3.10 Watershed management in arid-zones

Contents - Previous - Next

Professor, School of Natural Renewable Resources
Tucson - Arizona


The water relationships of arid zones are perhaps more critical to a greater number of people on earth than those of more humid regions. Water is always in critical balance with arid ecosystems and this balance is presently being upset by man and his animals at alarming rates over vast acreages of the earth.

Excessive livestock grazing, primitive agriculture and fuelwood cutting have depleted the naturally sparse vegetation on arid land watersheds in many countries to such an extent that the flow of rivers which they supply have become flashy with high sediment yields during the rainy seasons and with a trickle or no flow at all during the dry seasons. The results have been damage downstream to prime agricultural land, irrigation works, increased flooding of valleys and human settlements, the siltation of reservoirs and destruction of other civil works. Streamflow becomes insufficient during dry periods for the dilution of disease-carrying pollutants, the maintenance of irrigation works and for urban and industrial needs.

Of perhaps greater significance is the loss, through erosion, of the soil reservoir which not only is the principal means of controlling the discharge of water from watersheds upstream but more importantly, is also the basis for the production of the renewable resources on the watersheds. As soil is removed by water and wind from lands already in fragile ecological balance, there results a spiralling decrease in productivity as the resource base is progressively depleted, sometimes to the point of irreversibility, as is presently occurring in many dry regions of the earth.

As long as current land use patterns on arid zone watersheds continue, the livelihood of tens of millions will, at best, remain at its current dismal level. At worst,-and most probably, a prolonged drought will mercilessly rebalance the number of people with the available resources. This calamity is already taking place in many countries and it may be only a preview of the future for a much larger number of countries in the dry regions of the world.

Dry regions cover about one-third of the earth's land surface, and slightly over half of this area is inhabited by more than 850 million people. The remainder is climatically so arid and unproductive that it cannot adequately support human life. But the degradation of the land and water resources of arid zone watersheds by human activities is turning potentially productive dry lands into unproductive deserts in Asia, Africa and Latin America. This process is called desertification. It has been estimated that a collective area larger than Brazil, with rainfall above the level classified as semi-arid, has been degraded to desert-like conditions. This does not take into account the far greater degradation that is taking place within the potentially productive semi-arid zones.

Present land use patterns on arid zone watersheds must be reshaped in order that delicate water relations are not pushed beyond their limits. As the number of people and animals living in the arid zones climbs and the quality of the land on which they must live simultaneously declines, the impact will be global unless solutions are implemented.


As arid zones come under ever increasing pressures of human use, the ecological interrelationships between people, plants, animals and water have important implications for the role of watershed management in the formulation of policies that would lead to wise development and sustained productivity for the improvement of human welfare. This applies to the welfare of people living upstream on the watersheds who must deal with the severe natural constraints of arid ecosystems as well as to the welfare of people living downstream who are dependent upon reliable supplies of water for their livelihood.

Arid lands are usually characterized by abundant sunlight and mineral rich soils; with water, they can become the most productive lands on earth. It is not surprising that modern civilization has its origin in those arid regions where, with organization, water could be readily developed for irrigation on large scales.

The large-scale irrigation projects of today are more complex. They involve the technologies and economics of large reservoir construction and management, water distribution systems, irrigation schemes and drainage facilities. The role of the watershed, the ultimate source of the water being developed, has often in the past been given only a small part in the design of these systems. Greater attention is being given today to the importance of watershed management to human welfare downstream in sustaining flows during dry periods, controlling sediment yields, reducing flash flood flows, improving water quality and in providing groundwater recharge. Because of the great capital and maintenance costs of large irrigation systems watershed management is increasingly being viewed as simply good business.

Perhaps of greater importance is the role that can be played by watershed management in the lives of people living on the arid land watersheds of the developing countries. Water is the driving variable for the human-ecological interactions on these watersheds. Because of the great and seemingly capricious changes in water abundance, arid zone watersheds can be considered as pulse-reserve systems which turn on with water, store material for the dry interval and then shut down until the next rainfall event or the next rainy season. The indigenous plants and animals of these systems have adapted to this paradigm by various methods of drought evasion. Humans are not so well adapted. In the developing countries people must resort to either nomadism or to a sedentary life which is dependent upon primaitively developed water sources.

The purpose of watershed management in these areas is to understand the human, ecological and hydrological relationships of arid lands and to apply this understanding to rehabilitating depleted areas, to conserving the soil, water and other rational resources and to improving land use for increased and long-term productivity.


Watershed management in arid lands is faced with the complexities of developing and conserving food and fiber resources under extreme and variable climates in the face of ever-increasing human need. Specifically, this requires the use of strategies to improve rangeland for livestock, to develop arid land forestry and to improve dry land agriculture which are in turn dependent upon the development and protection of the basic soil and water resources.

3.1 Strategies for Rangeland Improvement

Livestock grazing is the principal and only economic use of vast areas of arid lands in the developing countries. Unrestricted and excessive grazing has led to severe desertification of the watersheds in many of these areas. The management, control and improvement of grazing lands is, therefore, the most critical aspect of watershed management in most arid lands. The primary strategies that must be considered are grazing management, range improvement and watershed rehabilitation.

Grazing Management: There are numerous systems of grazing recommended around the world, but any successful system must fit the local conditions if it is to be of any value. Too often grazing systems have been instituted which are not based on local needs; thus, they have not been an improvement over traditional methods.

The quantity of forage produced, the seasons it is produced and the dependability of the production are first considerations in developing a strategy. One animal unit, roughly 450 kg of live animal weight, requires about 400 kg of range forage per month. This requirement must be met with less than half of the herbage and browse on the area. At least half of the production should be left on the watershed to maintain plant vigor, favorable microclimate, and an adequate ground cover to protect the soil surface and to return organic matter to the soil.

If the major forage production on the watershed occurs during a single season, then forage must be utilized both to provide forage for livestock and to maintain soil cover on the watershed throughout the year. Livestock grazing capacity may be limited by the number of animals which can be carried during the dormant periods of vegetation. Herbage and browse production should be managed so that there is adequate provision for watershed protection at the end of dormant periods and prior to major growing seasons. Plants may not have enough vigor to recover full production after dry periods, especially if the period extends over more than one season. Stocking therefore, should be flexible to adjust to a variable forage supply.

The variability of the physiological and morphological characteristics of individual plants also influences management. Plants on the same range site may differ in growth characteristics and respond differently to grazing pressure in different seasons. It is important to know, at least for the major plant species, the periods of new shoot development, carbohydrate depletion and accumulation relationships, periods of major new root growth, and time of seed production. These characteristics influence the capability of different plants to benefit from periods of rest and their susceptibility to damage by close grazing during specific seasons.

Livestock distribution on ranges usually can be improved by well-planned water developments, because livestock can be attracted to areas where they would not use the forage because of a lack of water. Controlling livestock access to water and moving animals from one area to another is a means of regulating areas of concentration. Fencing may be used to restrict grazing areas, and trails may be constructed in rough or brushy country to help move livestock from one area to another or to obtain a more even distribution. Salt also may be used to some extent to attract livestock to areas which would otherwise be little used.

Periods of rest should be designed to enable certain plant species to improve in vigor by allowing them to make root growth, to seed, replenish carbohydrate reserves, initiate new shoots or to meet a combination of these requirements which could not be met under a year long grazing period.

Good livestock husbandry and correct management practices combine to provide efficient conversion of the range forage to marketable protein. The test of an efficient range livestock operation could well be the pounds of range forage consumed to produce a pound of marketable meat. In today's world of protein and energy shortages, this is an important contribution of well-managed watersheds.

The Grazing Plan: An inventory of forage available on the watershed is necessary to determine the livestock grazing capacity. Balancing livestock numbers with the forage supply at a level of use which maintains plant vigor and watershed stability is often the major problem of proper grazing management. The level of forage use which will maintain range condition is very dependent on the specific range site, season of use and plants being grazed. A proper level of grazing is usually consistent with long term goals of sustained production but may not be consistent with short term economic goals of private livestock producers. Educational programmes aimed at showing long term goals in terms of society needs and presented to livestock producers are necessary parts of this implementation. Livestock producers should participate in plans for proper stocking if it is to be successful, for they are responsible for the success or failure of grazing management programmes. Political and social pressures, however, may be necessary to accomplish proper levels of grazing on the watersheds of some areas.

Balancing livestock numbers with the forage supply should have flexibility in order to make adjustments to good and bad forage years. This requires provisions for heavy culling and sale of extra animals in drought years. The alternative is to stock at a conservative level, so that the watershed is protected even in drought years. The provision for flexibility is a preferred alternative.

A major objective of any grazing management system should be to even out the use of plants on the range, so that forage production comes from many plants and not just a few species which receive all of the grazing pressure. If one or more major species tend to receive a greater portion of the grazing pressure and are heavily grazed before other species are used, then some management strategy must be used to give these plants a rest. If use cannot be shifted to associated species, the total grazing capacity must be based on the level of use which can be tolerated by the most preferred species.

The uniform distribution of livestock over the range also should be a major objective of any management plan. If topography, vegetation and soils are greatly variable on a watershed, distribution of livestock use may be poor. Livestock tend to use some areas more heavily than others, especially near water, along level bottomlands, ridges, and certain range sites. They may avoid steep slopes, some soils, and dense brush.

The requirements of the livestock must also be considered. In addition to quantity, the forage must also meet nutritional needs in four major categories: protein, energy, phosphorus and carotene. All of these nutrients are generally at a satisfactory level when adequate green forage is available. Browse species, however, usually do not contain adequate energy, and dry grass is low in protein, phosphorus and carotene. If both palatable browse species and grass are present, nutritional requirements can often be met, otherwise some supplementation is necessary. A high nutrition place is especially important at the time of livestock breeding.

In general, the management plan is directed at providing maximum livestock production, yet maintaining or improving watershed stability. For any specific watershed, however, the goals of management should be outlined in detail, including the identification of the range sites on the watershed and even the specific species which are to be the principle plants for management. If a specific area of the watershed is in need of improvement this must be specified as an objective of the management system.

Range Improvement: Manipulating the vegetation on rangeland watersheds by reducing or eliminating undesirable species and increasing the number of desirable species is a useful management strategy. Any decision to manipulate vegetation on a watershed must be based on an understanding of precipitation patterns, the potential productivity of specific sites on the watershed, and estimates of successional or retrogressional trends which may result from different management strategies. With proper grazing management, the condition of vegetation on sites within a watershed may be improved to provide greater cover and an increase in desirable plant species.

In addition to the general improvement which can be accomplished by good management, there are also other methods, but each has constraints for different situations. The watershed manager is faced with the task of selecting among the alternative methods and then conducting trials to determine which are most suitable and/or must be modified for his specific conditions. The methods most frequently used in range improvement are the use of livestock, fire, chemicals, mechanical treatments and reseeding.


Livestock can often be used to influence changes in plant species. Cattle, sheep and goats have different preferences in forage, and these differences may be further accentuated by differences in seasonal preferences. If livestock preferences are known by different seasons and the plant growth requirements are known, a grazing system can often be designed to benefit or reduce the vigor of a given plant species. There are limitations, however. If the undesirable plants are long-lived, such as most woody species, and are not palatable to livestock, they will not be eliminated by grazing once they have become established.


Fire is an inexpensive method of removing undesirable vegetation, but there are inherent dangers of eliminating desirable species as well. In general, shrubby species are more damaged by fire than herbaceous species, but some shrubs may sprout after being burned. A knowledge of how different species respond to fire is necessary to plan for the efficient use of fire in vegetation manipulation. A major cost of burning is the herbage which is required as fuel and must be included in economic evaluations of burning. Also any burning treatment must be coordinated with grazing management programmes so that sufficient fuel is provided both for the fire and for the livestock. Precipitation patterns must also be considered so that burned watershed areas are not devoid of cover when there are high probabilities of rainfall.


Chemical control is effective but expensive and presents a danger of contaminating water supplies. The effectiveness of foliar applications of herbicides are greatly dependent on the phenology of the undesirable species. The herbicide must be applied when the leaves are at a stage when they have sufficient surface area but have not developed a cuticle layer. Granular herbicides applied to the ground at the base of the plants are less sensitive to the time of application but are dependent upon moisture conditions. Effective applications are also dependent upon the species to be controlled and upon soil conditions. Insufficient information is available on herbicidal control to determine the probable effect of herbicides on plants other than the ones to be controlled. Chemical control has the greatest chance of being a useful method when one or more species can be reduced in abundance to allow other native species to increase.


Mechanical control may be superior to herbicides in providing a seedbed if there is a lack of desirable species and particularly if reseeding is necessary. There are numerous methods of mechanical control dependent upon the species to be removed, soil, labour and equipment available, costs and the kind of seedbed which is prepared. Chopping or grubbing individual plants by hand is effective if the plants to be controlled are not too numerous and if cheap labor is available. If heavy machinery is available, bulldozing is an effective control of large woody plants on all sites except steep slopes. Cabling or chaining, accomplished by dragging a heavy chain or cable between two crawler type tractors, is also effective for large plants which will either be uprooted or broken off at the ground surface. Small flexible shrubs and sprouting species unless uprooted are not well controlled by these methods. The soil is disturbed but may not be loosened enough to provide a satisfactory seedbed. Rootplowing is an effective control of woody plants. The plow can be adjusted to cut just below the bud zone, and the tilt can be adjusted to lift plants and prepare a good seedbed. Disking is a suitable method for small brush control and to also prepare a good seedbed.


Grazing areas on many arid zone watersheds have deteriorated to the point where reseeding is necessary. The success of reseeding efforts depends upon rainfall patterns and site conditions. In general, successful seeding requires the following conditions:

  1. Soils and precipitation adequate to provide a reasonable probability for success. Higher precipitation is usually required to establish plants on clay soils than on more sand soils.
  2. Indigenous or adapted species. Selecting adapted ecotypes within a species may also provide an extra measure of success on some sites (see Wright and Streetman, 1960).
  3. Seedbed preparation to control existing vegetation and provide rapid infiltration. A loose, irregular seedbed is often desirable in range seeding operations.
  4. An adequate quantity of seed to insure a successful stand but not to the point of being wasteful. Seeding approximately two to three million viable seeds per hectare is a rough estimate of seeding intensity. For small-seeded species, like Lehmann lovergrass, the seeding rate is less than one kg per hectare.
  5. Proper depth of seed coverage. The smaller the seed, the shallower it should be seeded. Broadcasting of small seed on loose seedbeds is often adequate, but drilling is recommended for large seeds and crusted seedbeds.
  6. Seeding when favorable moisture and temperature combine to give the longest possible period for germination and early growth.
  7. Protection of seedlings until they have become well established. In very dry sites, this may require three or more growing seasons.
  8. Proper grazing after plant establishment. Reseeding is expensive and can only be justified if it provides returns in forage and for watershed protection over many years.

Remedial Cultural Treatments: These treatments are expensive and require machinery or large labor force. The expense is not easily justified unless run off and sediment from the watershed threaten important downstream developments or land reclamation is essential to the survival of people in the area who have no alternative means of livelihood.

The purpose of cultural treatments is to reduce surface run off and soil loss by retaining water on site until a vegetative* cover can become established. Cultural treatments have a limited life expectancy depending upon the amount of run off and sediment produced on the site. Therefore, it is important to evaluate the site for its potential to maintain a vegetative cover after the treatment has lost its effectiveness. Obviously, a management strategy must be formulated and enforced if the treatment is to have a lasting effect. The type of treatment chosen depends upon the run off potential of the site. Often a combination of treatments may be desired, but planting should follow treatment as soon as possible. Some of the more common treatments include contour furrows, contour trenches, pits and basins.

Contour furrows are small ditches 20-30 cm which follow the contour, and are usually constructed with a single blade furrowed plow. The furrows form miniature terraces which hold the water in place until it infiltrates into the soil. Studies have indicated furrows are effective if the spacing is kept to less than 2 metres. At greater spacings there is very little effect except just along the furrows (Barnso, et al 1966 and Wright, 1972). On slopes greater than 5% the soil is usually cast downhill to increase storage capacity. A number of early trials with furrows failed, due largely to the difficulty of following the contour and lack of seeding and follow-up maintenance. It is important that furrows be placed on the contour, otherwise they become drainage ditches which concentrate run off and may cause erosion rather than prevent it. Following the contour, it is a difficult process and one of the disadvantages of the method. The effectiveness of furrows can be increased, and the misalignment of furrows with the contour can be corrected, by constructing crossbars across the furrows at intervals of 1.5 to 10 metres. The furrows then become small basins and if one section should break, water from adjacent sections will be held in place on the watershed.

Contour trenches are usually required on slopes too steep for contour furrows. There are two types: 1) the shallow outside type where the excavated material forms the barrier to overland flow and 2) the deeper inside type where the excavation holds back most of the overland flow. The outside type is suitable for slopes up to 30% and the inside type for steeper slopes (up to 70%). Both types are expensive, usually require machinery and must be designed to handle large storm flows; failure of an upper trench could result in a cascading effect on trenches downslope. Hull (1973) reported that a good stand of vegetation once established holds the soil and prevents sedimentation downslope.

Fallow strips have proven successful on level to gently rolling land (Wright, 1972). The strips are about one metre wide, parallel to the contour and cultivated to destroy unwanted vegetation, loosen the surface, and prepare a seedbed. The original vegetation is left between the strips until the newly planted vegetation has become established. Then new strips are tilled and planted and the process continued until the watershed has been rehabilitated. Conventional tillage equipment can be used for this technique.

Pitting is a technique of digging or gouging shallow depressions into the soil surface to store surface run off. The treatment of extensive areas requires a tractor and pitter. A conventional disk plow can be readily modified into a pitter although a similar disk plow is made specifically for range use. Every alternate disk can be removed and the remaining disks off-set from the center, or rather than off setting, a half moon section can be cut from each of the remaining disks. The disks are arranged to strike the surface at different times, creating an alternate pattern of pits as the plow is pulled across the surface. The furrow is broken by the missing disk or by the missing section of the cut disk, forming a discontinuous furrow. A standard heavy 51 cm disk plow with holes 8 cm off center will produce pits about 20 to 30 cm wide by 45 to 60 cm long, 15 cm deep and 40 cm apart. Under the most favorable conditions of soil type, soil moisture and weight of machine, the pits will have a storage capacity of about .013 m. The number of pits required per hectare can be estimated by dividing the estimated run off produced from a design storm by the storage capacity per pit. Pitting is effective on slopes up to 30%, and it has been estimated that the technique can produce as much as 4 times the amount of forage as produced on untreated areas (Slayback and Reeney, 1972).

Basins are similar to pits but larger and are generally about 2 meters long, 1.8 meters wide and 15 to 20 cm deep. A basin forming machine that can be pulled by a 25 to 30 HP tractor with hydraulic connections has been developed by Hinez and Frost (1973). Equipment can also be attached to the machine to allow simultaneous seeding. Slayback and Renney (1972) reported that basins produce 5 times as much forage as pits and 9 times more than untreated areas. The large capacity and greater life expectancy allows more plants to become established and insures a longer lasting and more productive stand.

3.2 Dry Land Agriculture

Watershed management in the arid zones of the developing countries must be concerned with agriculture, particularly with food production on individual or village sized landholdings. Although the development of groundwater may be an important aspect of small scale agricultural development in some specific areas, groundwater is not often a renewable resource in arid lands. The management of watersheds is primarily focused on the conservation of soil and the development of renewable resources. For agriculture, this includes the development and control of surface water derived from ephemeral streams for small water impoundments or for water spreading operations, water derived directly from rainfall for water harvesting (see water harvesting, this consultancy), and dry land, rainfed farming.

Dryland agriculture is a strategy for growing non-irrigated crops in semi-arid areas which receive between 300 and 500 mm of annual precipitation. As the annual average precipitation decreases from 500 mm, the differences in productivity between areas become pronounced. However, the total amount of precipitation received annually is not the best indicator of where dryland agriculture might be appropriate. The determining factor is the percentage of the annual precipitation available for plant use (the effective precipitation). This is determined by the seasonal distribution of precipitation, rainfall intensity, temperature and the amount and velocity of the wind. Dryland farming can be profitable in areas receiving no more than 300 mm annual precipitation, depending primarily upon temperatures during the growing season.

Summer fallow: This technique was developed in western U.S. where annual production of wheat was very unstable from year to year. The practice involves leaving the land free from live vegetation during part of a crop season. This requires tilling the land during the normal growing season to prevent water loss through weed growth. The purpose of summer fallow is to store soil water and accumulate nitrate nitrogen in the soil for the succeeding crop. The production of winter wheat in the U.S., for example, encompasses a 14 to 15 month period starting with the harvest of one crop, usually in June or July, and continuing until planting during September of the next year. There are three mayor constraints: 1) the average annual precipitation should be sufficient to wet the soil throughout the root zone, 2) the soil should be deep enough and have a water holding capacity sufficient to retain the water made available during the fallow period, and 3) the crop grown should have sufficient root development to utilize the stored water. Large losses of stored water can occur if weed control is poor during the tillage period, but usually evaporation accounts for most of the losses. More water is stored during the first winter of the following season, primarily as a result of a crop being harvested the previous summer and because of lower evaporation losses during the winter.

Soil type affects fallow efficiency since the soil must have adequate water holding capacity to benefit from the additional water provided during the fallowing period. As a rule, soil that has a light sandy loam or loamy sand texture throughout the soil profile will not provide an adequate reservoir for moisture storage as will heavier textured soils.

The early practice of clean tillage which leaves the soil surface exposed and susceptible to erosion has largely been abandoned today. This was accomplished with plows which turned the soil and buried the crop residue. An older method of pulverizing the upper 5 cm of soil to break capillary continuity and thereby reduce evaporation has also been abandoned because of the hazard of wind erosion. Stubble mulching, which was developed primarily to reduce erosion, is now a common practice. The practice requires undercutting crop residues with some type of subsurface instrument usually V-shaped sweeps varying in width from 45 to 176 cm. The type of implements used in stubble mulch systems depends upon 1) the amount of stubble anchored in the soil at the first tillage operation and 2) the amount of stubble desired at seeding time. Turning under all residue at planting provides a good seedbed, but there may be greater advantage to having 1200 to 1800 kg of straw per hectare left at planting to protect the soil against wind erosion.

The principal disadvantages of stubble mulch are: 1) weed control may be less effective than plowing under; 2) correct implement adjustment and speed of tillage is critical for satisfactory tillage; and 3) seedbed preparations and planting requires the correct type of implement and considerable care by the operator.

Level Pans: Leveled areas constructed in broad natural drainages on gently rolling topography have been successful in some semi-arid areas. m e purpose of these level pans is to trap run off and store soil moisture for growing crops. The pans can vary from less than one-half hectare to several hectares. They can be arranged to drain any excess water from one pan into another at a lower elevation. Grassed waterways are required for drainage, and diversions are necessary to reduce high volume run off at times when excess water would be detrimental to crop production.

The amount of water collected in pans depends upon the number and type of run off events and the number of pans in the system. Mikelson (1966) reported that for pans installed on the great plains of Colorado, U.S., increases in the yield of grain sorghum was 115 to 278 kg/ha greater than conventional agriculture.

Terraces: Terracing is one of the most frequently recommended strategies for controlling run off and erosion on agricultural areas of depleted watersheds. But terracing is expensive, requires diligent maintenance and expertise both in construction and operation. The introduction of terraces into a culture having no previous experience with the method often leads to failure. In semi-arid lands the expense cannot often be justified on the basis of increased crop production alone.

Conservation Terraces: These terraces, sometimes called level bench terraces, are designed to catch run off water and spread it over a leveled area that is to be intensively cropped. They were designed for and have been successfully used in temperate semi-arid regions. The strategy requires a level area (bench) on the contour and a contributing area can also be farmed to some less intensive crop or possibly assigned a fallow rotation.

Level bench terraces have been used successfully in eastern Colorado, U.S. where the annual precipitation is less than 500 mm, part of which occurs as snow. In this area a 2:1 ratio of contributing slope to bench width was found efficient (Mikelson, 1968). The principal advantages of the terraces were collecting snow, preventing run off from snowmelt and torrential rains and increasing crop yields through water conservation.

There are several constraints to the use and construction of conservation terraces. The crop grown on the contributing area must be managed in such a way that heavy run off events will not present an erosion hazard, and at the same time, provide sufficient run off water to the bench area. Light run off events present problems in obtaining an even distribution of water across the width of the terrace. During large run off events, farming operations have been delayed and crops have been damaged. Construction on the contour requires considerable skill.

The period of the year when run off is received largely determines the crop to be grown. Hass, et al. (1966) found the yield of grain sorghum grown annually on benches was 4.88 kg/ha greater than bi-annual yields on fallow.

Strip Cropping: Alternate strips of crop and fallow can be used to control erosion, but there is no evidence that the use of strip cropping promotes better water storage or utilization by plants. Strips used to control wind erosion are straight strips situated as nearly as possible at right angles to the prevailing wind. The width of strips depends upon soil texture; strips on erodible soils should be narrower than those on less easily erodible soils. Where water is the primary agent of erosion, strips should follow the contour in order to break the slope length and lessen the momentum of overland flow.

Cultural Methods: Contour furrows, pits and basins, as discussed above are also suitable methods of improving agriculture production in areas where climate and soils permit.

Crops: Most drought hardy grain crops are suitable for dry land farming. Winter and spring wheat, winter and spring barley, sorghum, oats and millet are the most common. In southwestern Colorado, U.S., field beans are annually cropped in areas of 330 to 430 mm annual precipitation.

3.3 Arid Land Forestry

There have been shifts in the policies of many of the major international development agencies within the past decade away from large scale, industrial forestry toward small scale, rural development forestry. An important element in this shift is an emphasis on watershed and environmental protection and improvement of the welfare of rural peoples through social forestry programmes. The U.S. International Cooperation and Development Act of 1979, for example, directs that U.S. foreign aid programmes in forestry devote their emphasis to social forestry-type activities.

The reasons for this emphasis is the realization that as land resources become scarce, the concept of forest resources becomes socially unacceptable, environmental arguments to the contrary. This emphasis is nowhere better directed than to the needs of rural peoples in arid lands. The harsh climates of these areas preclude large scale projects devoted solely to forestry. A principal factor which has led to the natural adoption of multiple-use systems and further supports their future development is the greater product security in marginal environments as opposed to that of monocultures.

By far the greatest emphasis of international donor agencies involved in arid land problems has been on fuelwood production. However, it is argued that very few rural communities in the developing world make fuelwood shortage their main priority. Barnes et al (1982) report that "most analysts seem to agree that wood scarcity is a very location specific problem." They report that real prices for fuelwood in developing countries, which should reflect the degree of wood scarcity, are rising very slowly if at all even in semi-arid nations such as Pakistan, Morocco and Senegal. This situation seems a paradox to anyone who has experienced fuelwood shortges in the limited areas (mostly urban) where they are severe, but it may help explain the frustration of donors with less than successful fuelwood projects.

The assumption which seems to run through most of these programmes is that rural peoples are subsistence oriented and therefore will support forestry projects which assist them in dealing with fuelwood shortages and environmental degradation. This is true in some cases, but the rural farmer is more likely to be immediately concerned with his own income shortage and limited prospects for economic improvement. An example is given by a forestry project in India.

In Gujarat State, India, the World Bank became involved in a diversified state governmental campaign to grow fuelwood and other tree products on marginal lands, in village woodlots, and on small family-managed plots. Success of the project has been varied, with persistent problems in obtaining local participation. However, it become clear that many of those who did participate did so, not to grow fuelwood, but to harvest building poles to be sold as a cash crop.

Multiple use of trees that provide a basis for increasing individual income and lead to development rather than subsistence is given by an example from Sudan: in the Kordofan Region of this country, the Integrated Sahel Programme, a recent project of the Sudanese government and the International Union for Child Welfare, reported good success in convincing village people to participate in the planting of Acacia senegal (the gum arabic tree) on personal plots. In the project's first year 80% of all farmers in 27 villages applied for seeds and seedlings, and demand exceeded supply despite the production of more than 1,000,000 seedlings (Hammer, 1982). Initial success in securing villager participation stemmed from the locally well-known multiple-use value of the tree, especially the market for gum arable, and the project's stated intention of integrating tree production with improved services to the community in the area of agriculture, health, education and water development.

The three most promising means of encouraging the practice of arid land forestry for multiple use appear in agroforestry, shelterbelt planting and in encouraging the production of trees in riverine and riparian communities.

Agroforestry: A major problem facing subsistence farmers and herders in many arid land countries is to obtain a steady supply of food and/or income throughout the year. In areas where dryland agriculture may be possible, income and food is only obtained at widely spaced harvesting intervals. In areas where only livestock herding is possible, prices are subject to great fluctuations in markets. Conventional forestry is usually unattractive to farmers and herders because of problems in cash flow and long investment periods. agroforestry can offer opportunities to subsistence farmers and herders to diversify production of wood and non-wood products and to maintain regular employment and income during periods between harvest or livestock sales, in addition to the other obvious benefits that may be obtained from trees. It is only recently that scientific attention has been given to the practice of combining forestry and agriculture.

In the warmer, drier portions of arid lands in the developing countries, agroforestry primarily involves the integration of forestry and livestock production. This type of agroforestry has been practiced by arid land pastoralists for thousands of years. A menagerie of both browsing and grazing animals is herded together to artificially fill ecological niches in order to utilize trees and shrubs as well as grass and herbaceous plants.

This practice continues today but on a larger scale. The cutting of shrubs and trees, primarily for fuelwood, by increasing numbers of people and the cropping of foliage by corresponding increases in numbers of livestock has now diminished or completely eradicated the most valuable fuelwood and browse species on vast areas. These species are usually replaced by undesirable and non-palatable plants.

The use of trees for browse is not incompatible with forestry. Cropping shrubs and trees generally removes transpiring tissues, thus reducing water stress on the remaining plant, and in some cases, assisting in plant survival. Browsing animals also are the primary distributors of some tree seeds. On the other hand, excessive cropping removes vital photosynthetic tissue and structural material critical to life functions. The strategy for agroforestry, therefore, is to achieve an optimum mixture of trees and grazing plants compatible with numbers and types of livestock. In many cases this will simply be good range management but with a vertical component that includes trees and a variety of livestock preferences.

Depending upon local situations trees could be spotted on rangelands to provide forage and shade, or on the perimeters of fields or around dwellings to serve as fences and/or windbreaks. If moisture conditions are suitable, such as in or near ephemeral drainages, wood lots (which should be protected from livestock) might be established and managed to provide fodder, fuel and construction materials. Any surplus might be marketed.

In semi-arid areas where dryland agriculture is possible, nitrogen fixing legumes can be used in combination with agriculture. For example, it was found in Senegal that millet yields when grown under Acacia trees were increased as much as 250% and were 350% higher in protein (Ffolliott and Thames 1983). The phosphate releasing ability of some tree-root mycorrhiza can also be of advantage in providing essential nutrients to associated agricultural crops. In Niger it was found that under Acacia albida total nitrogen and exchangeable calcium was increased 100%. Assimilable phosphorus increased 134% and exchangeable magnesium increased 70% (Delwaulle, 1977).

The aim of agroforestry is to develop a desirable replacement that at least matches the productivity of any alternative system. Important benefits include: 1) lessening the danger of catastrophic losses that can occur in dry year sequences by having an alternate source of income and/or animal feed, 2) direct economic benefits of fuelwood, fence posts, poles and other tree products without having to buy or transport them from other sources, 3) increased opportunity to move from destructive land uses which may return profits only over the short term toward practices with long term benefits without diminishing productivity and 4) early reduction of the economic investment of establishing tree crops from the proceeds of thinning and crown manipulation to produce fodder, wood and fuel. Watershed protection can also be accomplished with agroforestry in modifying the microclimate by reducing temperature extremes, raising humidity, lowering wind velocities and reducing the energies of rainfall impact and surface run off.

There are constraints to agroforestry systems in addition to those imposed by climate. Agroforestry involves complex associations and, therefore, is not very amenable to experimentation. This problem is compounded by the scarcity of trained personnel to improve existing systems or develop new systems. In some cases, economic yields of agroforestry systems can be lower than herding or farming even though the long term environmental advantages may be great. In other cases, the combined value of trees and associated livestock and crops may eventually be higher than livestock or crops alone. But there may be resistance by the rural poor to planting and managing trees whose products can only be realized over long cycles. There is also a general lack of knowledge of the potentials of agroforestry. Without adequate experience, there is a danger of creating resentment at both the rural and decision-making levels from unsuccessful projects based on insufficient information. The development of projects based upon reports of "miracle trees" is an example.

An ultimate goal of agroforestry is the conservation of the soil and water resources of watersheds while satisfying the needs of rural people for food, fuel and income. Successful agroforestry depends not only upon the quantity and quality of joint products that may be produced, but also largely upon the socio-political strategies built into a project.

Shelterbelts: Wind erosion is a serious problem over much of North Africa and in other arid environments. Wind also accelerates evapotranspiration. The strategy for remedial action is to reduce the velocity of winds by increasing the roughness in the air layer near the ground and to protect the soil surface. Shelterbelts are the most common and, frequently, the most economical means of doing this.

Shelterbelts have been successful in semi-arid, temperate climates since the middle of the 19th century in the western world. In China their use extends back through thousands of years. They have been effective in improving the microclimate, reducing wind erosion, increasing crop and livestock yields, reducing heating costs, and providing fodder, fuelwood and other wood products. It has also been demonstrated that shelterbelts can be even more effective under the harsher conditions of warmer arid lands. On these lands, the value of thrifty tree species may be even higher than that of other products of land use.

Shelterbelts (no distinction is made here between windbreaks and shelterbelts) can consist of from one to about 12 parallel rows of trees. Shelterbelts of five rows are generally efficient and are not difficult to maintain. However, in considering economic worth, account must be taken of possible multiple uses. For example, wood products, shelter for animals and bees, food and cover for wildlife, and fodder for livestock may be important considerations. For these considerations, shelterbelts of more than 5 rows may be desirable. One-row shelterbelts are not effective and can sometimes funnel the wind. However, in areas where agricultural land is at a premium, single rows bordering small fields can be effective in increasing the overall surface roughness of the general area.

Shelterbelts are most often planned so that they will develop a triangular cross section, with the tallest trees in the center flanked by shorter trees and shrubs along the edges. The spacing within rows depends in part upon the tree and shrub species planted and the type of management to be followed once the plants mature. In general, seedlings are planted close together to obtain early closure. Final spacing within rows should be from 1 to 1.5 metres for shrubs and 2 to 3 metres for trees. Spacing between rows should range from 3 to 4 meters to allow for subsequent cultivation.

The design of shelterbelt systems largely depends upon the velocities and directions of local winds. If there are definite prevailing winds, a series of parallel shelterbelts should be established, preferably at right angles but no less than 45 degrees to the direction of the winds. More often, winds blow from various directions which would require a checkerboard pattern. In some cases, dense shelterbelts may be planted across the major wind directions and less dense belts planted across minor directions.

In irrigated areas, shelterbelts should be located mainly along irrigation channels. In rolling topography, shelterbelts are more effective if planted along ridgetops. Therefore, a compromise is sometimes necessary to take into account both the direction of winds and the cultural and physical characteristics of the area.

For sheltering livestock, a compact shelterbelt in a U,V,X or square configuration can be used. Shelterbelts around buildings are often planted in an L-shaped pattern across the prevailing winds.

In arid climates, the species that can be grown in a specific area without irrigation is usually the deciding factor in shelterbelt design. Ideally, shelterbelt species, in addition to drought tolerance, should possess the following:

Seldom will one species have all of these attributes. Often two or more species will be required to develop a shelterbelt that will provide adequate protection. For example, the low growth form of many Acacias make this species useful for planting in the outer rows: the inner rows may consist of taller tamarisk, casvania or eucalyptus. Single plant species, particularly those that sprout, can sometimes be managed to provide full vertical shelter by alternately cutting the outer rows of trees and allowing the cut trees to complete the shelter.

In dry climates it may be necessary to irrigate after planting. The number of waterings and the amount of water applied depends upon climate, species, and soils. For example, in an area of sandy loam which receives 150 to 200 mm of rainfall and has a dry season of 8 months, about 6 applications of 10 liters for each seedling is probably sufficient to assure survival. Contour furrows, trenches or other cultural treatments as described under Range Improvement are strategies that will assist survival and encourage growth. In addition to economic returns, the greatest benefits of shelterbelts are in moderating the effects of wind which causes erosion, damages vegetation, dessicates plants and soils, lowers the health of animals and reduces the comfort of humans.

Shelterbelts should be planted a suitable distance from buildings due to downdrafts on the leeward side of the shelter.

The effects of properly designed shelterbelts extend from about 10 to 25 times the height of the belt downward with greatest effect close to the leeward side. In hot and dry climates, dense shelterbelts placed too close to buildings may result in oppressive heat. These belts should be permeable and located at least 30 to 45 meters (but no greater than 90 to 120 meters) from the buildings.

The principal constraints to establishing shelterbelts, in addition to finding suitable species, is the use of water that might otherwise provide greater returns for increasing forage or agricultural production. Adverse environmental effects can also occur if trees harbor birds, insects, animals, or disease organisms which are harmful to crops or livestock. These factors should be weighed before shelterbelts are planned.

Riparian forestry: In many arid lands the only suitable areas for growing trees is in and along washes, streams or rivers where there is a source of water. However, where soils are suitable for farming and where flooding is not a great problem, conflicts with agriculture may arise. Furthermore, phreatophyte plant communities which depend upon shallow groundwater usually found adjacent to streams and rivers can consume up to 2.5 m annually in some areas.

In Pakistan, for example, forestry and agriculture in the flood plain of the Indus and its tributaries have supported the region of what is now Pakistan for hundreds of years. At present, the riverine trees of these phreatophyte communities make up over 12% of the country is productive forest lands. It is only recently that active reforestation and management of these forests have been undertaken where the emphasis is on the production of the multiple products, fodder, browse, fuelwood, timber, resins, gums, medicinal products, and wildlife (Shah and Thames 1985).

Species Selection for Arid Land Forestry: There are numerous examples from around the world of introduced species that have outperformed native trees in growth rates and in the production of quality wood products. However, the use of native trees is a strategy that offers less risk from biological, ecological and social acceptance standpoints, particularly for new projects in areas where exotic species have not been well tested. The selection or encouragement of one or more tree species should be based on a number of criteria, some of which are listed in Table 1. The list is not exhaustive. No single species will meet all of the criteria that may be desired for any specific area. The objective is to give the highest priority to those criteria which are most desirable for local conditions and needs, and rank prospective species accordingly. But the overriding criteria for species suitable for arid land forestry is their ability to grow, survive and produce the needed products. In most cases this product is fuelwood.

Table 1 - Criteria for Species Selection

drought wind salinity/alkalinity/acidity
temperature disease water logging
browsing insects/animals other species
growth rate form strength/density
heat content water content workability
fuel erosion control pharmaceuticals
wood products animal shelter/shade wind breaks
food soil improvement wildfire
fodder oils/resins/gums/nectar aesthetics
Seed Natural Regeneration Nursery Planting
availability seed local (small) direct seeding
quantity/viability coppice commercial or gov. (large) bare rooted
special treatments layering time to produce stock balled stock


The development and management of the soil and water resources of arid zone watersheds require knowledge of the hydrology of arid systems. Contrasted with the predictive relationships developed for watershed behavior in the humid, more densely populated temperate zone, quantitative relationships have not been well developed for arid zone watersheds. The studies that have been made point up some striking characteristics that have important implications to management. Some of the most notable are:

Among these characteristics, rainfall variability, run off, erosion and potential evapotranspiration have the greatest implications for the management of arid zone watersheds. The relationships between groundwater and surface water also have important implications for downstream irrigation works.

Precipitation: In temperate climates the standard deviation of annual rainfall is about 10 to 20 percent in 19 years out of 20. In more arid climates with mean annual rainfall of 200-300 mm, the standard deviation commonly ranges from 40 to 200 percent. Usually there are more years with annual values below the mean than above it. At a site in Northern Chile, as an extreme example, no rainfall was measured in four consecutive years, but 15 mm was received in the 5th year which gives a mean of 3 mm for the five year period.

The fact that no rainfall was measured at the station in Chile does not mean that it did not rain in the area. From simple statistics: if 50 storms of 1.5 km in diameter each fall at random locations over an area of 150 km on a site which has 50 raingage stations, also located at random, the probability of any one station recording rain is only 0.4 percent. The nomads in northern Sudan, where rainfall is recorded only every few years, have long been aware of this and have existed for centuries by following the storms and the forage they produce.

The strong link between erratic precipitation and vegetation phenology makes it difficult to optimize timing of human use of arid areas, including livestock grazing. Long term directional changes, such as desertification, are almost always linked to some alteration in water regimes. Even in contemporary systems water is the main ecosystem influent, with the exception of the physical impacts of man's activities.

The spatial distribution of rain, particularly that of convection storms which produce the bulk of run off from watersheds in many arid zones, is determined primarily by the characteristics of storm cell clusters and their movement relative to the terrain. Commonly, storm cells are from one to 5 km in diameter and move no more than a few kilometers before they dissipate. The result is a roughly ellipitial isohetyl pattern. Surface run off produced by the cells is maximum along the axis of the ellipse, and the isolated nature of the cells results in only partial area run off.

Run off: In most dry regions the watertable is well below the stream channel invert, especially in headwater areas, with the result that the streams are influent rather than effluent. As a consequence water flows in the channels only during large storm events, and then only for short periods of time. Thus, the channels are dry most of the time.

When storm flows occur in these normally dry and often highly permeable channels the volume of flow is reduced by infiltration into the stream bed, the banks and possibly the flood plain. These losses to infiltration, called transmission losses, reduce not only the volume of the flow but also the peak discharge. Transmission losses are usually a large component of the water balance of arid zone watersheds, and may account for as much as 15 to 20% of the total precipitation input.

Because of transmission losses and because of the spatial variability of rainfall, which leads to partial area run off, both peak discharge and volume of flow per unit area of watershed usually decrease with increasing size of watershed. This is in contrast to more humid areas where water yield per unit area usually increases with increasing size of drainage area (Renard, 1985).

Peak discharges in the headwaters of arid zone watersheds are often surprisingly large, particularly at locations near the center of intense convection storms. This is due to the short times (time of concentration) required for water to travel from the most remote points on the watershed to the watershed outlet. Times of concentration are short because the greater portion of storm run off in arid zones occurs as overland flow which travels downslope to the channels unimpeded by vegetation and surface litter as it would be in more humid zones. Furthermore, arid zone soils which characteristically have no organic horizons and shallow or undeveloped A horizons do not permit rapid infiltration of short duration storm rainfall and consequently enhance the volume of overland flow. For many arid zone soils, infiltration is also reduced by the presence of impeding layers near the surface, such as those cemented by calcium carbonate, and by sealing of the soil surface by raindrop puddling or by crusts resulting from the activity of algae or cryptograms. Walls of water travelling at high velocities down dry stream channels during a storm event are a common phenomena in arid zones and are the result of both the large quantity of rapid surface run off and high transmission losses.

Erosion and Sedimentation: Because of the large contribution of surface run off to the water balance of arid zone watersheds, erosion is a serious problem in all dry regions of the world, particularly in those regions where the naturally sparse vegetation is further reduced by excessive livestock grazing and/or fuelwood cutting. Erosion is greatest in the more habitable semi-arid regions of the world which receive from about 200 to 400 mm annual rainfall, and it is these regions which are presently undergoing severe desertification worldwide.

The high erosion rates characteristic of arid zone watersheds help explain the shallow soil profiles in many areas and often the lack in development of an A horizon. The most desirable soil, with its organic matter and nutrients, is often non-existent because erosion removes the soil from upland watersheds as fast as it is developed from the parent geologic material. Such erosion also explains the erosion pavements which dominate the soil surface in many sparsely vegetated areas.

The amount of eroded material passing a point on a stream system is called sediment yield. On many arid zone watersheds there is an almost unlimited supply of sediment available to the channel system from erosion. As a consequence the stream channels adjust to the sediment yield by developing a rectangular or trapezoidal cross section which are more efficient in transporting sediment than would be the parabolic channels of more humid areas. Furthermore, the longitudinal profiles of treambeds tend to be linear rather than concave as generally are the profiles of streambeds in humid areas. This is due to the decrease in water yields, and consequently sediment yields, per unit area which results in sediment being deposited to form a convex profile but which is balanced by tributary inflow.

Evapotranspiration: Arid and semi-arid areas are characterized by moisture deficiencies due to potential evapotranspiration exceding annual or seasonal precipitation. For example, in southeastern Arizona, pan evaporation of 2 meters or more is about 7 to 8 times the annual rainfall, with the result that vegetation must be capable of using the limited soil moisture rapidly and efficiently.

In most dry regions, the watershed vegetation is composed of mixtures of plant communities (grasses, brush and fortes). Because of limited moisture, the total amount of the land area covered by live plants is generally very low. In southern Arizona, U.S., for example, plant cover is generally less than 10%, and often as low as 2%. Much of the bare areas between plants support roots which in turn affect the soil moisture beneath and between plants, and in turn, the soil structure-erosion problem. Most warm-season grasses are classified as C4 plants, which are more efficient water users than the C3 brush plants. Most grass plants use about one third the water used by a brush per equal amount of biomass. Thus, in characterizing soil moisture regions, which is essential for run off prediction, not only plant density must be considered, but also the composition of the plants as well.

Surface/Groundwater Relationships: In planning for groundwater development the quantity of water available is not the only important variable; the quality may be equally or more important, particularly in arid zones where the natural input to the system from rainfall is small.

The problem in such areas is one of inputs from surface water flow and groundwater recharge that always contain some salts, dissolved from rocks and soil, and increased salt loads from fertilizers and sewage effluents and yet the main outflow - evapotranspiration - is of pure water only. The salts inevitably accumulate in the soil and groundwater unless there is some through-flow; and this is, in a sense, a loss of good water.


One of the greatest problems in managing arid zone watersheds is the lack of adequate predictive methods; methods that would provide reliable estimates of the effect of land use practices both on the watersheds and downstream from the watersheds. This is needed to recommend suitable land use practices and to determine the economic feasibility of these practices within the upstream-downstream system.

Better schemes are needed to match precipitation patterns and consequently plant phenology with human needs, particularly for growing agricultural crops and for livestock production. This will require better predictive methods for evaluating the spatial and temporal variability of precipitation and more knowledge of the requirements of economic plants.

Livestock production, often the principal economic use of arid lands, lacks information for managing herds of different livestock types where the vertical as well as the horizontal component of grazing lands must be considered. The ecological-livestock interrelationships of this type of system are not well understood.

A better understanding is needed of the variability of rainfall in relation to agricultural crop growth. Crop-water models are available but the need now is for more specific information on short term rainfall variabilities; while further research is also needed on longer time-scale changes, measured in terms of years.

More information is needed on plant species and varieties which can be adapted to the social, economic and site conditions of specific arid land situations. Species trials, plant improvement, utilization methods and market development for indigenous and exotic grass, shrubs and trees needs expansion.

More work must be accomplished in the social-forestry area to determine better methods for gaining social acceptance of forestry projects. It is often more important to know why a project failed than why it was successful. Unfortunately, failures are not often reported, or for various reasons are not given high visibility. An international clearing house would be helpful.

Techniques for rehabilitating depleted areas abound. There are very few situations where one or several of these techniques are not applicable, but the economics and sustainability of these techniques under the varying conditions found in the developing countries have not been established for most situations.

More information is needed on the development and management of riparian and riverine forests. These are often the principal or only source of forest products on many arid zone watersheds.


Bonner, Douglas F., Julia Collen and William Ramsay Social Forestry (1982) in Developing Nations, Resources for the Future. Discussion paper D-73F, Washington, D.C.

Branso, F.A., R.F. Miller and I.S. McQueen Contour Furrows, Pitting and (1966) Ripping on Rangelands of Western United States. Journal of Range Management 19 (4):182- 196.

Delwaulle, I.C. Species, Techniques and Problems of Semi-arid Zones (1977) (The Sahel). In: Savanna Reforestation in Africa. FAO Rome. pp. 160- 167.

Ffolliot, Peter F. and J.L. Thames Environmentally Sound Small-Scale (1983) Forestry Projects. CODEL-VITA. 109 p.

Hammer, Turi Reforestation and Community Development in the Sudan. (1982) Resources for the Future. Discussion paper D-73M, Washington, D.C.

Hass, H.J., W.O. Willis, and G.O. Boatwright Moisture Storage and Wheat (1966) Yields on Level Bench Terraces as Influenced by Contributing Area, Cover, and Evaporation Control. Agro. Jour. 58:297-299.

Hinz, W.W. and Frost, K.R. Reseed Rangelands with Basin Farming Machines. ((1974) Agri-File Q-134. University of Arizona Cooperative Extension Service.

Mickelson, R.H., M.B. Cox and J. Musick Run off Water Spreading on Leveled (1966) Crop Land. Jour. Soil and Water Cons. 20:57-60.

Mickelson, R. Level Pan System for Spreading and Storing Watershed (1968) Runoff. Soil Sc. Soc. Amer. Proc. 30:388-392.

Renard, Kenneth G. Water Resources of Small Water Impoundments in Dry (1985) Regions. In: Small Water Impoundments in Semiarid Regions. J.L. Thames (editor). University of New Mexico Press (in press).

Shah, Bashir and J.L. Thames Role of Riparian Vegetation in Pakistan. (1985) Paper presented at the North American Riparian Conference, University of Arizona; April 16-18, 1985.

Slayback, R.D. and Renney, C.W. Intermediate Pits Reduce Gamble in Range (1972) Seeding in the Southwest. Jour. Range mgmt. 25(3):224-227.

Wright, J.R. and F.H. Siddoway Improving Precipitation-Use Efficiency on ((1972) Rangeland by Surface Modification. Jour. Soil and Water Cons. 27(4):170-174.

Wright, Nel and Streetman, L.J. Grass Improvement for the Southwest (1960) Relative to Drought Evaluation. Ariz. Agr. Exp. Sta. Tech. Bull. 143. 16 p.

Contents - Previous - Next