feet over the mountains to Goose Lake on the Oregon-California border and then to Shasta Lake behind Shasta Dam in northern California before its final transfer to Lake Mead. Along the way substantial amounts would be available for irrigation.
By far the most audacious scheme yet advanced is the one proposed in 1964 by the Ralph M. Parsons Company of Pasadena, CA., one of the engineering firms that built Hoover Dam in the early 1930's. Known as the North American Water and Power Alliance (NAWAPA), the scheme would divert waters from Alaska and northern Canada to many parts of Canada, the U.S. and Mexico. Hydroelectric plants along the way would generate substantially more power than what would be needed for pumping. The total drainage area envisioned by the plan covers about 1.3 million square miles where precipitation is plentiful. Of a total runoff of more than 800 million acre-feet per year NAWAPA would divert some 160 million acre-feet southward for consumption and waterway control. The completed system, estimated to take 30 years to build and to cost more than $200 billion, would provide a surplus hydroelectric capacity of about 70,000 megawatts, equal to nearly 25 percent of present U.S. average production.
Although the magnitude of the NAWAPA plan is staggering and the plan would have to surmount formidable political hurdles before it could be implemented, it is in my opinion the only concept advanced so far that will enable the lower reaches of western rivers to achieve the salt balance necessary for the long-term health of western agriculture, on which the entire U.S., and indeed the world, has much dependence. Unless the lower rivers are allowed to reassert their natural function as exporters of salt to the ocean, today's productive lands will eventually become salt-encrusted and barren.