Previous Page Table of Contents Next Page

Chapter 4 - Sustainable crop rotations

African examples
Asian examples
Australian and mediterranean examples
Rotations and soil productivity

Papendick et al. (1977), Beets (1978; 1982), Steiner (1984), Palaniappan (1985) and others have brought together information on cropping patterns, including crop rotations and crop arrangements within multiple crops. Much has been written on the theoretical and practical benefits of mixed cropping and on its problems (e.g., competition). This work is not reviewed here but the aim is to describe briefly some of the cropping patterns currently used in the semi-arid tropics and contrast them with systems used in semi-arid temperate regions. The aims are to draw attention to practical constraints and to identify opportunities for more-sustainable dryland crop rotations in wet-and-dry environments.

The systems chosen are in Africa (the Sahel, east Africa and Ethiopian highlands), south Asia (India and east Indonesia) and in temperate Australian and Mediterranean regions. All include livestock (Table 36). The climate varies within all these areas. Rainfall increases north to south in the Sahel, and changes with relief in the Rift Valley, and in the mountains and rain-shadow areas in E and NE Africa. It increases from east to west in Indonesia and is lower inland than on the coast in India and Australia. In mountainous country in Ethiopia, tropical crops are replaced by temperate species as altitude increases. In contrast, temperate species only can be grown in the wet, cool season climates of much of Australia and the Mediterranean.

In labour-intensive Africa and Asia, villages are the units of production. They can be either traditional or, as in Ethiopia, reformed government-directed settlements. In Australia, the farm family is the decision-making and labour unit, whether they own or manage the properties. These differences in organization and ownership are relevant to sustainability. Where a collectively-determined cropping strategy is not sustainable, migration from the village is accompanied by necessary (often slow) changes in land use, or the village is abandoned. In extensive, mechanized, family-managed systems, economic or biological non-sustainability causes changes in ownership which are quickly followed by changes in cropping practices.

Within labour-intensive systems food security has high priority. This and family under-employment encourage the focusing of labour near the house. The importance of the household garden thus varies with the size of the property (Harwood 1975). This is more apposite in Asia and the wet tropics than in the wet-and-dry tropics, where sequential cropping of vegetables under a household tree crop (using, if necessary, hand-carried water), gives way to a seasonal vegetable crop followed by fallow in the dry season. When villages are moved and land use is changed to commercial agriculture, recreation or forestry, as has occurred to a marked extent in Thai highlands, the intensively-used household gardens are abandoned and the overall intensity of management and the fertility of the system is reduced.

Components of cropping systems illustrated in this chapter


Major crops

Major animals

Feed source

W. Africa Sahel

Millet/sorghum, maize, groundnuts, cowpea, sesame, cassava, yams, tree legumes

Cattle, goats, sheep, poultry

Fallow weeds, crop residue, vines, tree crop cuttings, oil cake, cull tubers

E. Africa

Maize/barley, sorghum, millet, tef (Ethiopia)

Donkeys, cattle, goats

Crop residues, fallow weeds, tree cuttings

S. Asia

Sorghum/millet (India), maize/rice (other); cassava; kenaf, wheat, groundnut, soybean, chickpea

Cattle, buffalo, goats, sheep

Bran, oil cake, crop residues

S. Australia

Wheat/barley; lupins, peas (subterranean clover, annual pasture medics)

Sheep, cattle

Pasture ley, crop residues

Land utilization by district, Sine-Saloum, Senegal (Source: Pieri 1992)








Total area (km2) (ST)


2 959

2 330

11 847

1 880

2 277

23 939

Rural population

186 973

98 629

139 775

283 868

136 602

154 829

1 000 678

Density (population/km2)








No. of farms

20 393

9 781

15 351

29 635

15 486

13 969


Mean pop./farm










71 300

57 800


157 500

62 000

114 500

528 000


36 700

8 100

22 400


1 250

5 600

140 300


13 300

2 900


11 900


2 500

88 460


176 000

207 500

75 000

100 500


122 500

751 500


Cultivated (ha) 83/84

93 649

96 575

151 379

285 386


145 965

870 956

Proportion cropped:

Millet and sorghum
























Area/inhabitant (ares)








Area needed for wooda (km2)

37 395

19 725

27 955

56 774

27 324

30 966

200 139

Area needed for livestockb (km2)

256 520

184 460

172 120

435 860


265 700

1 489 820

Total Areas (T)

387 564

290 760

351 454

778 020

310 484



Degree if exploitation (T/ST%)








a Norm at 0.2 ha forest to cover fuel needs for cooking and timber for building.
b Based on 1 ha pasturage per month of the dry season per LU (livestock unit). As animals are partly fed from fodder crops and harvest residues, values used are: 2 ha/LU (cattle and horses), 0.4 ha (mules). 0.2 for sheep and goats.

African examples

Key resource issues for sustainability in Africa

In substantial areas of semi-arid west Africa and the highlands of Ethiopia the present land use is not sustainable. The areas needed for food crops, for livestock and fuel exceed the actual land available. This is illustrated in Table 37 for selected districts in Senegal (Pieri 1992). Pearson (1992) and Pol (1993) draw similar conclusions for some districts in Ethiopia and Mali.

It is unrealistic to suggest that changing crop rotations will provide the quantitative answer needed in areas such as those in Table 37. The most obvious answer is to remove livestock from the system. According to Table 37 they need one to two times the area currently under food crops. Animals are kept for prestige, for their products and for traction. Substitutes for their products (e.g., soy milk and meal) might be provided more cheaply, in real terms, if soil sustainability is assigned a value. They could be traded for crop surpluses. Grain legumes can be substituted for some livestock products and increased planting of a wider variety of legumes would improve the sustainability of the cropping system. This has already be done elsewhere, for example in parts of east Indonesia.

The value of draught animals is also debatable. Sandford (1989) interprets conflicting data to suggest that farmers often use animal traction to increase the area cropped, rather than for more timely management of existing land to increase yields. If this is so, livestock reduce productivity in two ways: directly through their need for food and through their traction being used to increase the area, but not necessarily the management skill.

There are two obvious difficulties when recommending removal of livestock from a district. First, the livestock may be owned by people other than those who cultivate. In these circumstances, integration of livestock with cropping is unlikely. Livestock removal usually involves relocation or re-employment of the herdspeople, neither of which is easy. Secondly, it is not easy to identify and foster peasant-based industries to generate cash for fuel and livestock products such as milk.

Conventional wisdom is turned upside down by the suggestion that, in densely populated regions where fallows are short (<4 years) and land use is not sustainable, land used for livestock and fuelwood should be turned to food crops. For example, recent inventories and land assessments for east Africa place equal priority on livestock, fuel and food (Kassam et al. 1991). Such equal emphasis is surely inconsistent with the belief that the priority product of a cropping system is food security (Chapter 1). Similarly, and perhaps underlying the notion that fallow-for-forage has equal status with cropping, is the widely-repeated belief that soil degradation is caused by shortened fallow periods (e.g., Brown and Thomas 1990) and, therefore, that sustainable cropping depends on some minimal fallow period between crops. Many current land assessments include calculated 'fallow requirements' (e.g., Kassam et al. 1991). Fallow as such, however, does not restore soil structure, biological activity or nutrient status. A fallow, long enough to allow re-establishment of deep-rooted perennial plants crops that draw nutrients from depth and maintain soil surface cover, will improve soil productivity provided that nutrients are not removed (e.g., as milk or meat) from the system. This rarely holds nowadays. Most commonly, a short fallow of weedy annual species leaves the soil bare and liable to erosion in the dry season, there is structural degradation through trampling and plant uprooting, and insufficient soil water is conserved for the next crop. There is no reason why a well-managed continuous cropping system should be less sustainable than one including unmanaged short fallows from which nutrients are removed by livestock or as fuel.

FIGURE 31 - Seasonal calendar of a peasant association in Wollo, Ethiopia (Source: Ethiopian Red Cross Society, pers. comm. 1988)

As well as taking up land for their feed, animals are a key constraint on crop rotations. Poulain (1980) describes a typical example, in the Serere, Senegal: 'the system comprises a typically triennial rotation and the integration of cropping with cattle rearing and fodder trees'. The herd is enclosed in one field during the crop-growing season and immediately after harvest the cattle are released to graze on crop stubble and Acacia albida trees. Around the village there are permanent fields of millet, cowpeas and spices fertilized with household refuse while more distant from the settlement the rotation is typically groundnut-millet-fallow.

Seasonal cropping and grazing patterns for mid-altitude Ethiopia are shown in Figure 31. The specific food crops favoured in Ethiopia vary with altitude (Pearson 1992), but the seasonality of crop growth, livestock needs and labour requirements, shown in Figure 31, are broadly typical. This figure also illustrates the pressure on human resources in the semi-arid tropics during the period around the start of the wet season, when the highest demand for labour (for tilling and planting) occurs at a time when human diets are deteriorating and there is a high incidence of debilitating diseases.

Local distributions of cropping largely radiate from the house. As shown in Figure 25, gradients of organic matter and nutrients from the house create semi-permanent gradients in crop rotations. Continuous cropping nearest the house is sustained by replenishment of nutrients using animal and human manures and food residues. Long-season crops (e.g., maize), and/or those favoured socially (e.g., t'ef, in Ethiopia) are grown nearest the house. Examples of spatial variations in crop rotations are illustrated in Figures 32 and 33.

There is a wide diversity of cropping systems in dryland Africa. There are 'parkland systems' in semi-arid west Africa. On the arid margin of cropping in north and east Africa there are traditional cultivated fields among bush-fallow as well as parklands. The parklands of the Sahelian zone are predominantly Acacia albida with food crops, particularly pearl millet and groundnuts. These are structurally and functionally similar to the Prosopis cineraria/pearl millet parklands of parts of India.

One crop or, at best, two in relay, can be grown in the seasonally wet-and-dry climate of the parkland that has a perennial tree component. The cultivated ground layer carries an annual food crop, the stubble of which is grazed leaving the ground more or less bare through the dry season. The land used for food crops is traditionally cultivated, often by first cutting-back trees and bushes at the start of the wet season. The soil is ploughed and soil and organic matter piled into mounds, which are then burnt. The ash from the mounds is then spread and the soil re-tilled as often as necessary. For t'ef in Ethiopia, there may be as many as nine tillage operations spread over three months, with consequent loss of cropping opportunity (soil water which is not used). Tillage and mounding are labour-intensive. They destroy topsoil structure and disrupt the continuity of macropores (Chapter 2). Nonetheless, and despite the scientifically-demonstrated benefits of maintaining soil surface cover, traditional tillage can give higher yields than minimum- or zero-tillage. As shown in Figure 15, the benefits from zero tillage may take several years to accumulate so it is not surprising that one-year experiments in the Sahel failed to demonstrate the benefits of minimal soil disturbance. Ridging may also be practised. This helps to reduce wind erosion and wind-caused seedling death, improving establishment by 50% in experiments. While the benefits of ridges seem intuitive, it does not follow that the ridges need to be broken and re-formed each year. Permanent ridges have the advantages of both reducing wind erosion and maintaining undisturbed topsoil.

FIGURE 32 - Homestead and bush field land use and cropping pattern at Dalung, Benue Plateau (Netting 1968)

The main food crop is usually pearl millet but sorghum and maize are grown in more favoured areas. The cereal crop is usually planted between the trees in hills or pockets of up to ten seeds. The seedlings are thinned or compete with each other so that commonly only 2-3 survive to maturity. A low plant density, commonly 5000 hills, is equivalent to between 10 000 to 15 000 plants per hectare. In these circumstances, on-farm experiments suggest that inorganic fertilizers may give only marginal economic returns. Again, the dilemma raised in Chapter 1 is illustrated: biological sustainability may not equate with short-term economics. As Figure 34 illustrates, plant populations need to be increased substantially (with attendant risks of population-induced water deficits, in dry years) to get economic benefit/cost ratios from applied inorganic nutrients, at least in the year in which the fertilizers are applied. However, this and similar analyses, highly relevant to short-term decisions of individual farmers, do not allow for: (a) residual benefits from fertilizers in the seasons following application of phosphorus and nitrogen, as found by Bationo et al. (1992); or (b) indirect benefits from the fertilizers, such as, say, reduction of the incidence of parasitic weeds (e.g., Striga) and the increased effectiveness of nodulation by rhizobia in subsequent legume crops.

Millet and sorghum yields are generally higher when grown the year following a grain legume. Though, on poor soils, legumes such as cowpea in rotation with millet have little effect on the millet yields if phosphorus fertilizer is not added (Fussell 1992). The beneficial effects of grain legumes are somewhat disputed. There is substantial evidence showing relatively small residual nitrogen benefits from groundnuts and soybeans. Australian systems are similar in this respect. The benefits, associated with relatively high nitrogen fixation, only accrue when nutrients, particularly phosphorus, are adequate for both rhizobia and crop growth. The systems require small inputs of inorganic phosphorus for sustainability.

FIGURE 33 - Cropping schedules and pattern of planting for some crop combinations in northern Nigeria (Source: artificial composite of figures from Norman 1973)

Fertility is increased under tree legumes by: (a) fixation of dinitrogen gas from the soil atmosphere; (b) the gathering of livestock and humans beneath their shade so concentrating their manure; (c) recycling of nutrients from the deep subsoil by leaf fall; and (d) trapping and accumulation of nutrients moved laterally by wind or water from elsewhere. Since soil fertility is higher beneath the trees, crop yields may also be higher (though in dry years yields can be expected to be lower because of the water used by the trees). Thus, farmers often sow sorghum under the trees to take advantage of the improved soil conditions which are not so well exploited by pearl millet, with its lower yield potential.

Figure 35 illustrates the nitrogen benefits from the prunings of tree legumes. Juo and Payne (1993) review the use of tree legumes as windbreaks to alter the crop micro-environment. They conclude that there are beneficial effects of the trees other than nitrogen enrichment, but go on to question the use of tree legumes as alley crops in the semi-arid tropics: 'In West Africa, drought stress and acute P deficiency are major factors limiting the growth and establishment of leguminous trees and shrubs at closer (than naturally-occurring) spacings. For example, even at row widths of 6 to 8 m, nearly double those recommended for humid and sub-humid conditions, marked yield depression was observed when alleys are pruned up to 3 times during the short cropping season in semi-arid India. Thus, the adoption of alley cropping systems in the semi-arid tropics would largely depend upon the identification of specific locations where soil texture and fertility level, groundwater table and rainfall distribution are more favourable for alley establishment'. A similar view led to targeting existing fencelines as the most realistic sites for the introduction of tree legumes in Ethiopia (Table 38).

FIGURE 34 - Effect of plant density and fertilizer use options on millet yield in farmers* fields where Y = grain yield kg/ha, D = plant density (103 pockets/ha) and PN represent P fertilizer only and P + N fertilizer codes with values of 1 and 0 represent treatments with and without fertilizer respectively (Source: Bationo et al. 1992)

FIGURE 35 - Partitioning of nitrogen from prunings of Leucaena leucocephala to crop, tree and soil in an alley-cropping system on an Lixisol at Ibadan, Nigeria (K. Mulongoy, unpublished data)

Strategies for introducing forage legume into the Ethiopian crop-livestock systems (Source: Pearson 1992) (Species are speculative until further field evaluation.)


Low altitude <2000 m

Medium 2000-2400 m

High >2400 m

Forage strips


Leucaena (Leucaena leucocephala), Sesbania (Sesbania brachycarpa), pigeon pea (Cajanus cajan)

Tagasaste (Chamaecytisus prolifer), sesbania Pigeon pea



Siratro (Macroptilium atropurpureum), silver leaf desmodium (Desmodium intortum), greenleaf desmodium (Desmodium uncinatum), Verano stylo (Stylosanthes humilis)

Desmodium (greenleaf)Axillaris (Axillaris spp.) white clover, vetch


Backyard forage


Leucaena, sesbania, pigeon pea

Tagasaste, sesbania pigeon pea



Desmodium, alfalfa (Madicago sativa)

Alfalfa, vetch


Undersown and relay legumes in crops


Lablab (Lablab purpureus), cowpea (Vigna unguiculata), siratro, greenleaf desmodium, Verano stylo, vetch (Vicia sativa), Cassia rotundifolia

Vetch, desmodium


Oversown legumes in pasture and fallow


Verano stylo, Siratro



Livestock exclusion areas


Leucaena, sesbania

Tasasaste, sesbania



Siratro, stylos, Axillaris

Axillaris, siratro


Conventional pastures and special-purpose forages


Stylo (Verano), siratro, desmodium (greenleaf and silverleaf)

Desmodium (greenleaf and silverleaf), vetch

Alfalfa, vetch

Others (e.g., Kerkof et al. 1990) rightly draw attention to the need for technical packages which incorporate various options for tree legumes: intercropping and alley cropping; planting along contour lines and farm boundaries; as wood lots; as wind-breaks and the promotion of natural regeneration of trees in pastoral areas. It is best for the farmers to decide which option fits their goals.

FIGURE 36 - Scheme to illustrate how observed deterioration in agriculture may be linked back to root causes, which are themselves interrelated. Here, observed declining productivity and stability are linked back to key attributes which were discussed in earlier chapters.

To conclude this overview of African dryland systems, it is appropriate to consider their constraints. Obviously, the main constraint is the rainfall amount, its distribution and reliability. These severely limit the opportunities for increasing productivity per unit area. The limitations, nonetheless, can be tackled effectively using a range of strategies. For example, in a region of Somalia, with a rainfall of 580 mm in a good year, four practices for better soil water use were tried (ridging; bunding; intercropping with a grain legume; incorporation of residue from the previous cereal crop). Even in a season of above-average rainfall, all increased the yield of sorghum, the main crop, by 30-50% (Eagleton et al. 1991). Clean fallow (a crop- and weed-free season), with the aim of conserving water for the next crop, did not improve the sorghum yield. In areas with higher and more reliable rainfall, opportunities for water conservation and relay cropping increase.

Key resource issues for sustainability in Africa

In regions where single crops are the most realistic expectation, the seven aspects outlined below may be considered 'key local resource issues which are involved in production and sustainability' (Figure 4). They are not listed here for prescriptive purposes. Any prescriptive list should be compiled by people working and living in the relevant regions. They can, however, be used as realistic examples of the resource issues/problems/opportunities related to sustainable rotations for the semi-arid tropics. It is useful to link these issues schematically, and attempt to make their root causes explicit, perhaps through a flow diagram as in Figure 36. Similar diagrams could provide a basis of discussion and, perhaps, action which may lead to changes in cropping practice and agreement with farmers as to how to monitor the sustainability of their system (Figure 4).

1. Too large a proportion and area of land is used for livestock or fallow which is marginally beneficial for livestock and fuel production. Reducing livestock numbers and/or keeping them penned in a cut-and-carry forage system, rather than allowing them to graze freely for most of each year, frees land for food cropping. Fallow as such confers no positive benefits to the sustainability of cropping. Indeed, contrary to the implication in Kassam et al. (1991), poorly-managed short fallows grazed by livestock can be associated with degradation of soil structure and chemical depletion.

2. There are too few local markets to generate cash so that farmers aiming for food security could purchase livestock-products and fuel. There is a need for more markets to overcome the constraint described in 1 above. There are opportunities to add value, for example small-scale processing of soybean meal. Such steps are immediate and create local employment, and are presumably more realistic than suggesting that semi-arid Africa tries to compete internationally with, say, the USA for markets in soybeans. Subsistence farmers competing with mechanized agribusiness may be likened to a runner competing against a fast car.

3. The health and strength of human (and animal) labour varies seasonally. Labour is needed for tillage when debilitating disease is most likely and when seasonal deterioration in diet starts. Sowing and weeding coincide with the seasonal minimum in diet. Thus, human labour constraints enhance traditional conservatism with respect to maintaining uncropped areas and, combined with traditional high intensity tillage, reduce the area of crops sown and make sowing untimely.

4. For cropping systems to be sustainable, nutrients other than nitrogen need to be imported. This constraint cannot be overcome even if recycling of manures is made more effective and degrading practices such as mounding, burning and ploughing are stopped. It follows that cash generation (2, above) is essential to increase purchases of fertilizers.

5. To sustain nitrogen levels, crop legumes are essential. Their absence near homesteads needs to be balanced by giving legumes a dominant role in rotations in outer fields.

6. Parts of the cropping system with herbaceous legumes do not currently dominate cropping systems in dryland Africa nor are they likely to with present species and cultivars. Groundnuts are highly competitive and have poor water-use efficiency. They, however, have an appreciable, largely un-exploited, variation in nitrogen-fixing capacity (Giller et al. 1987). Cowpeas have a longer growing season and are more susceptible to pests and diseases than is desirable. For legumes to dominate part of a cropping system there is a need to move, for example, to 1:1 intercropping with cereals and 3:1 years legume/cereal rotations to maximize the net benefits of legumes to soil structure and fertility. The adoption of such rotations requires: (a) legumes that can be used as an alternative to groundnuts; (b) legumes which are promiscuous, that is able to be infected by many naturally-occurring rhizobia; and (c) legumes which are more effective at nodulating and can benefit more from their rhizobia. These opportunities are developed in Chapter 5: they can be seen as part of a system that focuses on the improvement of a few selected grain legumes, and their processing. In this way the present causes of non-sustainability listed in 1 to 4 above can be addressed

7. Naturally-occurring tree legumes that have been sustainably managed by trimming and browsing can be augmented with the same species or introduced species, particularly as part of stable land systems, along fencelines or on permanent soil ridges.

Traditional and improved crops and cropping systems for selected locations in the semi-arid tropics of India (Source: Singh and Reddy 1988)



Stable cropping systems






Jodhpur (380 mm, 11 weeks)

Pearl millet
Moth bean
Cluster bean
Mung bean

Hybrid, pearl millet
Improved mung bean
Castor bean
Cluster bean

Green gram or cluster bean/pearl millet (BJ 104)
Cenchrus ciliaris mung bean (T 44) (normal rain)
Cenchrus ciliaris cluster bean (FS277) (for >500 mm rain)

Pearl millet-fallow
Pearl millet (BL 104)-mustard (T 59) (for >500 mm rain)

Hisar (400 mm, 13 weeks)

Pearl millet
Cluster bean
Mung bean

Hybrid pearl millet
Cluster bean
Improved mung bean

Pearl millet/mung bean
Pearl millet/cowpea (fodder)

Pearl millet-chickpea
Mung bean-mustard

Shallow Luvisols

Anantapur (570 mm, 14 weeks)

Foxtail millet

Pearl millet or sorghum
Mesta (rozella)

Groundnut (Kadiri-1)/pigeonpea (PDM 1)
Groundnut/castor bean
Pearl millet/pigeonpea


Medium Vertisols

Rajkot (625 mm, 17 weeks)

Pearl millet


Groundnut (J-11)/ Groundnut (J-11)/ pigeonpea
Cotton/green gram


Shallow Luvisols

Hyderabad (770 mm, 17 weeks)

Castor bean
Pearl millet

Castor bean
Foxtail millet
Pearl millet

Pearl millet/pigeonpea
Castor bean/cluster bean
Pigeonpea/mung bean


Deep Luvisols

Bangalore (890 mm, 32 weeks)

Finger millet
Horse gram

Finger millet

Finger millet (PR 202)/ soybean
Finger millet/maize or pearl millet (fodder)

Cowpea-fingar millet

Shallow to medium Vertisols

Solapur (722 mm, 23 weeks)

Pearl millet

Hybrid pearl millet

Pearl millet/pigeonpea

Pearl millet-chickpea
Mung bean-rabi sorghum

Bijapur (680 mm, 17 weeks)

Pearl millet

Hybrid pearl millet
Foxtail millet
Green gram

Pearl millet/pigeonpea
Chickpea, safflower

Green gram-rabi sorghum
Green gram-safflower

Deep Vertisols

Hyderabad (770 mm, 25 weeks)




Sorghum-saff lower

Bellary (500 mm, 8 weeks)

Rabi sorghum

Rabi sorghum
Field beans


Asian examples

The largest areas of seasonally wet-and-dry climate in south Asia are the 12 zones (Hutchinson et al. 1992) in India and the II and 12 zones in Laos and Cambodia (Figure 6). There are also much smaller, densely populated, areas, in east Java and the eastern islands of Indonesia.

Venkateswarlu (1993) and others review rotations in the Indian semi-arid tropics. As in west and north-east Africa, tree and shrub legumes are sometimes present and planting of elite crops such as pigeonpea, (Cajanus cajan) are encouraged. Annual crops, however, are dominated by pearl millet and sorghum, as in the Sahel. Table 39 lists traditional crops and their association with the length-of-growing season. As explained earlier this varies substantially with soil water-holding capacity in areas having the same annual rainfall. Maize is grown on favoured microsites, on soils with relatively high water-holding capacity (cf. Figure 9) and on the wetter margins of the region.

Animals are important in Asian farming economies, cut-and-carry systems of feeding being more common, for example, in Indonesia than in India or Africa. It is common practice at the end of the wet season, before the crops are harvested and residues available to livestock, for farmers to strip leaves from the maize or sorghum during grain filling. As Pearson and Hall (1984) have shown, lower leaves can be removed to provide nutritious feed without harming the food crop, but removal of upper leaves, which is not uncommon, reduces grain yield.

Management of two cropping patterns commonly used in upland areas of Indonesia

Cropping pattern


Spacing (m)

Population per ha (1000 plants)













- - - maize - - -



2.0 x 0.4


- - upland rice - -


upland rice

0.4 x 0.2


- - - - - - - cassava - - - - - - -



4.0 x 0.4


- - - maize - - - -



2.0 x 0.4


- - - upland rice - - -


upland rice

0.4 x 0.2


- - - - - - - - cassava - - - - - -



4.0 x 0.4




- - - rice beans - - -


0.2 x 0.2




0.4 x 0.2


Relative yield of transplanted and dry-seeded rice obtained from on-farm trials in Iloilo and Pangasinan, Philippines, 1977 (Source: Gomez and Gomez 1983)


Seedling establishment

No. of fields

Grain yield (t/ha)






Interdisciplinary Research Team 1978




Interdisciplinary Research Team 1978





Gines et al. 1978




Gines et al. 1978

As in Africa, trees for firewood are important in India. Shortage of fuelwood has caused other communities (e.g., in east Java) to trade crop surpluses for petroleum products. Budgets of fuel and forage needs for farms and regions are published by Srivastava and Rao (1989) and others. Where community populations are sufficiently sparse, it is realistic to recommend one-fifth of available land for silviculture and a substantial proportion for tree alley crops (16%). Planting of trees for fuel and forage on 'alternative land' which has been degraded and is unstable, should be seen as a conservation measure and not as a cost against food cropping.

Tandon and Rego (1989) review crop responses in India to inorganic nutrients other than nitrogen. They conclude that responses are often large and that phosphorus and zinc are of major importance, though yield responses to potassium and sulphur are also common on some soils and iron chlorosis is an emerging problem.

Published figures giving areas of grain legumes (mostly chickpea, cowpea and gram), suggest that cereal-legume rotations and intercropping mixtures are, in practice, about 4:1 cereal and oilseeds: legumes (Venkateswarlu 1993). Thus, cropping systems in India have similarities with the Sahel though differing soils, social factors and markets should caution against extrapolating, or trying to transfer, solutions from one to the other.

Semi-arid cropping systems in India, and to a greater extent in Indo-China and east Indonesia, use dryland or upland rice when possible during the wet season. Table 40 shows typical dryland cropping patterns in east Java, with rice the preferred wet season crop if the length of the growing season (duration of rainfall), or the availability of runon water, permit. Otherwise, monoculture maize is grown, with little interplanting of grain legumes. The grain legumes are used as a second, relay crop, often hand-broadcast into the rice stubble or as a marginal wet-season crop, planted late and on less-favoured fields.

The advantage of semi-arid south-east Asia over, say, the Sahel, is its varied topography, so that many areas are able to capture run-on water which allows extension of the growing season. Thus, dryland rice grown with a significant period in standing water is possible. In India, ICRISAT has successfully popularized water harvesting to create local areas suitable for rice or relay cropping in less hilly regions. This increases both crop diversity and economic and ecological stability.

Traditionally, rice has been grown largely by transplanting seedlings from nurseries. Societal change, driven by women preferring to do piece-work for cash instead of maintaining transplant gardens, is encouraging direct seeding. This has several advantages. It shortens the crop life cycle allowing two crops in some situations (Gomez and Gomez 1983). It is less labour intensive and avoids transplant shock. Transplanted rice has no inherent yield advantage over directly-seeded rice (e.g., Table 41).

A list of constraints on and opportunities for dryland cropping in south Asia might identify similar key issues to those listed in the previous section. Livestock is segregated from cropping in parts of the region, though notably not in India. The absence of livestock and dearth of trees for fuel elsewhere in Asia, however, illustrate that communities may move from devoting land to livestock and fuel crops to essentially food-crop only systems. Recognition of the benefits of tree legumes and their adoption seems broadly feasible. Within the food crop component of the cropping system, there remains the problem of having insufficient legume cropping to maintain a neutral nitrogen budget. The productivity of the herbaceous legumes currently grown is of the order of 0.3-0.9 t/ha only (Venkateswarlu 1993). Thus, to achieve sustainable cropping in south Asia attention is needed to the following constraints (Compare with Item 6; previous section):

1. The proportion of herbaceous legumes in crop mixes, and the effectiveness of nitrogen fixation (residual value) of the legumes need to be increased.

2. Visual symptoms should be used to diagnose the need for inorganic fertilizers, particularly those containing phosphorus and zinc.

Soybean and mungbean are currently grown, and may be more nitrogen-effective and less degrading than groundnuts. They are, however, less extensive than cereals, which suggests that more legumes should be grown as intercrops. Table 42 summarizes potential intercropping systems, including crop legumes, shown by ICRISAT to be agronomically viable.

Promising intercropping systems that include legumes in India


Row ratio

Potential areas

Sorghum + pigeonpea


Semi-arid red soil regions of southern Telengana, semi-arid black soils of Vidarbha region and semi-arid black soils of Malwa plateau

Pearl millet + pigeonpea


Semi-arid black soils of Deccan region of Maharashtra as well as Karnataka and semi-arid black soils of Saurashtra

Pigeonpea + soybean


Semi-arid/sub-humid black soils of Madhya Pradesh and sub-humid red soils of Chotanagpur region

Sorghum + mung bean


Semi-arid black soils of Bidarbha and Marathwada region

Pigeonpea + sunflower


Semi-arid black soils of Deccan region of Maharashtra

Finger millet + pigeonpea


Sub-humid red soils of Orissa and semi-arid red soils of southern Karnataka

Groundnut + pigeonpea


Semi-arid red soils of Rayalaseema

Maize + mung bean


Sub-humid submontane regions of northwest Uttar Pradesh and Jammu and semi-arid black soils of Malwa plateau

In parts of Asia, for example northeast Thailand and east Indonesia, root crops, particularly cassava, are grown extensively for local consumption, local processing (e.g., into packaged chips) and for export (e.g., as stock-feed to Europe). The profitability of root crops for export varies widely. Their local use, however, and their good storage properties and transportability once processed, suggest that root crops will remain a component of semi-arid cropping. Inter-planting of a non-competitive legume crop shortly after planting cassava stakes could substantially increase the area sown to grain or forage legumes. The low yields of legumes (e.g., Venkateswarlu 1993) suggest a need to improve both inorganic nutrition and rhizobial effectiveness.

Australian and mediterranean examples

Cropping in temperate dryland climates (E2, Figure 7) is based on cereals, mainly wheat and barley, as is cropping in the wet-and-dry regions of north east Africa above 2400 m and in the Near East. Although the climates in Australia and the Mediterranean areas under discussion are similar they have different traditions, the Mediterranean systems being aimed towards village food security until recently while the Australian systems, developed since the 1880s, are aimed at producing bulk grain for export. Farm size differs, Australian mechanized farms averaging 2800 ha (1983/4). Mediterranean systems recognize traditional practices and farmers' values, and there is more local consumption of food crops, so there is a variety of cropping systems around the theme of cereal/fallow or cereal/hay crop/fallow (Table 43). Australian systems have developed more limited options around crop/legume pastures or crop/volunteer weedy species rotations (Table 44). On the Australian dry margin and on fine-textured soils the systems usually include a long fallow (9-12 months), primarily to conserve water but also to assist the mineralization of nitrogen, and control weeds and diseases. The efficiency of the system, as assessed by water use (WUE), generally ranges from 10 to 16 kg grain/ha/mm, which is very similar to that of cereal crops in the United Kingdom and Syria.

Main rotation patterns in Mediterranean cropping systems (Source: Lopez-Bellido 1992)


Main rotation


Cereal-hay crops-cereal

Fallow-cereal-cereal-fallow (has grazing value)

Cereal-tobacco, sugar beet, grain legumes


70% cereal-cereal

10% cereal-hay crops

2% cereal-grain legumes

18% with other alternatives (cotton, sugar beet, tobacco)


Cereal-fallow (falow has grazing value)


Fallow-cereal-subterranean clover

Annual cropping:


Cereal-cereal-hay crop

Cereal-sunflower (grain legumes are being abandoned due to broomrape)

Spain (rainfed)

40% fallow-cereal

25% annual cereal

10% cereal-sunflower

10% cereal-hay

5% cereal-grain legumes

10% others (mainly grazed forage)


80% fallow-cereal (fallow has grazing value)

15% cereal-hay crops

5% cereal-grain legumes


Usually 3-year rotation including cotton, maize, rice



low rainfall

25% fallow (has grazing value)

75% continuous cereal


high rainfall

15% continuous cereal

79% grain legumes

15% fallow (grazed)


16% fallow-fallow-cereal

42% fallow-cereal

17% cereal-cereal

14% cereal-hay crops

11 % cereal-grain legumes


20% fallow

6% continuous cereal

20% cereal-hay crops. Nigligible grain legumes


Fallow (no grazing value) exists only in low rainfall.

The majority of the area is cereal-grain legumes or cereal-vegetables


Little remaining fallow. Where it exists it has grazing value except in low rainfall zone

Rotations are wheat-barley-wheat-alfalfa and wheat-lentils



low rainfall

95% fallow-cereal

5% grain legumes


high rainfall

50% continuous cereal

10% hay crops

10% grain legumes

30% others

Major small-grain production systems of Australia subdivided on rainfall and soil texture

Winter rainfall

Summer rainfall

Low (<300 mm)



Coarse-textured soils

Fine-textured soils


















Key: C, cereal crop; GL, grain legume; SF, summer fallow; VP, volunteer pasture; F, fallow; P, legume pasture.

FIGURE 37 - Rhizobium trifolii soil populations over four seasons with wheat (1983-84) and direct-sown (and inoculated) subterranean clover (1985) and regenerating clover (1986).

Land use in Mediterranean climates presents problems for soil sustainability, particularly on hill land. These are not so much directly related to cropping practices (although contour ploughing, for example, is not uniformly practised) but to deforestation (Table 18) and depopulation of hill country (S. Orsi, personal communication). In Australia, land preparation methods vary greatly with location and season and may confer non-sustainability. Land to be cropped is usually heavily grazed shortly before the wet season; residual vegetation (pasture or weeds) is then killed with desiccant herbicide. The crop is then sown four hours to four days later either with minimum soil disturbance, or conventionally using 3-5 passes of a tined implement. There is some attendant loss of topsoil organic matter and structure. It is noteworthy that, in a system depending on the survival and growth of rhizobia, direct drilling creates lower bulk densities (Carter and Steed 1992) and maintains higher rhizobia populations than tillage (Figure 37).

Estimates of equilibrium values of the amount of soil organic matter nitrogen which builds up under various cereal crop-legume pasture rotations, and its mineralization in the crop year, in Australia (Source: Perry 1992)

Equilibrium soil N (kg/ha)

Mineralization (kg/ha/yr)1

Continuous pasture

3 350


2 crop: 4 pasture

1 925


1 crop: 1 pasture

1 450


Continuous crop



1 A decomposition constant of 0.04 is assumed, but this would vary with season, crop and cultural condition.

Australian cereal-based cropping systems need to maintain a neutral nitrogen budget through the non-cereal phase. They require addition of phosphorus and other elements as inorganic fertilizers. Organic matter levels need to be kept up. Dunne and Shier (1935) criticized the rotations used then as not maintaining fertility for cropping. As a result pasture legume periods were lengthened. From about 1950 to 1980, pastures were widely included in rotations, the period in pasture commonly being one and a half to three times that of the crop. The accumulation of nitrogen under pasture is illustrated in Figure 21. The average annual increment of soil nitrogen from 22 estimates was 67 kg N/ha/year, the highest value being 182 kg/ha/year (Perry 1992). Perry argues that the key to the functioning of the system is the rate of mineralization which makes the nitrogen available to the crop. Table 45 shows a 4-fold range in rate of mineralization among four common crop-pasture rotations.

Nutrient elements other than nitrogen, particularly phosphorus, sulphur, zinc and molybdenum, are routinely added at sowing in quantities sufficient to compensate for their removal in the grain. As well as giving economic crop yield responses in the year of application, these nutrients have indirect benefits. Heavy applications of inorganic nutrients reduce the incidence of root diseases (e.g., Brennan 1992).

The cereal-legume pasture system, so widely used from 1950 to 1980 and publicized internationally, depended on relatively favourable international prices for grains and sheep products, especially wool, and a balance between them. Successive years of drought with high grain and relatively low wool prices in the early 1980s led to intensification of the cereal crop phase. Continuous cropping became widespread and fewer legume pastures were sown. Then, in the early 1990s, following a short period of artificially high wool prices and increased sheep numbers, wool became marginally economic so sowing and management of legume pastures were again largely abandoned. Recently, concern about erosion associated with overgrazing and the extent of bare soil that occurs under annual pasture species such as Trifolium subterraneum and medicago, has shifted attention towards the inclusion of grass, particularly perennial grass, in the non-cereal phase of rotations. This is supported by research which shows that the rate of soil acidification beneath perennial grass is slower than under annual ryegrass (Lolium rigidum) pasture in a seasonally wet-and-dry environment (Ridley et al. 1990).

FIGURE 38 - Linkages between major agronomic problems or issues identified from the conduct of a topical rapid rural appraisal in the Forbes Shire in Australia (Source: Ison and Ampt 1992)

Impact of lupins on soil nitrogen under cereal and cereal-legume rotations: mean soil mineral nitrogen (NH4+NO3+NO2 kg/ha, 0-20 cm) after growing wheat (W) and lupins (L) in rotation (Source: Reeves et al. 1984)

Soil mineral nitrogen after

Year 1

Year 2

Year 3





























1 Within columns, values not followed by a common letter differ significantly (P<0.05).

The present system is arguably not sustainable, because of the changes in markets and attitudes, depending as it does on nitrogen fixed by a pasture phase that is increasingly botanically-impoverished. This is illustrated by farmers' perceptions that the nitrogen concentration of the cereal grain is declining, and that pasture and soil factors are causing this (Figure 38).

The Australian system described above is a good example of how biological and soil sustainability depends on markets, in the absence of policies of land management agreed by land managers. It shows how sustainability and agronomic practices can change rapidly. Another issue raised by recent changes in Australia is the need for adapted grain legumes. The system is likely to return to sustainability only by introduction of legume crops, particularly field peas and lupins (Lupinus angustifolius and L. albus). These fix more nitrogen than legume-based pastures and, depending on the proportion of nitrogen removed from the field as grain, usually give a residual benefit of about 20-50 kg N/ha (field peas) and 50-90 kg N/ha (lupins). The latter value exceeds that for many tropical grain legumes (Table 24). As Table 46 shows, the soil nitrogen responds positively, quickly and substantially to the inclusion of a temperate grain legume (in this case, Lupinus angustifolius in NE Victoria). Grain legumes currently occupy only 12% in west and 4% in the east of the cereal areas. These areas need to be increased to help maintain general levels of soil nitrogen, organic matter and good soil structure.

One aspect, perhaps better documented for semi-arid Australia than elsewhere is that of timeliness. As Loss et al. (1990), Perry (1992) and others have described, there is a window of opportunity for sowing at the beginning of each wet season. Any delay in sowing causes substantial reductions in grain yield. This is a universal phenomenon that emphasizes the need for:

· knowledge of the probabilities of an effective beginning to the growing season, so that sowing will be neither premature and unreasonably risky or, as is common in developing countries, delayed so much by late, excessive cultivation that yield potential is routinely lost;

· sufficient knowledge by the farmer of the best timing and type of soil treatment at sowing for plant establishment. This requires ready availability of machinery and the use where possible of minimum tillage. Where cultivation is necessary because of local conditions, it should be coupled with sowing, so that at least some fields are sown in a timely manner and are not delayed until cultivation of the last field is complete.

Sowing is only one of several recurring operations in both crop and fallow phases (Figure 39). For each, timeliness may be critical. In some cases it is doubtful whether the operation should be undertaken at all. For example, ploughing and grazing at unsuitable soil-moisture contents can be detrimental to soil structure. Grazing is most beneficial if timed so that the livestock eat immature weed seeds, but it is detrimental if delayed until after seed fall and causes uprooting of plants when the topsoil is friable and dry.

Within both Australia and Mediterranean regions there are examples of family farmers and large corporations with an ethic of land care who manage their soils for sustainability, and others who do not. Table 47 summarizes some of the characteristics of these extremes in Australia. Unsustainable management is almost always oriented to short-term market prices without knowledge or interest in the biological mechanisms which underlie soil and crop behaviour, and the patterns which characterize sustainable crop rotations. Table 47, although necessarily general, illustrates the fifth premise of this Bulletin (Chapter 1), namely that tailoring an understanding of generic mechanisms to local situations is a powerful way of creating sustainable cropping systems.

FIGURE 39 - Scheme to draw attention to the need for clear time-lines, and critical timeliness of operations, in cropping in the semi-arid tropics

Figure 38 illustrates a non-sustainable situation in east Australia. The key indication that the system is not sustainable (declining protein concentration in grain) is linked to its root causes within the system. Similar schemas, together with local knowledge of seasonal work demands and social constraints (e.g., Figure 31), can form the basis for debate about any action to be taken (Chapter 5).

Management aspects of dryland cropping systems from an Australian perspective

Not biologically sustainable


· Cropping decisions made annually only on grain and livestock prices

· Cropping broadly based on rotations and ethic of land care

· Focus on few products: 2 crops, 1 -2 crops, 1-2 livestock (e.g., wool, sheep meat)

· Diversity of products

· No rotations; grasses dominate system

· Rotations with balance between grasses (cereals and legumes)

· Full utilization of crop residues for livestock, hay, etc.; bare soil in dry season

· Organic matter and soil biology management: soil surface coverage, residue incorporation, permanent contour ridges/bunds

· Inorganic nutrients added (sometimes) according to product prices

· Inorganic nutrients added regularly to maintain zero-loss budget, i.e., to compensate for nutrients removed in grain and livestock products

· Annual conventional cultivation (>2 passes)

· Minimum tillage

· Fallow not managed; without legume

· Fallow perceived as integral part of crop rotation: contains pasture legume which is managed to maximize seed products and self-regeneration

· Uncontrolled seeding of weeds including annual grasses during fallow

· Conscious weed management to minimize weed seed rain

· No perennial grasses or trees

· Some perennial vegetation: grasses in waterways, exposed hillsides; tree legumes

· Livestock populations high; grazing pressure uncontrolled, causing, e.g., soil structure degradation in wet areas

· Livestock populations conservative, flexible for surviving drought

Rotations and soil productivity

On the balance of experimental evidence, those cropping elements that help to maintain soil productivity in parts of Africa, Asia and Australia have been emphasized. These include: (a) the need to retain organic matter and soil structure; (b) the need for appropriate tillage which, in many situations, involves minimal soil disturbance; (c) the need to replenish the nutrients removed in grain, straw and livestock products; (d) the essential need to maximize ground cover; (e) the desirability of mixing crop species, particularly to include legumes with effective nodulation (grain crops, forage crops, trees and shrubs); and (f) the need for any fallow period to be managed rather than to be left bare to be exploited by livestock, or for weed seeds and pests to accumulate in the soil.

The question arises: how are these conceptual elements best exploited in crop rotations? Table 47 provides a framework, and Figure 39 illustrates the dynamic, and recurring, nature of decisions involving the choice of rotations and techniques for the maintenance of soil productivity.

Figure 21 illustrates the value of effectively-nodulated legumes (in this case, a pasture ley) in increasing soil nitrogen, as long as the nitrogen is not removed as a product. Figures 37 and 28 show associated increases in soil rhizobia and vesicular arbuscular miccorhizae under undisturbed pasture legumes. Tables 13 and 14 show the impact of plant architecture and ground cover on soil loss and mineral leaching. All the above give estimates of the benefits, to soil biology and chemistry, of the crop elements in a rotation. The aspects they illustrate are reflected by some Australian farmers' practices (Table 44), under which the rotations chosen appear to maintain soil productivity (Table 46) though soil nitrogen levels at equilibrium may be only half those under continuous legume pasture (Table 45). However, there are very few long-term studies which quantify the effects of different crops on soil conditions. Indeed, such studies seem to be a priority to encourage changes in cropping systems. For example, there are few data on nitrogen fixation by legumes on farms in Africa and Asia, and no known studies, that compare the effects of groundnuts on soil structure and fertility (which are possibly degrading) with other legumes such as soybeans.

There is little information on the effects of fallow management on soil biology, particularly on weed seeds, pests and diseases. On many farms, weeds spread uncontrolled during the fallow period, while experiments often unrealistically contrast weed-free extremes. Caporali and Onnis (1992), in re-affirming the importance of crop/legume rotations, interestingly also show that weed density increased more during ten years of sunflower-based cropping under sunflower monoculture than under a sunflower-lucerne rotation. They show too that there are more weeds after mechanical- than chemical-control.

There is much information available on the effects of tillage and litter/stubble treatments on soil properties. The data, however, are not always consistent. This is not surprising as the timeliness of an operation is relevant to its effects on the soil and there are rapid changes in soil properties in the days following sowing (Addae et al. 1991; Ghuman and Lal 1992). Moreover, several years of treatment are often needed before there are lasting effects upon soil characteristics that influence crop growth, for example, root length density (Figure 15).

Notwithstanding the variation, and at times the ephemeral effects of management on soil properties, it is clear that there are benefits from those crop rotations that enable crop residue retention. Three examples follow. Gill and Aulakh (1990) show that zero tillage for three years gave the lowest soil bulk density and highest crop yields. Radford et al. (1992) found that zero tillage plus stubble removal gave the least soil water storage but that zero tillage plus stubble retention gave the greatest soil water capacity. Chan et al. (1992) record that ten years, conventional (three-pass) cultivation and stubble burning gave least soil organic matter whilst direct drilling plus stubble retention gave most organic matter. In this study, tillage had a cumulative effect, causing 31% differences in organic matter (2.42% as against 1.68% organic carbon); a loss of 1% organic carbon was calculated to result in a loss of almost 3 cmole of negative charge per kilogram of soil. As a working hypothesis, it seems that stubble and litter retention (and incorporation if necessary to avoid removal by grazing or wind) is the key to sustainable rotations. Bearing in mind soil organisms, however (Chapter 3, section Organic matter), it may be necessary occasionally to till and mix the soil layers (Alegre et al. 1991).

Mineral nutrition is, together with crop choice, stubble/litter treatment, and soil management, a key to sustainable rotations. For example, the removal of a 600 kg grain crop requires 1.8 kg phosphorus to be replaced using manure or mineral fertilizer. Nine kilograms of nitrogen are also needed (most realistically, from the breakdown of a previous legume crop or pasture, or by application of manure). A long fallow, in which deep-rooted perennials draw nutrients from depth, may help to balance the loss of nutrients. However, nitrate may move down the profile during the fallow, commonly giving a net annual decline in soil nitrogen in the absence of legumes (Martin and Cox 1956). Standley et al. (1990), however, found that there was an increase (of 30 kg N/ha) during a short (7-8 months) fallow which balanced the loss during the crop phase. If weeds or pasture grow during the fallow and are grazed, then nutrient losses of about 20% are common. These losses are by removal of animal products and volatilization of urine. Grazing can also redistribute nutrients as dung from one field to another.

Previous Page Top of Page Next Page