Previous Page Table of Contents Next Page

Agroforestry in the semi-arid tropics

R.J. Vandenbeldt

Rick J. Vandenbeldt is Field Team Leader of Winrock F/FRED, Bangkok, Thailand.
Note: This article has been in pan excerpted from K.G. MacDicken and N.T. Vergara, eds. 1990. Agroforestry: classification and m an management. New York, Wiley.

Early literature on agroforestry generated a lot of interest, the consequence of which was a plethora of development activity, often with insufficient planning and research backup, inadequate time frames and very mixed results. This article examines agroforestry technologies for the semi-arid tropics and provides suggestions that may enhance their implementation.

The SAT environment

The semi-arid tropics (SAT) cover an area of about 20 million km2. Kampen and Burford (1980) estimated that 700 million people live in this zone, nearly half of them in India. This population undoubtedly has increased, perhaps by as much as 50 percent, in the intervening 11 years. The SAT cover most of West, East and the southern part of central Africa; most of India, northeastern Myanmar, northeastern Thailand and northern Australia; most of Mexico; and large parts of eastern and central South America. The SAT environment is characterized by high atmospheric water demand; a high mean annual temperature (>18°C); and a low, variable annual rainfall (400 to 1900 mm) (Swindale, 1982). The climate of most of the SAT is monsoonal, with over 90 percent of the rainfall occurring in the period of April-October in the Northern Hemisphere and October-April in the Southern Hemisphere. This article focuses on agroforestry in the dry semi-arid tropics, i.e. those areas in which rainfall exceeds potential evapotranspiration for less than 4.5 months of the year.

Agroforestry systems for the SAT

Parkland systems

Although no definitive estimate exists, it is probable that the parkland system, characterized by mature trees dispersed in cropped fields, is the largest single agricultural land use in sub-Saharan Africa. Across the entire Sudano-Sahelian zone of West Africa, one finds crops planted under varying densities of mature trees. It is quite possible that some of these older parklands are remnants of elaborate pre-colonial agricultural systems, with cropping intensities and input levels not found in the region today (Hervouet, 1991).

The ability of these parklands, or two tiered systems, to enhance and stabilize crop production has been much studied over the past 20 years in West Africa and to a lesser extent in arid parts of India. The systems that have received the most attention are the Prosopis cineraria/millet mixtures of eastern Rajasthan, India (Mann and Saxena, 1980), and the Faidherbia albida/grain systems predominating throughout the Sahelian zone and in some parts of East Africa.

Many authors have shown an enhanced effect of these species, particularly F. albida, on grain crops growing underneath. Estimates of increased yields range as high as 100 percent compared with crops grown away from the trees (CTFT, 1988). While this effect is generally attributed to improved soil conditions under the tree because of litterfall, recent evidence suggests that much of this fertility may in fact precede the tree (Geiger et al., in press). There is also new evidence that emerging crops benefit equally from lower soil and leaf temperatures under the light shade of the tree, which is defoliated at the start of the rainy season (Vandenbeldt and Williams, in press).

Although there are no comparative yield data for P. cineraria/cereal associations from India, it is believed that crop yields are also higher when grown under this system (Mann and Saxena, 1980). Limited evidence suggests that the tree has an unusual rooting pattern that does not seem to compete with crop growth. Muthana et al. (1984) excavated a 20-year-old P. cineraria tree to a rooting depth of 7.5 m. The horizontal extension of lateral roots of this specimen was less than I m from the trunk. The local practice of lopping the trees annually for the highly valued fodder soon after crop harvest ensures that the canopy never gets large enough to shade subsequent crops.

Despite the fact that such traditional systems are beneficial, it is rather difficult to improve on them or extend them into other areas. There are two reasons for this: the first is that other tree species create high rates of competition for crops; the second is that it has often proven difficult to establish these systems in areas where farmers are not accustomed to them.

Parkland agroforestry system in the Niger (above) with Faidherbia albida and in Rajasthan, India (below) with Prosopis cineraria during the dry season P. cineraria trees are lopped frequently for fodder, while F. albida trees are rarely pruned In the Niger. Also note crop residue in the Nigerian example. Residues in India are almost always removed for animal fodder



There are not believed to be definitive reports of species in SAT other than P. cineraria and F. albida that exhibit the ability to increase crop yields under their canopies. In fact, Prajapati et al. (1971) observed suppression of yield by Prosopis juliflora growing near a cropped field. Roots of the trees extended 30 m into the field, reducing sorghum yields by 80 percent over this range. Similar crop yield reductions can be readily observed under most other tree species found in cropped fields of the SAT. However, in some cases, the Parkia biglobosa/cereal mixtures of the West African Sudanian zone, for example, crop losses are more than recovered by products produced by the trees.

Canopy management may be one way to decrease this competition. Mann and Saxena (1980) stated that a sixfold increase in mung bean was obtained when it was grown under 12-year-old Acacia tortilis spaced at 4x4 m when the trees were pruned. Singh et al. (1986) showed the importance of intensively lopping competitive trees in tiered systems. Although plots under pollarded Leucaena leucocephala spaced at 2x6 m had just half the sorghum grain yield of sole sorghum plots, these yields were ten times those in lightly lopped treatments. Therefore, lopping could be used to expand the range of species suitable for parkland systems, and the prunings could be used for organic fertilizer, mulch, fodder, fuel and other purposes.

Projects that have attempted to establish parklands by extensively planting trees on regular grids have not been successful.

This is because of variable and/or slow tree growth and often a lack of meaningful farmer participation and interest. Some success has been met with natural regeneration schemes, where farmers are provided incentives to protect young trees that emerge in their fields (Sumberg, 1990). However, this reduces control over tree species, seedling location and spacing. Given the high microsite variability in many SAT soils, an alternative strategy may be to identify "islands of fertility" in farmer's fields and plant locally favoured species on them. The subsequent rapid growth and high survival rate would be attractive to farmers and would lessen the amount of time needed to protect small young trees. Another possible option would be to introduce more valuable cash crops in the favourable microenvironment found under the shade of existing mature F. albida or P. cineraria trees.


Tree rows as wind-breaks have long been used in semi-arid temperate regions of North America, Europe and Asia for protecting crops and soil against wind damage and wind erosion. More recently, their effectiveness in increasing crop production has been demonstrated in the drier parts of the SAT, particularly sub-Saharan Africa.

The few critical studies that have been done in the SAT confirm findings of increased crop yield behind wind-breaks in semi-arid temperate regions. In northern Nigeria, Ujah and Adeoye (1984) reported a 14 percent average increase in millet yield behind a Eucalyptus camaldulensis windbreak. In the arid zone of India, Sur (1986) reported a 21 percent average increase in yield of protected cowpea over six years. El-Kankany (1986) reported a yield increase of 36 percent for cotton, 38 percent for wheat, 47 percent for maize and 10 percent for rice in Egypt, where over 100000 ha of croplands are protected by windbreaks. In the Majjia Valley of the central Niger, double-rowed Azadirachta indica (neem) wind-breaks spaced 100 m apart showed a 20 percent increase in crop yield in two separate studies (Vandenbeldt, 1990).

The great interest generated by these data and experiences led to an increased emphasis on wind-breaks in West African development assistance projects during the late 1980s. These invariably met with mixed results. A major problem was uneven tree growth in wind-break lines caused by soil variability, which creates unevenness in wind-break heights and reduces their effectiveness. Too much faith was put in results from trials located on research stations or from sites of very successful wind-break projects. An example of this was the Majjia Valley project, where the trees benefited from a very shallow water-table which is unrepresentative of most other agricultural lands of the Niger. Finally, wind-breaks in many of the projects were designed to run in straight lines perpendicular to damaging winds. From a technical point of view this is ideal, but in practice it can lead to inequable losses among small farmers of lands occupied by wind-breaks.

Competition between the trees in a windbreak and the crops they are supposed to protect can also be a problem (Vandenbeldt, 1988). In a study at the University of Agricultural Science (Bijapur, Karnataka, India), safflower was planted on either side of a line of six tree species. In order of decreasing competition, the species were Acacia nilotica, A. catechu, Eucalyptus camaldulensis, Dalgergia sissoo, Leucaena leucocephala, Causarina equistifolia (G. Radder, University of Agricultural Science, pers. comm.).

Other than species selection, there are few management options for reducing underground competition. Kort (1986) advocates root pruning in temperate regions to a depth of 60 cm close to the wind-break. This is hardly possible in less developed nations, where draught power, even animal draught power, is lacking. As with two-tiered systems, canopy management may be effective in reducing competition for solar energy, provided it does not interfere with the aerodynamic efficiency of the windbreak. This was accomplished in the Majjia Valley by pollarding one of the two windbreak lines at a height of 2.5 m. The wood harvested from this operation has provided income for the cooperative which manages the wind-breaks.

The successful establishment, protection and management of the Majjia Valley plantings have required nearly 15 years of sustained effort and funding from the Government of the Niger and donor agencies. Similar long-term commitments from donors and governments are rare among development projects in general, let alone wind-break projects. Therefore, it is imperative to select sites wisely, i.e. those that will give a reasonable return on investment for wind-break establishment. Further research into the ecophysiological processes, favourable and unfavourable, that govern wind-break/crop interaction in the SAT is important. Imaginative ways to interact with small farmers in anticipating costs and benefits of such projects and in actually distributing them are just as necessary in wind-break extension as in that of other agricultural technologies.

Another factor that bears consideration is the effect of wind-breaks on the air temperature on the down-wind side of the trees. A number of studies indicate that increases in air temperature in the lee of shelter-belts of semi-arid and arid areas may be as much as 5 to 10°C and may shrivel or burn crops (Guyot, 1967).

Competition of Leucaena hedgerow on growing millet crop. The left side has a 50 cm deep polyethylene barrier separating the Leucaena roots from the crop root zone

Alley cropping

Based on positive results from the humid tropics, particularly at the International Institute of Tropical Agriculture (IITA) in Ibadan, Nigeria (Kang, Wilson and Lawson, 1984), alley cropping has generated much interest among donors. This is particularly so in West Africa where a collaborative network has been established by the IITA and the International Council for Research in Agroforestry and under various national programmes.

It remains to be seen whether alley cropping, which has been so successful on-station, will be widely adapted by farmers in West Africa, particularly in and to the north of the Sudanian zone where a reduced incidence of trypanosomiasis now allows extensive animal husbandry. The fodder needs of these animals will compete with the use of hedgerow loppings for green manure, a key component of the alley cropping technology. Furthermore, it is now clear that alley cropping has definite drawbacks in areas where water is a limiting factor. Narrow hedgerows, spaced at distances recommended for the humid and sub humid tropics (i.e. 4 m), are particularly competitive with crops in the SAT. Trials in India have shown that grain yield of millet grown in 3.6 m wide Leucaena alleys systematically decreased with alley age from a yield reduction of about 17 percent the first year to over 80 percent by the fifth year (Singh et al., 1989). This is probably because of underground competition for water.

Despite these limitations, the concept of carefully placed and closely pruned rows of trees in farmers' fields may have a role to play in SAT farming systems (Kang and Vandenbeldt, 1990). Such vegetative barriers could be located on field boundaries, providing feed for livestock and delineating cropped fields. Similarly, if situated at suitable contour intervals, they may serve to protect highly erosive soils. Finally, as low wind-breaks, they could reduce wind erosion and actually trap saltating soil particles on sandy soils (Renard and Vandenbeldt, 1991).

The future role of alley cropping in the SAT will depend on the development of management methods to control the sources of competition between hedgerows and crops, and a greater selection of woody or perennial grass species to use in the hedgerows. A better chance of success is likely if alley cropping is only considered for areas where the assured annual rainfall exceeds 750 mm on soils with good water holding capacity.

Silvipastoral systems

There are many animal husbandry systems in the SAT, but two extremes may be recognized. The first is where animals are a permanent fixture in the farming system (as in much of India), and are primarily stall-fed with supplemental pasturing. The other extreme is the nomadic and transhumant systems of the Sahel, where animals are moved from one site to another according to changing pasture conditions throughout the year, or herded to distant pastures during the rainy season to protect crops. In both cases, livestock play an important role in sustaining field fertility through manure production. In India, farmyard manure is normally collected and spread on the fields. In the Sahel, herders are paid to bed animals on arable fields.

In both systems, livestock pressure must be balanced with rangeland vegetative productivity if it is to be maintained. However, most grazing lands in the SAT of Africa and India are public lands and are therefore freely accessible (Jodha, 1985; Jahnke, 1982), a situation that may be leading to their overuse and irreversible decline. Under communal ownership, the incentive is to expand one's herd, without concern for the effect it has on the land, in order to reap maximum short-term personal gain. At the same time, the long term impact of such individual misuse is shared by all users, and therefore the individual's long-term loss is buffered. In this situation, there is no personal incentive to make improvements on the pasture (Jahnke, 1982). In India, this decline of communal pastures is exacerbated by the growing tendency toward private ownership of grazing land, forcing poorer livestock owners to share less and less land (Jodha, 1985).

In the face of this emerging regional crisis, it is clear that meaningful changes in land-use policy and policy enforcement is an important first step. Before this is done, agroforestry interventions, or any other production-side intervention, will have little effect on a broad scale. However, they can be useful on smaller scales: excellent examples from semiarid Australia (Most and Reid, 1985), India (Deb Roy and Pathak, 1983) and elsewhere demonstrate the usefulness of fitting the potentially excellent feeding value of SAT trees and shrubs (Le Hourerou, 1980) into pastoral and agropastoral systems.

It is possible to think of all sorts of grass/tree/crop designs to produce more fodder at the farm level, a task far beyond the scope of this paper. However, some guidelines can be proposed. First, fodder trees and other suitable perennial vegetation should be considered as components of almost every agroforestry system for the SAT. Examples would be the use of fodder trees in wind-breaks, or the management of parkland systems to produce a fodder component through pod production or lopping. Second, more effort must be made by scientists to incorporate studies on increasing fodder production into farming systems research. Finally, the overall direction of research and development work at the farm level in this area should be the designing of efficient cut-and-carry systems.

Miscellaneous systems

Where irrigation is available, or in river valleys where water-tables are close to the soil surface, many traditional agroforestry systems closely patterned after those in the humid tropics can be observed (Vandenbeldt, 1990). Such sites offer rare opportunities to maximize the output of tree products on a year-round basis and to take advantage of their many uses as vine supports, living fences, shelter and the like.

In non-irrigated areas, tree planting for specific farm management objectives is little studied but of significant potential (Weber and Hoskins, 1983). Vegetative fences, either live or in the form of thorny prunings, are widely used in the SAT. Similarly, boundaries between fields and farms could be made productive by tree or shrub planting for browse and other uses. Vegetative livestock driveways are sometimes established in and around villages in the Sahel to control animal movement and define village boundaries. Trees can play an important part in river and stream bank and waterway stabilization efforts, vegetative contour strips for soil conservation and simply for shade. Tree plantings can fill production "gaps" in farming systems by making use of farm areas that are not usually cropped - for example, bunds, field corners and rocky and unproductive field sections.

SAT trees and shrubs have an excellent potential value as kidder and should be fitted Into pastoral and agropastoral systems


The limited success (or even outright failure) of many agroforestry development projects has led some authors to question whether such interventions can have any positive impact at all (Kessler and Breman, 1991). Part of the problem has been overambitious expectations on the pan of those planning such programmes. Experience has shown that agroforestry is not a panacea for the SAT, just as any other single technology is not a panacea.

Better results will be obtained when agroforestry is seen as merely one of the many tools available to agriculture and forestry development workers and researchers. For maximum success, agroforestry technologies should be addressed to a limited range of problems and products, and in the future must be designed to mesh closely with local farming realities, needs and constraints. The forestry aspect of the discipline must stress that a given tree species, like any other plant species, has specific site requirements. When these are lacking, the growth of outplanted trees will be retarded, thus obviating any beneficial functions that the agroforestry system might have had. Similarly, the agronomy aspect of the discipline must stress that there are numerous options available for improving farming systems, and these should be considered along with (or even in place of) agroforestry interventions.

Betel vine (Piper belie) grown in Maharastra, India on a support of nitrogen fixing tree species


CTFT. 1988. Faidherbia albida (Del.) (monograph). Nogent-sur-Marne, France, Centre technique forestier tropical.

Deb Roy, R. & Pathak, P.S. 1983. Silvipastoral research and development in India. Indian Rev. Life Sci., 3: 247-264.

El-Kankany, M.H. 1986. The importance of shelterbelts in Egyptian agriculture. In D.L. Hintz and J.R. Brandle, eds. Proc. Int. Symp. Windbreak Technol., Lincoln, Nebraska, 23-27 June 1986. USDA Great Plains Agric. Council Publ., No. 117. USDA, Lincoln, Nebraska. 303 pp.

Geiger, S.C., Vandenbeldt, R.J. & Manu, A. Variability in the growth of Faidherbia albida (Del.) A. Chev. (syn. Faidherbia albida (Del.)): the soils connection. Soil Science. (in press)

Guyot, G. 1967. Intérêt et dangers des brise vent en agriculture. 10e Colloque sur les plastiques en agriculture, Angers, France.

Hervouet, J.P. 1991. Faidherbia albida: a witness of agrarian transformation. Paper presented at the Int. Workshop on Faidherbia albida: State of the Art and Goals for the Future, 22-26 April 1991. Niamey, the Niger. ICRISAT Sahelian Center, Niamey, the Niger.

Jahnke, H.E. 1982. Livestock production systems and livestock development in tropical Africa. Kiel, Germany, Kieler Wissenschaftsverlag Vauk.

Jodha, N.S. 1985. Population growth and the decline of common property resources in Rajasthan, India. Pop. Devel. Rev., 2(2): 247-264.

Kampen, J. & Burford, J. 1980. Production systems, soil-related constraints and potential in the semi-arid tropics, with special reference to India. In Priorities for alleviating soil-related constraints to food production in the tropics. Proc. Workshop, 4-8 June 1979, Los Baños, Philippine IRRI.

Kang, B.T. & Vandenbeldt, R.J. 1990. Agroforestry systems for sustained crop production in the tropics with special reference to West Africa In E. Moore, ed. Agroforestry Land Use Systems. Proc. Special Session Agroforestry Land Use Systems. American Society of Annual Meeting, 28-29 November 1988, Anaheim, CA. Honolulu, Hawaii, Nitrogen Fixing Tree Association.

Kane, B.T., Wilson, G.F. & Lawson, T.L. 1984. Alley cropping: a stable alternative to shifting tiny cultivation. Ibadan, Nigeria, IITA. 22 pp.

Kerkhof, P. 1990. Agroforestry in Africa, a survey of project experience. London, PANOS 216 pp.

Kessler, J.J. & Breman, H. 1991. The potential of agroforestry to increase primary production in the Sahelian and Sudanian zones of West Africa. Agroforestry Systems, 13: 41-62.

Kort, J. 1986. Benefits of windbreaks to field and forage crops. In D.L. Hintz & J.R. Brandle, eds. 1986. Proc. Int. Symp. Windbreak Technol., Lincoln, Nebraska, 23-27 June 1986. USDA Great Plains Agric. Council Publ., No. 117. Lincoln, Nebraska, USDA. 303 pp.

Le Hourerou, H.N. 1980. Chemical composition and nutritive value of browse in West Africa. In H.N. Le Hourerou, ed. Browse in Africa, the current state of knowledge. Addis Ababa, International Livestock Centre for Africa.

Mann, H.S. & Saxena, S.K., eds. 1980. Khejri (Prosopis cineraria) in the Indian desert its role in agroforestry. CAZRI Monograph No. 11. Jodhpur, India, CAZRI. 93 pp.

Mott, J.J. & Reid, R. 1985. Forage and browse the northern Australian experience. In Plants for arid lands. Kew, Royal Botanic Gardens.

Muthana, K.O. et al. 1984. Root system of desert tree species. Myforest, 20(1): 27-41. (India)

Prajapati, M.C., Verma, B., Mittal, S.P., Nambiar, K.T.N. & Thippannavar, B.S. 1971. Effect of lateral development of Prosopis juliflora DC. roots on agricultural crops. Myforest, 20(1): 186-193). (India)

Renard, C. & Vandenbeldt, R. 1991. Bordures d'Andropogon gayanus Kunth comme moyen de lutte contre l'érosion éolienne au Sahel. Agronomie tropicale. (accepted for publ.)

Singh, R.P., Vandenbeldt, R.J., Hocking, D. & Korwar, G.R. 1989. Alley farming in the semiarid regions of India. In B.T. Kang and L. Reynolds, eds. Alley farming in the humid and suhhumid tropics. Proc. Int. Workshop, Ibadan, Nigeria, 10-14 March 1986. Ibadan, Nigeria, IITA.

Sumberg, J. 1990. Protecting natural regeneration in agricultural fields. CARE Agriculture and Natural Resources Technical Report Series, No. 2. New York City, CARE.

Sur, H.S. 1986. Role of windbreaks and shelterbelts on wind erosion, moisture conservation and crop growth - an Indian experience. In D.L. Hintz and J.R. Brandle, eds. Proc. Int. Symp. Windbreak Technol., Lincoln, Nebraska, 23-27 June 1986. USDA Great Plains Agric. Council Publ., No. 117. Lincoln, Nebraska, USDA.

Swindale, L.D. 1982. Distribution and use of arable soils in the semi-arid tropics. In Managing Soils Resources - Plenary Session Papers. Proc. 12th Int. Congress Soil Sci., 816 February 1982, New Delhi.

Ujah, J.E. & Adeoye, K.B. 1984. Effects of shelterbelts in the Sudan savanna zone of Nigeria on microclimate and yield of millet. Agric. Forest Meteorol., 33: 99-107.

Vandenbeldt, R.J. 1990. Agroforestry in the semiarid tropics. In K.G. MacDicken and N.T. Vergara, eds. Agroforestry: classification and management. New York, John Wiley & Sons.

Vandenbeldt, R.J. & Williams, J.H. Faidherbia albida shade, soil temperature and the growth of millet (Pennisetum glaucum). Accepted for publ. in Agric. Forest Meteorol.

Weber, F. & Hoskins, M.W. 1983. Agroforestry in the Sahel. A Concept Paper based on the Seminar on Agroforestry 23 May-9 June 1983, Niamey, the Niger, sponsored by the CILSS, AID and Virginia Polytechnic Institute. Blacksburg, Virginia Polytechnic Institute. 102 pp.

Previous Page Top of Page Next Page