3.11 Water harvesting

Contents - Previous - Next

Professor, School of Natural Renewable Resources Tucson - Arizona


Water harvesting is a technique of developing surface water resources that can be used in dry regions to provide water for livestock, for domestic use, and for agroforestry and small scale subsistence farming.

Water harvesting systems may be defined as artificial methods whereby precipitation can be collected and stored until it is beneficially used. The system includes: 1) a catchment area, usually prepared in some manner to improve run off efficiency and 2) a storage facility for the harvested water, unless the water is to be immediately concentrated in the soil profile of a smaller area for growing drought-hardy plants. A water distribution scheme is also required for the systems devoted to subsistence farming for irrigation during dry periods.

The earliest evidence of the use of water harvesting are the well publicized systems used by the people of the Negev Desert perhaps 4000 years ago. Hillsides were cleared of vegetation and smoothed in order to provide as much run off as possible; the water was then channelled in contour ditches to agricultural fields and/or to cisterns. By the time the Roman Empire extended into the region, this method of farming encompassed more than 250,000 hectares and had become quite sophisticated.

In the New World, about 400-700 years ago, people living in North America in what is now the state of Colorado in the United States, and those living in what is now Peru in South America employed relatively simple methods of water harvesting for irrigation (0 'Bryan, Cooley and Winter, 1969).

Although the practice of collecting water from rooftops is a very ancient type of water harvesting, practiced since the earliest of times up to the present, the harvesting of rain water for agriculture and ranching was pioneered in Australia during the 1920's. Galvanized rooflike structures were built near the ground surface solely for the purpose of water harvesting (Kenyon, 1929).

A renewed interest in the technology of water harvesting occurred in the 1950's in Israel, Australia and the United States. In Australia, 'roaded' catchments based on the concept of compacted earth were constructed over more than 2,000 hectares in order to collect water for agricultural purposes. In the United States, at about the same time, catchments were constructed primarily of sheet metal for watering livestock. Also during this period, experimentation was undertaken with plastic and artificial rubber membranes for the construction of both catchments and reservoirs (Lauritzen, 1960). Since the 1950's the technology of water harvesting has developed variety and sophistication through experimentation and demonstration projects. Not all have been successful but some have worked very well.

The technology and experience gained has now reached the point where some form of water harvesting system can be designed to fit within the physical, climatological and economic constraints of almost any dry region of the world.


The development of water resources in arid lands at present and in the recent past has been concentrated on large-scale irrigation projects. Some have water supplied by river systems such as the Indus, Nile, Yellow River and Tigrus and Euphrates. Others depend upon ancient and finite groundwater bodies. However, the availability of modern technology to societies unprepared for it, combined with the common political focus on urgent and short-term goals, has created long-term and irrevocable consequences in many countries.

Presently, primarily because of salinization, as much land is being permanently retired from agriculture along major river systems as there is new land being added. Electrical and engine-driven pumps and improved methods of well construction have greatly increased the capacity of wells. But in many arid lands the absence of effective controls on drilling and pumping has resulted in groundwater mining with the consequences of decreasing yields, deteriorating water quality, saline intrusion and wastage of water.

Large irrigation projects are essential to many national economies in arid lands. They can be effective over the long term if there is not only a technical competence but an informed social and political structure as well. However, these projects provide few direct benefits to the small land holder or nomad who must exist within the constraints of his environment without the benefit of new technology appropriate to his needs.

Water harvesting offers one method of improving the livelihood of these people. It is not a panacea, but depending upon the rainfall regime in a specific area, it can be used to augment existing supplies, improve range utilization and provide food for better nutrition, wood for fuel and, perhaps more importantly, help towards economic stability by reducing the uncertainty of human life in arid ecosystems.

Much of the economy of arid lands depends upon livestock; so it is not surprising that most of the work that has been accomplished in water harvesting has been aimed at providing water for livestock. Many of the systems devised have been very effective.

Many rangeland areas are overgrazed around water sources, whereas large portions of the land and forage are unused because of inadequate water. A uniform distribution of watering spots could allow these lands to be used for maximum benefit. Water harvesting can provide an adequate distribution of watering spots in most situations, and, with water, even dead forage can be utilized. However, there is an inherent danger that, in the absence of grazing controls, each new watering source can lead to increased herd size, overgrazing and a new nucleus of expanding desert. Just as with other forms of water resource development, in order to be lastingly effective, water harvesting projects and particularly those designed for livestock production must be preceeded by comprehensive planning and controls with provisions for perpetual operational support.

The purpose of water harvesting is to either augment existing water supplies or to provide water where other sources are either not available or would entail prohibitive developmental costs. The aim is to provide this water in sufficient quantity and of a suitable quality for the intended use.

Arid zones are often described as pulse-reserve systems that turn on with water, store material for the dry interval and then shut down until the next water event. The natural ecosystem has developed under these constraints but the strain on human populations is considerable, particularly on sedentary populations. The idea of water harvesting is to smooth out the peaks between want and abundance by collecting and storing water for the period of want, and in addition, to work within the arid ecosystem using a renewable resource.


3.1 Configuration and Use

The geometric configuration of water harvesting systems depends upon the topography, the type of catchment treatment, the intended use and personal preference. Microcatchments, strip harvesting, roaded catchments and harvesting aprons are some of the more common types. Illustrations of these systems can be found in Frasier and Myers (1983) and Thames and Fischer (1981).

Microcatchments and strip harvesting can be successful in years of normal or above normal rainfall. However, in dry years most annual crops will fail. They are best suited for agroforestry where drought resistant trees or other drought hardy perennial species are grown. The use of microcatchments involves the preparation of small catchment areas in which one to several plants are grown on the low side. The collection area may range from 20 to 1000 m depending upon the precipitation in the area and plant requirements. The microcatchment procedure may be used in complex terrain or on steep slopes where other water-harvesting techniques may be difficult to install.

Strip farming is a modification of the microcatchment method. Berms are erected on the contour and the area between them prepared to serve as the collection area. Run off occurring between the berms is then concentrated above the downslope berm to irrigate the vegetation planted there. Only very drought hardy plants should be grown with this type of system.

Apron type water harvesting systems are used primarily for livestock, wildlife and domestic water supplies. The catchment area (apron) is treated to obtain a high run off efficienty, unless an existing impermeable surface is used. Gravel covered, asphalt impregnated fiberglass is a common treatment in the United States. The systems are designed for minimum maintenance and must be fenced. A storage tank with evaporation control is required with the necessary pipes and valves to conduct the water to drinking troughs or to households. The apron type system is the simplest to design. As a first approximation for the size of apron required, the following equation is helpful:

A = 1.13 U/ p where:

A = catchment area mē
U = annual water requirement litres
P = average annual precipitation mm

Roaded catchments are well suited to agroforestry and for growing high value horticultural crops (e.g. fruit and nut trees, grapes etc). They are also suitable for providing livestock water. These catchments are best adapted to very gently sloping ground.

A roaded catchment consists of parallel rows of drainages 100 m or less long and spaced 15 to 18 meters apart. Trees or horticultural species are planted in the drainages. Grapes have proven an excellent crop in Arizona, U.S.

The areas between drainages are shaped much like high-crowned roads to serve as catchments. Side slopes of the catchment roads and longitudinal slopes of the drainages should be no more than about 2% to prevent erosion. The catchments are cleared of vegetation and smoothed and are treated to reduce infiltration. Sodium chloride has been an effective treatment in Arizona. If high value, horticultural crops are to be grown, water storage is necessary to provide supplemental irrigation water. This is easily accomplished by diverting excess water from the drainages into a storage facility.

Water harvesting for agriculture requires a more complex system than do the other systems. The size of the catchment area in relation to that of the agricultural area must be balanced against crop demands topography, provided there is a level area to farm and care is taken with catchment construction to provide low slopes, or, in steep terrain, short slope lenghts broken by diversions.

Compartmental reservoirs consisting of three storages is recommended. Electric, engine or wind driven pumps are usually necessary to transfer water between ponds and to drive the irrigation system. In some steep terrains, gravity systems might be possible. Since fairly large quantities of water are usually required, efficient catchments are also necessary. But treatments can be expensive. Sodium chloride is one of the least expensive treatments in some areas and is effective in locations where the soil has a sufficient quantity (about 10% or more) of expanding clays.

Water harvesting for subsistence farming holds much promise in alleviating the food and nutrition problems of arid lands. Successful systems have been installed and shown great promise. However, training in new techniques is necessary, more information is needed from around the world on crop phenology and water requirements, and additional demonstration projects need to be installed under varying economic, social and climatological conditions in order to develop universal prescriptions.

3.2 Catchment Areas

The catchment area of any water harvesting system is an area that is reasonably impermeable to water which can be used to produce run off. Some examples are: l) natural surfaces such as rock outcrops, 2) surfaces developed for other purposes such as paved highways, aircraft runways, or rooftops, 3) surfaces prepared with minimal cost and effort such as those cleared of vegetation or rocks and smoothed, or both smoothed and compacted, 4) surfaces treated chemically with sodium salts, silicones latex or oils, 5) surfaces covered with asphalt, concrete, butyl rubber, metal foi, plastic, tarpaper or sheet metal.

The particular surface treatment selected will depend upon the availability of materials and labor and budget allotments. Generally, the greater the run off efficiency and life of treatment the greater the cost. At one end of the scale, simple smoothing and compact on of non porus soils is effective but requires annual maintenance. On the other end of the scale is asphalt impregnated fiberglass covered with gravel which may last 20 years or more (see Thames and Fischer, 1981).

Myers (1975) lists some desirable characteristics of catchment treatments as:

  1. Run off from the surface must be non-toxic to man and animals.
  2. The surface should be smooth and impermeable to water.
  3. The treatment should have high resistance to weathering and should not deteriorate because of chemical or physical process.
  4. The treatment need not have great mechanical strength but should be able to resist damage by hail or intense rainfall, wind, occasional animal traffic, moderate flow of water, plant growth, insects, birds and burrowing animals.
  5. The material should be inexpensive on an annual cost basis and should require minimum site preparation.
  6. Maintenance should be simple.

Obviously, there is no single treatment that would have all of these characteristics. Some trade off is necessary, but lowest cost over the long term is often the overgrazing objective. Estimates of the costs of water using various catchment treatments in the U.S. are given in Table 1.

Selection of the lowest cost catchment can sometimes be a mistake. For example, simple smoothed catchments produce water at very low cost, but they do not provide run off from small storms that characterize rainfall periods in many arid zones. A large and expensive structure might have to be built to store water for use during the period when no run off occurs. The cost of a simple smoothed catchment, plus the large storage required could be greater than the cost of a more expensive catchment which provides run off from small storms. A procedure has been developed that can be used for livestock and domestic water systems to determine the lowest cost for any given combination of unit construction costs of catchment and storage, catchment efficiency, water demand schedules and precipitation patterns (Frasier and Myers, 1983).

Table 1. Water costs for various water-harvesting treatments (Frasier, 1975)

Treatment Run-off (%) Estimated life of treatment (yr) Initial treatment cost ha ($ US) Annual amortized cost ha-1
($ US)
Water cost in a 500 mm rainfall zone ($ US MI-1)
Rock outcropping 20-40 20-30 <120 <240 50-100
Land clearing 20-30 5-10 120-240 <120 67-100
Soil smoothing 25-35 5-10 600-840 120-240 55-158
Sodium dispersant 40-70 3-5 840-1440 120-240 ?0-100
Silicone water 50-80 3-5 1440-2160 240-480 50-158 repellents
Paraffin wax 60-90 5-8 3600-4800 600-1200 110-330
Concrete 60-80 20 24000-60000 2040-5280 420-1450
Gravel covered 70-80 10-20 6000-8400 480-1200 100-282 membranes
Asphalt fiberglass 85-95 5-10 12000-24000 1680-5760 290-1110
Artificial rubber 90-100 10-15 24000-36000 2520-4920 415-890
Sheet metal 90-100 20 24000-36000 2040-3120 335-570

3.3 Water Storage

There are basically three types of storages: 1) the soil profile, 2) excavated ponds and 3) tank or cistern containers. The early, ancient water harvesting systems were simple arrangements where water was directed from hillsides onto cultivated areas with the idea of immediately storing the water in the soil for plant use. The problem was whether sufficient water could be stored to offset a prolonged drought. However, the method is still valid today and can be used in agroforestry to grow drought resistant varieties of trees and other economic plants.

Excavated ponds are often the only economical means of storing the large quantities of water which are needed for run off farming. But evaporation and seepage are serious problems. Evaporation suppression on water impoundments is still in the experimental stage. Surface area reduction, reflective methods, surface films, mechanical covers, floating styrofoam balls, and even empty, plastic film canisters have all been used. Most have been effective to some extent, but a simple economical method has yet to be developed. Surface films are generally not economical on small impoundments. Reflective films are generally not economical on small impoundments. Reflective methods (beads or dyes floating on the surface) are ineffective in windy conditions. Some types of floating mechanical covers and floats have worked well in experimental situations, but most are short-lived and all are expensive.

Two inexpenisve methods for reducing evaporative surface area that have been effective in some situations are the compartmented reservoir system (Cluff, 1981) and sand or rock filled reservoirs. The compartmented reservoir is a system of pumping water between two or three ponds so as to minimize total surface area. Sand or rock reservoirs are structures either deliberately filled with rock or designed to capture gravel and sand alluvium, as well as water. The dam is sometimes built in stages so that only coarse sediments are deposited. However, about 50% of the capacity will be lost after filling. A well is sunk behind or through the dam to draw out the stored water. After the water level has sunk to about one meter below the surface of the fill, evaporation effectively ceases.

In a water balance study of three ponds in southern Arizona, seepage acounted for 58 to 87% of the total annual water loss. Seepage control is simpler and less expensive than evaporation control. Chemical dispersing agents, bentonite, soil cement, membrane liners, asphalt, salt and simple compaction have all been used successfully to seal impoundments. Treatment method and application rate is determined by soil type, purpose of the impoundment, severity of wetting and drying cycles, and economics.

Tanks or cisterns can be used effectively for livestock watering and domestic supplies. Seepage and evaporation suppression are less difficult and less expensive. Any container capable of holding water is a potential water storage facility. External water storages are a necessary component for drinking water supply systems, and may also be a part of a run off farming system where the water is applied to the cropped area by some form of irrigation system. In many water-harvesting systems, the storage and water distribution facility is the most expensive single item, and may represent up to 50% of the total cost.

There is an almost infinite number of types, shapes, and sizes of wooden and reinforced plastic storages. Costs and availability are primary factors for determining the potential suitability of these storages. One common type of storage is a steel tank with vertical walls and with a concrete or other type of impermeable bottom. Storages constructed from concrete and plaster are relatively inexpensive, but require considerable hand labor. Roofs over the storages are a common technique of suppressing evaporation although they are usually expensive. Floating covers of low density synthetic foam rubber are an effective means of controlling evaporation from vertical walled, open topped storages, and are not expensive.


The particular strategy to be taken in developing a water harvesting system depends upon a number of constraints. Some of the more important that must be considered include:

  1. The need for acceptance by the local community whether the system is to be used for livestock, domestic purposes, agroforestry, or farming.
  2. The quantity and quality of water required to meet the need.
  3. The availability of alternative, less expensive sources that could be developed.
  4. The amount, seasonal distribution and variability of rainfall.
  5. The materials, labor and machinery available, suitable for installing a water harvesting system are within budgetary limitations.
  6. The provisions for maintenance.


4.1 Need and Acceptance

The need for water resource development in almost all arid lands is patent. The problem is to reconcile this need with what may be accomplished. Furthermore, the user must be fully aware of the potential benefits as well as the limitations of a proposed system. Rural people in arid lands are understandably conservative. They cannot take chances with their survival on unproven methods. But the ultimate success of a water harvesting project depends on the full support of the user for proper operation and maintenance. The user must believe that the system is the best for his needs.

In areas where the concepts of water harvesting and run off farming are not fully accepted, the first system installed must be constructed from materials which require minimum maintenance and have maximum effectiveness. m e extra cost encountered in building a substantial system may be necessary to insure acceptance of the concept by the user. Once the concepts are accepted, it is often possible to utilize lower cost materials and techniques on subsequent units even though these lower cost systems may have a greater chance of failure or require additional effort from the user. If the user has been shown the ideas are valid, he is more likely to expend the extra effort to operate and maintain the system properly.

4.2 Water Quantity

With the exception of water harvesting systems that do not require water storage facilities (e.g. microcatchments and strip catchments), other systems must be designed to supply the quantity of water needed at the times it is needed.

For livestock the need depends upon the time of the year the system is expected to supply water, and this is determined by the grazing systems employed and the monthly distribution of rainfall. For most installations, many combinations of catchment and storage sizes will provide the desired quantities of water. The problem is to find the most economical combination. A method of doing this can be found in Frasier and Myers (1983).

For domestic supplies, people require from 20 to 40 liters per day for cooking, drinking and washing, depending upon how they conserve water. me system should be designed to account for this minimum requirement in addition to any losses that would occur by evaporation or seepage from storage.

Water harvesting systems designed for agriculture are more difficult to design. There is little information on the minimum total water requirements of agricultural crops although consumptive use data abound for crops abundantly supplied with water. Of equal importance to the total water requirement is the timing of the water needs. Erie, French, Bucks and Harris (1982) and Doorenbos and Kassam (1979) provide the most useful information to date on consumptive use. The seasonal pattern of water use from initial establishment to harvest must be satisfied by the design of the water harvesting system. This type of information has been developed for many crops under extensive irrigation practices and is probably higher than needed for many run off farming applications. Relationships of this nature must be developed or estimated for proper design of agricultural water harvesting systems and matched with the water supply to determine frequency and amount of irrigation.

4.3 Water Quality

Water collected from a catchment can contain organisms and water-soluble impurities from windblown dust deposited on the surface or chemical pollutants directly from the treatment (e.g. salt, silicone, tars, oils, etc) or from weathering by-products created by deterioration of the treatment materials. Asphalt and certain plastics are deteriorated by sunlight and heat into water soluble products. Animals feces can be a source of bacteria and virus contamination if the area is not protected against entrance. However, the quality of water from most surface treatments is usually adequate for livestock, but filters are needed in most cases if the water is for human consumption. None of the surface treatments (even sodium treatment) that have been successful appear to affect plant production.

4.4 Alternative Sources of Water

Although water harvesting is not necessarily an expensive means of developing water, there may be other sources of water at or near a particular site that can be developed more cheaply or that would insure more reliable supplies. For instance, undeveloped or underdeveloped springs, a shallow groundwater table which may receive reliable recharge along a mountain front, perched water that might be developed with horizontal wells may offer possibilities. These types of sources should be thoroughly investigated before embarking on a project. All potential water sources should be evaluated with respect to number, location, yield, dependability, and quality. If there are other convenient sources that can be developed economically but are deficient in yield or dependability they may be used to supplement a water harvesting system. If water quality if poor (high salt content, for example) harvested water might provide sufficient dilution for the purpose of the need.

Incorporating intermittent water sources, such as ephemeral flows in waddles, into the total water supply system can in some cases permit the installation of a smaller water harvesting facility. The harvested water can be saved for periods when the ephemeral sources are insufficient or dry up entirely. This type of combination not only saves time and money but can be of major importance during extended drought periods.

4.5 Precipitation

The amount, timing and variability of rain which occurs during a season or year are the key factors that must be evaluated in designing a water harvesting system. Long term daily records are the most desirable. In arid lands, at least 15 to 20 years of record are needed. If there are large variations between years, data from the two wettest years should be eliminated. If sufficient long-term data are available, stochastic methods can be used to determine the probabilities of extreme periods (Duckstein and Fogel, 1972). Mean annual rainfall is not a very good indicator of available water because there will be more years with rainfall less than the mean than there will be years with rainfall greater than the mean.

To compensate for dry years, the size and efficiency of the catchment areas and storage can be increased. But regardless of the design, there will be risk involved because of the uncertainty of rainfall. The user must decide the amount of risk that can be accepted should there be insufficient rainfall during some periods.

In very general terms, if the annual rainfall is no more than about 150 mm, agroforestry, with plantings of indigenous or proven drought resistant species, using microcatchments or contour strips is suitable. Systems designed for livestock watering can also be used, but will require storage tanks protected from evaporation losses as well as efficient catchment surfaces. In areas with annual rainfall greater than 250 mm and if there are good water storage facilities, farming systems are possible. However, between 250 and 300 mm mean annual rainfall (unless the rainfall period coincides with the growing season) drought resistant crops should be used. At 300 mm and above almost any conventional crop can be grown. In locations with an average annual rainfall less than 50 to 80 mm, water harvesting will probably never be economically feasible.

The final size of the catchment area should be determined by computing a weekly or monthly water budget of collected water versus water requirement to help insure that there are no critical periods when there will be insufficient water.

Smaller systems can frequently be used when the periods of maximum rainfall coincide with periods of maximum use. Larger systems, with large storage capacity are necessary when the periods of greatest precipitation occur after the periods of greatest water needs when it may be necessary to store water for 6 to 9 months.

4.6 Materials and Labor

There is no universally best material for catchment and storage. The cost of alternative water sources and the importance of the water supply determine the costs which can be justified in a system. Ordinarily, the lowest cost of locally available materials are used. Usually, water-harvesting systems for supplying drinking water are constructed from materials which are more costly than can be economically justified for run off-farming applications. One must balance the cost of materials to the cost of labor. Some materials and installation techniques are labor intensive, but have a relatively low capital cost. Other materials may be higher in initial cost, but require minimum labor for proper construction.

4.7 Maintenance

Failure to provide for maintenance will result in early failure of the system. Failure to repair minor damage can result in complete destruction of the system. A maintenance programme must be followed even when the water collected is not being used. Some types of catchment treatments and storages require more frequent and intense maintenance than others. However, most water harvesting systems can be adequately maintained with once a year inspection and repair visits if any problems that appear at other times are immediately repaired. All elements of the system should be inspected, including checks for leaks in any valves, pipes or water storages, as well as the condition of the catchment, weed, animal and insect control. Inspections and repair usually require only a few hours, but are as essential to the system as the initial installation.


Numerous water harvesting systems have already been installed in various arid lands of the world, and many more are being implemented. Although much of the work is in the experimental stages and it is perhaps too soon to evaluate the ultimate results. Most of the installations have reported success. Only a few representative cases are cited here.


There are a number of water harvesting projects in Mexico varying from village water supply systems to run off farming. Two are given as examples.

Nuevo Leon

This system is composed of a set of 248 microcatchments of 70 m each for growing pistachio trees. Each tree is planted in a microcatchment. A major objective is to evaluate different catchment treatments. They are 1) compacted soil, 2) soda ash (Na2CO3), 3) road oil, 4) gravel covered polyethylene, 5) gravel covered asphalt and 6) smoothed soil. Soil moisture was monitored under each tree at depths of 15, 35, and 55 cm. Also included were tests of various soil coverings immediately around the trees to reduce evaporation (Velasco and Carmona, 1980).

This is a long term experimental demonstration project because of the slow growth of the trees. From preliminary observation it appears that soil eroded from the salt treated catchment is deposited around the tree. The deposited soil reduced infiltration before it could infiltrate into the soil.

Techo cuenca

A water harvesting system was constructed in 1975 to augment the domestic water supplies for 30 families (about 180 people) for the village of Laguntia y Ranchos Nuevos in the state of Nuevo Leon. The system consists of an inverted galvinized metal roof (269m) supported on a steel framework above an 80,000 liter steel tank. Village labor for constructing the system was 36% of the total cost of 143,000 1975 pesas. The system provides drinking water to the entire village for 4 and 1/2 months of the year, based on an allotment of 20 liters per day per family at about one third the cost that is incurred in hauling water (Carmona and Velasco, 1981).


Trials with okra were carried out to study the effects of different water conservation methods on plant growth and yield. It was found that out of 8 methods tried, the highest yield was obtained with plants in 2 rows 45 cm apart in furrows running east/west with ridges 60 cm apart. The beneficial effect of this method was attributed to the rainwater drainage from the ridges to the furrows (Vashistha et al., 1980).


Asphalt coverings were applied aw 1 liter/mē to slopes above the experimental run off plots and terraces (2 m wide) on a hillside near Tehran, Iran for harvesting rainwater for growing trees. Rainfall run off over a 5-year period was substantial. Robinia pseudoacacia, Cupressus arizonica and Fraxinus rotundifolia were planted. The data showed that there was a significant increase in height, diameter of stem and crown development of plants in asphalt treated plots as compared to controls (Mehdizadeh et al., 1978).


The Pakistan Forest Institute installed microcatchments on more than 40 hectares in 1982 in an area that receives 250 to 300 mm of annual rainfall. The purpose of the project was to establish forest plantations. Species of Acacia, Prosopis, Tecoma and Parkinsonia were planted. Survival of species planted as control, without microcatchments, was only about 10%. Survival of species planted in microcatchments was 80 to 90%. Acacia tortilis had the highest survival rate and grew more than 4 feet per year.


Australia was among the first of the western countries to install operational water harvesting systems. The systems were designed to provide water for livestock and domestic needs. Much of this work was done in southwest Western Australia.

Harvesting from natural surfaces (250 mm rainfall)

Advantage has been taken of natural outcrops of granite which commonly occur on the crests of broad, sand plain ridges in the area. Run off from the 250 mm of annual rainfall on many of these outcrops is collected by means of low concrete or masonry gutters built around the lower slopes of the rock. The water so collected is conducted along the gutter to a concrete storage tank built on or close to the rock, or to an earth tank sited downslope from the rock. The water from these rock catchments is noted for its excellent quality. The Public Works Department has constructed harvesting systems of this sort on some of the larger outcrops to provide public supplies of potable water.

The Public Works Department of Western Australia initiated in 1948 a programme of construction of Roaded Catchments (Carder, 1970). These catchments consisted of clearing, shaping, and contouring to control length and degree of slope and compacting with the aid of pneumatic rollers. An estimated 2500 roaded catchments have been installed principally to supply water for livestock use. These average approximately two acres in size. There are also 21 roaded catchments totaling 1745 acres (706 hectares) and ranging in size from 30 to 175 acres (12.1 to 70.8 hectares) presently being used to furnish domestic water for small towns in Western Australia (Burdass, 1975).

United States

The Indian reservation of southwestern United States have much in common with the developing countries of arid regions: remoteness, grazing cultures, lack of water and poor economies. A number of water harvesting systems have been tested by various agencies on these reservations. Three are cited here.

The Arizona Strip

The Arizona Strip lies across the Colorado River from the Hualpai Indian Reservation and south of the Arizona-Utah state line. The land is under the jurisdiction of the U.S. Department of the Interior, Bureau of Land Management, and is leased to cattlemen" for grazing by livestock.

Perennial streams and springs are rare, and groundwater is inaccessible due to depth and isolation of perched aquifers. Earthen reservoirs are often used, but are rarely dependable because of high seepage, evaporation losses, and low run off.

Two water harvesting systems were installed in September, 1974 to evaluate the potential of this technique for supplying the necessary animal drinking water. The catchment areas (0.3 and 0.4 hectares) were treated with a refined paraffin wax sprayed on the prepared soil surface. The collected water was stored in a 300,000 liter steel rim, concrete bottom tank with a foam rubber floating cover for controlling evaporation. Total cost (1974) was $ 8,925 and $ 9,150, including labor and miscellaneous items such as fencing and drinking troughs. The systems are maintained by the Bureau of Land Management.

During a drought in 1976-1977, these systems provided the only water supply. All other water sources went dry. Without this water, the permittees would have had to move their livestock. Ranchers have remarked that these systems were as good as, or better than, a spring (Cooley et al., 1978). Since that time, the Bureau of Land Management has installed over 60 more units of various types and treatments, and several local ranchers are installing units on their own. There have been some failures of the later units, but this has not deterred the ranchers from accepting this method of water supply. This attitude has been developed because it was demonstrated that, with proper installation and maintenance, water harvesting can be effective method of water supply.

Black Mesa

The Black Mesa water-harvesting facility is located on the Navajo Indian Reservation in northeastern Arizona on displaced overburden from a strip coal mine. This is one of the largest systems in the United States. It consists of (1) three water storage ponds with a total capacity of slightly over 3 million litres, (2) two leveled agricultural terraces of 1 hectare each, (3) a "road" catchment for an orchard of 0.5 hectares, 84) a fiberglass asphalt-gravel catchment of 3.2 hectares, and a 2.9 hectare salt-treated catchment. A pump system is used to transfer the collected water between ponds and to lift the water to irrigate the crop areas. Initially, flood irrigation was used, but later, a prinkler system was installed.

Annual crops grown and evaluated were beets, onions, turnips, potatoes, chard, lettuce, cabbage, tomatoes, squash, beans, pumpkins, melons, manger and corn. All crops, except tomatoes, did well, with some producing at levels above the national average. The value of the corn produced was the lowest of all crops. This was not unexpected, but corn is a traditional food in the area, and was planted for social reasons. Fruit trees had never been grown in the area before. All trees were growing well after 3 years, but it was too soon to determine the potential production of the varieties planted. The water-harvesting project yielded about $ 1,700 net revenues per cultivated hectare (1981). Revenues are expected to increase when the orchards reach maturity (Themes and Cluff, 1982).


The village of Shungopovi is located upon Second Mesa, on the Hopi Indian Reservation in northern Arizona. The village, built on top of a sandstone rock mesa, had no source of water. From the time of first establishment, the villagers had carried water up from the valley, initially on foot and, later on, on the backs of burros. In the early 1930's, a small water-harvesting system was installed to partially relieve the water shortage of the village. An area of approximately 1/3 hectare was set aside, cleared, and the loose soil removed to expose the sandstone bedrock. Below the area, a deep cistern was hewed into the rock, and a concrete roof constructed. This system was a functional part of the village water supply for about 30 years, at which time, a community well, pump on the valley floor, and water distribution system was installed Chiarella and Beck (1974).


Researchers in Israel were the first to experiment in the application of new techniques to water harvesting. They have found various methods effective in increasing run off from land surfaces, either singly or in combination, including land smoothing and compaction, formation of sodic crusts spraying applications of various asphaltic materials. Among the asphaltic formulations, heavy fuel oil diluted with kerosene proved to be both effective and economical.

They found area ratios between contributing and receiving areas on the order of 3:1 to 6:1 in a rainfall zone of 200 to 250 mm. The accretion of water received in the planting zones of experimental plots provided soil moisture equivalent to the entire normal winter rainfall in the mediterranean climatic zone of the country, in northern Israel, where unirrigated orchards provide a livelihood for an appreciable number of farmers (Hillel, 1967).


The economics of water harvesting have never been fully documented, particularly over the long term, even in Israel, Australia and the United States where most of the modern technology of water harvesting has been developed. The need is for a better understanding of the economic viability of different methods in different economic environments, particularly in those of the developing countries.

A primary function of water harvesting system designed for livestock watering, domestic supplies, and run off farming, in addition to providing more useable water, is to smooth out the variation in natural rainfall by providing water during interrain periods. Never-the-less, the reliability of a system and the degree of risk associated with a system depends upon the reliability of rainfall and water demand. Thus, rainfall records are essential to design and operation. They are often insufficient in the developing countries and especially in the less dense populated arid zones of those countries. Furthermore, because of the high variation in rainfall from year to year in arid zones, records are needed from longer historical periods than would be required for similar analysis in temperate zones. More adequate coverage by weather stations and maintenance of daily records should be encouraged by international development agencies.

Information is badly needed on the minimum water requirements of agricultural crops in different climatic regions of the arid zones in order to reduce the risk in designing water harvesting systems for agriculture.

Evaporation from storage is a serious problem. Two meters or more water loss from free water surfaces is not uncommon in arid lands. As yet, except for tanks and cisterns, no economical and effective method has been developed to prevent evaporation from storage ponds despite the great variety of materials tried.

An important technical-research need is to reduce the costs of catchment treatment and to make the treatment practical for a wider variety of soils and situations. Industry is constantly formulating new materials which should be continually monitored and evaluated for use both for catchment treatment and in evaporation control.

The technology of water harvesting will largely be advanced through empirical experimentation. Thus, more information and hard data are needed from a wider range of climatic, soil, economic and social conditions. Information, of course, is needed on successful projects, but just as valuable and much more difficult to obtain is information on failures and the reasons thereof. Efforts should be made toward establishing an accessible international outlet for such information.

As more information becomes available, more work is needed on the modeling and synthesis of water harvesting systems. Until adequate prediction schemes are developed the design of water harvesting systems will remain dependent upon the experience of a limited number of experts who have learned the hard way.

Although there are a large number of techniques available and many variations within the techniques, new ideas are needed. for example, the use of plastic greenhouses, where water is continuously recycled, has been developed to the extent that it could be coupled with a water harvesting system to produce high value crops. This has not yet been done.

Quantitative information on the quality of water collected by water harvesting systems is limited. The information is needed for various catchment treatments to compare with safe standards for livestock use, human consumption and for growing plants. Water quality analysis should be an integral part of any water harvesting project.


Water harvesting offers a method of effectively developing the scarce water resources of arid regions. As contrasted to the development of groundwater, which is usually a finite water resource in arid zones, the method allows use of the renewable rainfall which occurs, even though in limited amounts, year in and year out. It is also a relatively inexpensive method of water supply that can be adapted to the resources and needs of the rural poor. It is necessarily small scale, and as such it can provide stability and improve the quality of life in small rural communities and that of small land holders who are several stages removed from the benefits of large scale development projects. Despite this, water harvesting is not a panacea. It involves some risk, dependent upon the vagaries of climate. New skills, though simple, are required, maintenance is a constant necessity, and good design is imperative.

There is no universally "best" system of water harvesting. However, there will be some type of system that can be designed to best fit within the constraints of a given location. Each site has unique characteristics that will influence the design of the most optimum system. All factors, technical, social, physical and economic must be considered. During the past two decades, there have been many water harvesting systems constructed and evaluated at a number of different places in the world. Some of the systems have been outstanding successes, while others were complete failures. Some of the systems failed, despite extensive effort, because of poor design or the materials used. Other systems failed despite good design and proper materials because social factors were not integrated into the systems. These systems failed because of poor communication and lack of commitment by the local people both in planning and financing the projects. Unfortunately, one failure in a traditionally conservative social system, as are many rural societies in arid lands, can offset the effects of 10 successes. A successful system must be:

  1. Technically sound, properly designed and maintained.
  2. Economically feasible for the resources of the user.
  3. Capable of being integrated into the social traditions and abilities of the users.

Much has been learned over the past two or three decades. Much more remains to be learned, but sufficient knowledge and experience has now accumulated to put into operation water harvesting projects throughout the arid lands of the world. Empirical information and documentation is needed from successes as well as failures on which to build a more exact technology.


Burdass, W.J. Water harvesting for Livestock in Western Australia. (1975) In: Proceedings of the Water Harvesting Symposium, Phoenix, Arizona, 26-28 March, 1974. Editor G.W. Frasier, ARS-W-22, U.S. Department of Agriculture, Agricultural Research Service, pages 8-26.

Carder, D.J. Roaded Catchments - A Method of Increasing Run Off into (1970) Dams and Reservoirs. Technical Bulletin No. 5. Western Australian Department of Agriculture.

Carmona-Ruiz, G., and H.A. Velasco-Molina Rain Harvesting for Human and (1981) Livestock Consumption in the Semidesert High Plains of Mexico. In: Rainfall Collection for Agriculture in Arid and Semiarid Regions. Editors G.R. Dutt, C.F. Hutchinson, and M. Anaya Garduno. Proceedings of A Workshop, university of Arizona, Tucson, 10-12 September, 1980. Commonwealth Agriculture Bureau, Farnham House, Farnham Royal, Slough, United Kingdom, pages 77-81.

Chiarella, J.V., and W.H. Beck Water Harvesting Catchments on Indian (1975) Lands in the Southwest. In: Proceedings of the Water Harvesting Symposium, Phoenix, Arizona, 26-28 March, 1974. Editor G.W. Frasier, ARS-W-22, U.S. Department of Agriculture, Agricultural Research Service, pages 104-114.

Cluff, C.B. "Surface Storge for Water Harvesting Agrisystems", in (1981) Rainfall Collection for Agriculture in Arid and Semi Arid Regions, ed. G.R. Dutt, C.F. Hutchinson, and M. Anaya Garduno, Commonwealth Agricultural Bureau, Farnham House, Slough, UK.

Cooley, K.R., G.W. Frasier, and K.R. Drew Water Harvesting: An Aid to (1978) Range Management. In: Proceedings of the first International Rangeland Congress. Editor D.N. Hyder, Denver, Colorado, 14-18 August, 1978, 1978, Society for Range Management, Denver, Colorado. pp. 292-294.

Doorenbos, J. and A.H. Kassam Yield Response to Water. FAO Irrigation (1979) and Drainage Paper, Number 33. United Nations Food and Agriculture Organization. Rome. Italy.

Duckstein, L. and M.M. Fogel A Stochastic Model of run off Producing (1972) Rainfall for Summer Type Storms. Water Resources Vol. 8, No. 2, pages 410-421.

Erie, L.J., O.F. French, D.A. Bucks, and K. Harris Consumptive Use of (1982) Water by Major Crops in the Southwestern United States. U.S. Department of Agriculture, Agricultural Research Service, Conservation Research Report Number 29, Washington, D.C., 40 pages.

Frasier, G.W. Water Harvesting: A Source of Livestock Water. Journal (1975) of Range Management, 28, 429-433. Frasier, Gary and L. Myers Handbook of Water Harvesting. USDA, ARS (1983) Handbook No. 600, pp. 45.

Hillel, Daniel Run off Inducement in Arid Lands. Final Tech. Report. (1967) U.S. Department of Agriculture, Pro]. A 10-SWC-36. Renovot. 142 pp. Kenyon, A.S. The "Ironclad" or Artificial Catchment. Victoria Dept. (1929) Agr. Jour. 27:86-91.

Lauritzen, C.W. Ground Covers-for Collecting Precipitation. Farm and (1960) Home Science 2:66-67, 87.

Mehdizadeh, P. Howsar, A. Vaziri, and L. Boersme. Water Harvesting for (1978) Afforestation, Efficiency and Life Span of Asphalt Cover 11, Survival and Growth of Trees. Soil Science Society of America Journal 1978. 42, 4, 644-657.

Myers, Lloyd E. Water Harvesting - 2000 B.C. to 1974 A.D. In: (1975) Proceedings of the Water Harvesting Symposium, Phoenix, Arizona, 26-28 March, 1974. Editor G.W. Frasier, ARS-W-22, U.S. Department of Agriculture, Agricultural Research Service, pages 1-7.

O'Bryan, D., Colley, M.E., and Winter, T.C. Water, Population Pressure, (1969) and Ancient Indian Migrations. Bulletin 10, International Hydrologic Decade. June 1969. pp. 438-442.

Thames, J.L., and C.B. Cluff Water Harvesting on Surface Mine Spoils. (1982) Paper presented at the Wisconsin Mined-Land Rehabilitation Research Workshop, Ft. Collins, Colorado, 1-11 June, 1982, 4 pages.

Thames, J.L. and J.N. Fischer Management of Water Resources in Arid Lands (1981) in Arid Land Ecosystems: Structure Functioning and Management, Vol. 2. Editors Goodall, Perry and Holmes, Cambridge University Press, pp. 519-547.

Vashistha, R.N., M.L. Pandita, and R. R. Batra Water Harvesting Studies (1980) under Rainfed Condition in Relation to Growth and Yield of Okra. Haryana Journal of Horticulture Science 9, 314:188-191.

Velasco-Molina, H.A., and M.C.G. Carmona-Ruiz Cosech de Agua de Lluvia (1980) Para Consumo Humano, Consumo Pecuario y Agricultura de Secano Primer Reporte de Evaluacion, Comision Nacional de Las Zonas Aridas, Instituto Technologico y de Estudios Superiores des Monterrey, 33 pages.

Contents - Previous - Next