Contents - Previous - Next

Land husbandry

[Plates 24 and 25]

Since 1975-80 numerous research experts, social scientists, economists and agronomists have voiced criticisms of the frequent failure of water management schemes implemented too hastily and without reference to local people's views (Lovejoy and Napier 1986).

In the United States, despite 50 years of remarkable work by the Soil Conservation Service and the annual expenditure of millions of dollars, 25% of arable land is still losing more than 12 t/ha/yr, the tolerance limit for deep soils. Although there have been no sandstorms as catastrophic as those of the 1930s, pollution and siltation of dams remain major problems. With a view to improving the effectiveness of the purely voluntary efforts of farmers hoping to protect the productivity of their land, federal laws (on cropping grasslands, wetlands and fragile land) now force farmers to respect rules for conservation-minded land use, failing which they lose the right to any of the financial incentives intended to support the American farm sector.

In the Maghreb and West Africa, farmers often prefer to abandon State-improved land rather than maintain protection works of which they do not understand the purpose, and whose ownership is unclear (Heusch 1986).

Many reasons have been advanced for these partial failures (Marchal 1979, Lefay 1986).

• choice of techniques ill-suited to soil, climate or slope;

• bad planning, incorrect implementation or lack of follow-up and maintenance;

• no training or preparation of beneficiaries, who reject the project because loss of surface area is not balanced by increased yields;

• poor organization of production units (fragmented and isolated plots).

A STRATEGY BASED ON AGRICULTURAL DEVELOPMENT

Given these failures, a new strategy had to be developed taking better account of the needs of those in direct charge of the land, both farmers and herders, by offering methods that would improve soil infiltration capacity, fertilization, and yields - or better, farmers' profits (Roose 1987a). This method was named "water, biomass and soil fertility management" by Roose in 1987, then "land husbandry" by Shaxson, Hudson, Sanders, Roose and Moldenhauer in 1989. It starts from the way farmers experience soil degradation problems, and comprises three phases:

1. Preparatory discussions among farmers, research scientists and technical support staff. This phase covers two surveys to identify problems and assess their importance and causes and the factors that can be brought into play to reduce runoff and erosion. It also includes field visits to the village community to foster their sense of communal responsibility, learn how degradation problems touch them, and discover the strategies they already have for improving water use, maintaining soil fertility, renewing plant cover and controlling wandering livestock. Also looked at are social and economic constraints, limiting factors, land ownership, credit, training and availability of labour.

2. On-farm field trials are set up to measure and compare the risks of runoff or erosion and the higher yields resulting from various types of development or improved cropping techniques. This procedure will establish a technical layout and determine the feasibility, profitability and effectiveness of the erosion control methods recommended: evaluation must be carried out jointly by farmers and technical experts.

3. A comprehensive land-use plan should then be established after one to five years of dialogue, with a view to rationally intensifying farming on productive land, characterizing the terrain, controlling gullies and stabilizing soil, preferably through the use of simple biological methods that farmers can handle themselves. Nothing can be done without the prior agreement of the farmers, who must be encouraged to manage their land as a unified whole.

Answers to the different problems will vary according to local social and economic conditions (large modern landholders or small subsistence farmers), even when the physical environment is the same. This is the main difference from previous approaches.

FROM SOIL CONSERVATION TO WATER, BIOMASS AND SOIL FERTILITY MANAGEMENT

It has become very clear in recent times that soil conservation schemes confined to reducing the amount of soil carried away by erosion cannot answer the needs of farmers in tropical regions. Indeed, experts have been saying for a long time that soil has to be conserved so as to maintain the productivity of the land; thus, the title of the fifth ISCO conference (Bangkok 1988) was "Land Conservation for Future Generations." This is a duty to society and a long-term investment!

Farmers (not always of their own volition) have undertaken to devote considerable efforts to schemes to control erosion on their land, but have been disappointed to see that their land still deteriorated and crop yields still fell. The erosion control structures imposed (drainage ditches, diversion channels, bunds) have often reduced the arable area (by 3% to 20%) without any equivalent improvement in the productivity of "protected" plots. If farmers are to be motivated, it is not enough to keep the soil in place: water must be managed and soil fertility restored in order to see a significant increase in yields from these tropical soils, the majority of which are already very poor (especially tropical ferralitic and ferruginous soils that are sandy on the surface).

Land husbandry must show immediate returns: the challenge is to double production in twenty years so as to keep up with population growth. SWC is essential for stopping loss of water and nutrients through erosion and for preserving the soil's storage capacity. But SWC is not enough, for the farmers need to receive immediate rewards for their labour in protecting their land. This is possible - at least with sufficiently deep soils - if improvement of both nutrient and surface-water management (drainage in cases of waterlogging, subsoiling of calcareous crusted or sealed horizons, rough tillage or mulching if the surface is crusted) are undertaken together.

In traditional systems it is the long fallow period that allows the recovery of good soil structure, ensures an adequate level of organic matter, and the availability of nutrients for crops. Burning can raise the pH by a couple of degrees and counter aluminium toxicity, particularly in humid zones. With population growth and expanding needs, however, fallow periods have been shortened so much that they can no longer restore soil fertility. The mechanization of farming has expanded the amount of cultivated area rather more than it has increased yields (Pieri 1989). In many regions all the arable land has already been cleared, so now the productivity of land resources has to be intensified.

Initially, farmers understood intensification as meaning a reduction of the fallow period and an expansion of cropping to all arable areas: average yields (600 kg/ha) were maintained by clearing new land.

Then rural organization and training services recommended animal-traction tillage and use of selected disease-resistant seeds from field stations. Only small amounts of mineral fertilization were extended (less than 100 kg/ha of NPKCa). Yields rose from 600 to 1100 kg/ha (cereals, groundnut, cotton), but as the balance of organic matter and nutrients was negative, soils quickly deteriorated, as did yields. Attempts were then made to improve the fallow period and fodder production.

Finally, development corporations suggested intensive cropping systems: cotton and maize in Sudanian areas, and groundnut and millet in drier, sandier areas. These systems combine larger inputs of mineral fertilizer (over 200 kg/ha on cash crops), tillage (turning under and hoeing/ridging), oxen-traction (which implies fodder and manure production on each farm), rotation with no fallow for ten years or fallow under a fodder crop (often legumes), and selected varieties with good response to fertilizers and the regular use of pesticides and herbicides.

Results were encouraging, but varied greatly according to rainfall, soil type, and socioeconomic conditions (Pieri 1989). Crop yields increased two- to fourfold (1500-2500 kg/yr) and up to tenfold in field stations with deep, even-textured soil. However, after five to ten years yield improvements from mineral fertilizers were falling annually by 10%. Cropped soil receiving only mineral fertilizers is deficient in organic matter. In savannah areas the amount of humus in the soil declines by 2% a year on loamy-sandy soil, 4% on very sandy soil (Clay + Loam < 10%) and up to 7% where there is considerable erosion and/or drainage.

Ploughing in crop residues - or what is left at the start of the rainy season (less than 10%) - is not enough, especially when such residues can be put to better use by cattle or craftsmen. Ploughing in coarse straw (Carbon/Nitrogen < 40 will produce nitrogen lock-up) or green manure does stimulate microbial activity for some months, accelerating mineralization of reserves of stable humus. The only way of maintaining soil productivity seems to be applications of manure or well-decomposed compost (C/N < 15) (3-10 t/ha/yr), supplemented by essential minerals to correct soil deficiencies. This maintains the level of organic matter in the soil (and thus its structure and its water- and nutrient-storage capacity), prevents acidification, and fosters deep rooting and biological activity (micro- and mesofauna) (Chopart 1980).

Erosion, poor tillage (carried out too late or crushing the soil too fine), plus applications of nitrogen fertilizer hasten depletion of the soil's stores of organic matter. By contrast, rotation of different types of crop, use of full mineral fertilization, tillage leaving a rough surface, minimum tillage along seed lines plus a litter of residue spread on the surface, and fallow crops producing a large amount of root biomass (Andropogon, Pennisetum or cultivated legumes) all delay depletion of organic matter.

On tropical ferralitic and ferruginous kaolonitic clay soils, organic matter plays an important part in protecting soil structure and its ability to store water and nutrients. Kaolinitic clay which has a cation exchange capacity of only 14 milliequivalents per 100 grams will give only 1-2 meq in horizons colonized by roots (Clay + Loam £ 20%), whereas humus will give up to 250 meq per 100 grams.

Although crop yields may not be directly linked to levels of organic matter in the soil below certain threshholds (Organic matter/[Clay + Loam] < 0.07), soil structure breaks down, runoff and erosion accelerate, rooting is less effective as the soil becomes compacted, and nutrients are less easily accessible. Degraded soil gives a poorer return for fertilizer as there is less water available in compacted soil (Pieri and Moreau 1987).

It was believed at one time that massive mineral applications, including the dose needed to correct soil deficiencies (applied every 4 to 10 years) plus the replacement dose (taken up by crops), would solve all these problems: increasing both yields and the available biomass to improve the level of organic matter in the soil. What was forgotten was the risk of acidification from nitrogen and other acid fertilizers (sulphates and chlorides), as well as losses through erosion and drainage and, above all, the rapid mineralization of organic matter, which is further accelerated in tilled soil. Even if such a huge input of fertilizer is technically feasible, it is not always economically viable. For example, it was seen that on an intensively irrigated banana plantation in southern Côte d'Ivoire on highly desaturated ferralitic soils, erosion and (especially) leaching led to losses of 9% of the phosphorus, 100% of the lime and magnesium (1 tonne of dolomite) and 60% of the nitrogen and potassium (at least 300 units), although these were spread out over ten applications a year around the foot of each plant (Roose and Godefroy 1977). A tendency to acidification of sandy soils has also been noted in the case of nitrogen, sulphate and chloride abuse.

A major new development was the insistence of World Bank economists on cost-pricing fertilizers in order to reduce wastage. Fertilizer subsidies were intended to offset the huge costs of transport, and their withdrawal meant that small farmers scattered in thousands of villages were denied access to this modern technology and thus the possibility of increasing returns on their labour. They therefore had to turn back to regional resources (crushed natural limestone and phosphates) and more or less converted local biomass. However, it very soon became clear that a fatal imbalance was fast being reached between inputs of nutrients and losses through erosion, drainage and uptake by crops.

Forest fallows are able to draw nutrients from deep down (the product of weathering of minerals and recovery of solutions that have drained down below crop roots) and recycle them at the surface (8-15 t/ha/yr added to the litter). It takes 8 to 20 years to reconstitute soils in subequatorial forest zones, 15 to 30 years in Sudanian forest zones, and 30 to over 50 years in Sahelian zones. By contrast, degradation of the surface horizon is much faster; nutrient reserves are depleted after 2 to 6 years of intensive cropping, and after 15 to 20 years the macroporosity has decreased and a sand horizon is all that is left after selective erosion. Furthermore, if soil is deficient because the parent material is poor in a given element (for instance phosphorus), the vegetation will also be so, equally the litter and humus, so that a supplementary mineral input becomes essential.

In savannah areas, where the biomass is mostly composed of grass which burns each year, fertility is concentrated by animals who harvest the biomass dispersed over rangeland (often very poor land unsuitable for cropping) and return it in night paddocks in the form of dung. This is not real manure (which ferments at 80°C, killing seeds), but sun-dried dung trampled into powder by livestock kept in paddocks with no straw. This mixture of muddy earth and poorly decomposed organic matter contains many seeds of weeds and fodder shrubs ready to germinate. The rather poor organic matter has unfortunately lost much nitrogen through gasification in the sun, since there is no straw to trap the nitrogen and form humus. It would be easy to improve and increase manure production through some kind of system of stabling animals on straw litter (which would collect liquid waste and reduce loss from drainage), shaded by a rudimentary roof, until such time as a tree canopy can form. The role of trees in the management of a dung-compost heap is that of providing a more temperate atmosphere, protecting the fermenting biomass from direct sunlight, reducing evaporation (and hence the need for water), recovering some of the nutrients lost in drainage, and producing a litter richer in nutrients than grass straw can provide [Plates 30 and 31].

The contributions of manure from livestock, however, are limited. In an extensive system, one cow gives 0.6 t/ha/yr of dung, whereas it takes 3 t/ha/yr of manure to keep soil carbon above critical level. Five cows are therefore needed to produce 3 tonnes and maintain an hectare of crops, and since it takes four hectares of extensive rangeland to feed one cow, 20 hectares of extensive rangeland are needed for the upkeep of one hectare of cultivated land through the use of organic manure. This performance can be improved by intensive animal husbandry: one cow can be kept on the crop residues from one hectare; one cow can moreover produce 1.5 tonnes of manure if she is kept on litter over-night and during the hot hours of the day; and, lastly, one hectare of an intensive forage crop can in fact keep the two cows needed to produce three tonnes of manure.

From the soil standpoint, however, only 30-40% of nutrients in the biomass digested by animals are excreted back to the soil. All conversions have an efficiency rate. The best animals fix up to 70% of the nutrients in their diet to form bone (Ca + P), protein (N-P-S, magnesium and various trace elements), most of which are lost to the soil. It would be a good idea if powdered bone, blood, horns, hooves and other animal products not used elsewhere were returned to the soil. In the intensive systems of Rwanda, where population density is over 250 per km², organic manure can maintain only a third of the farm (often less than one hectare to feed 4-10 people), with the rest of the land being used for very undemanding crops such as cassava and sweet potato. Small animal husbandry with the stock feeding on communal lands and along roads is often the only way small farmers can survive and build up a modest nest-egg to cope with life's emergencies (illness, accidents) and social relations (marriages, funerals, etc.) (Roose et al. 1992).

Composting is an even longer way of transforming the biomass (6-18 months) with returns no greater than from manure. However, it is a valid technique for those without livestock (the poorest farmers) or with large amounts of industrial waste available (coffee husks, brewery draff, town sewage, etc.). The main problem is the amount of work required to produce good compost. Compost pits dug in the fields have been tried, in order to avoid double transportation of straw and crop residues, but most of them stay empty, and any compost is poor. The only effective ones are the compost-manure-rubbish pits close to the house, which receive all available residues plus ashes and domestic waste. To help the mixture ferment with minimum waste, small pits (4 m × 2 m) are recommended, planted with trees, which will give shade, a cool, damp environment and a biomass rich in minerals, and whose roots will recover solutions leached from the compost by drainage water. Since the maximum is 5 t/ha/yr per family (i.e. 0.2-0.5 ha manured per farm unit), additional solutions must be sought in order to fertilize the whole farm. However, it is a good basis for starting to grow vegetable crops.

Turning in residues and weeds: people often tend to overlook the mass of crop residues, roots and particularly weeds which farmers turn in when tilling and hoeing. However, it is a short process (1-3 months), allowing speedy recycling of the nutrients in the biomass. There are also various traditional methods of gathering weeds into piles to dry, then covering them with a mound of earth in which sweet potatoes are then planted. After the sweet potato harvest, the organic-matter-rich earth is then spread. Repeatedly turning in fresh organic matter throughout the year in this way does allow maintenance of a certain level of organic carbon in the soil, but its effect on soil fertility and resistance to erosion is limited. Moreover, farmers are increasingly using this biomass for their livestock, since fallows are disappearing. Also, a 1% increase in the amount of organic matter in the soil brings a mere 5% reduction in soil erodibility (Wischmeier, Johnson and Cross 1971). Sizable applications of broken-down organic matter are needed for a 1% rise in the amount of carbon in 10 cm of soil (1% of 1500 tonnes of soil). Simply ploughing in 15 tonnes of barely broken-down straw only leads to lock-up of the nitrogen fixed by the microbial mass, and lower yields.

A thick mulch (7-10 cm, or 20-25 t/ha) is a very effective way of reducing evaporation and weed growth, maintaining soil moisture during the dry season, and halting erosion. It is also a short-cycle means of restoring the whole of the biomass and its constituent nutrients (K, Ca, Mg, C, initially by leaching, and N and P by mineralization and humification through the action of meso- and microfauna). The litter on the top of the soil disappears 30% more slowly than when the organic matter is ploughed under, and there is less risk of nitrogen deficiency. Under forest, where the soil is often best, litter is never ploughed in but is left to the action of earthworms, termites and other mesofauna: soils which are not degraded are fully capable of absorbing organic matter deposited on the surface. Mulching has been successfully tested on coffee and banana plantations which proved to be the least eroded and degraded plots on hillsides long cultivated. Unfortunately there is never enough plant residue to cover all the cultivated land. Nevertheless, a light mulch (26 t/ha), spread at the start of the rainy season, once the soil has been tilled and sown, cushions the force of raindrops and runoff and maintains good infiltration longer, as well as encouraging good mesofauna performance. Even if only 50% of the soil is mulched, erosion risk can be reduced by 80%. On crusted soil mulching is effective in reducing erosion, but less so runoff. In any case, mulching improves soil structure by providing the surface soil with nutrients and fresh organic matter.

None of these recycling techniques is perfect, but must be used in combination to draw the best from each [Plates 28 and 29].

Agroforestry [Plates 18 and 19], and especially planting hedges every 5 to 10 metres, gives a mass of fodder and mulch which can be returned to the soil during tilling. Usually, deep-rooting leguminous shrubs capable of producing 4 to 8 tonnes of dry organic matter/ha/yr are used (Balasubramanian and Sekayange 1992, König 1992, Ndayizigiyé 1992). Despite all this organic matter, a supplementary mineral application will still be necessary, both to condition the soil, raising the pH above 5 in order to suppress aluminium toxicity and allow legumes to grow, and also to offset soil deficiencies by directly giving plants the necessary nutrients where they need them and when they can store them.

It is often too expensive to correct soil mineral deficiencies with direct applications, and it also makes little sense unless the soil storage system is improved (organic matter and clay content). In many cases there is also a retrogradation of phosphorus in the presence of iron and lime, or of potassium if the environment contains swelling clays (montmorillonites). The soil systems in high-rainfall tropical zones also suffer from many kinds of loss, first through erosion, then through drainage, and finally through gasification. The risks of leaching can be reduced by observing the following rules:

• split applications of fertilizer (1/3 at sowing, 1/3 at shooting or heading, and 1/3 at ear-emergence);

• liming after the period of heavy rains;

• gauging dosage to soil and plant storage capacities;

• choosing nutrients in a form that plants can assimilate;

• enhancing soil storage capacity by applying organic matter or clay with a high fixation capacity (swelling smectite);

• spreading fertilizer over the whole area penetrated by roots;

• encouraging a certain amount of weed cover, which is then cut down at the right time to form litter (such vegetation will temporarily hold nutrients that could easily be leached);

• balancing applications according to plant needs and availability in the soil.

While trials of biomass production in plant containers (control + NPK + NP + PK + NK) reveal the relative importance of soil deficiencies and soil potential, the level of uptake indicates the minimum nutrient applications needed to attain a production goal (Chaminade 1965). Monitoring plants (leaf analysis) and soil (soil analysis) is expensive, but it does show what the plants are consuming and what is lacking in the soil. Sampling is essential for meaningful results (Boyer 1970, Pieri 1989).

SOIL RESTORATION AND LAND REHABILITATION

Population pressure, drought and overgrazing in many semi-arid tropical regions have degraded plant cover and impoverished soil in organic matter, nutrients and fine particles (selective erosion). Eventually, the soil may be acidified, then scoured, destructured, crusted by rain-splash and compacted in depth following tillage and the mineralization of organic matter. There are therefore significant sterile areas (5-20% of the land) which are nonproductive but give rise to considerable runoff that then causes gullying problems on good arable land further downstream.

Until now such degraded land has been turned over to foresters for restoration. Their response is to declare the land off limits, (i.e. protect it against bush fires, herders responsible for overgrazing and farmers who have cleared these fragile soils), planting it with some pioneer tree species and managing surface water through the use of bunds or diversion ditches. People and animals are asked to live elsewhere or be liable to fines ... which leads to the classic tension between foresters, herders and farmers. Closing land off is often a very effective method if plant degradation has not progressed too far, but it is difficult to enforce under strong population pressure, and its effectiveness decreases in low-rainfall areas.

From a land husbandry perspective, it seems more effective - and cost-effective too - for farmers to concentrate on the sound management of good, productive land before it becomes degraded, since there is a faster and larger increase in yields on deep soil than on exhausted stony soil. Prevention is better than cure! However, there are cases where restoration of degraded land is a priority for the population:

• rehabilitation of stony land (what Haitians call "finished land") on hilltops, for runoff from them will gully good cultivated land lower down;

• restoration of land that is degraded but still has an agricultural future, with a possibility of water- and nutrient-storage in a profile thick enough to ensure a cropping cycle despite climatic vagaries (more than 30 cm of clayey soil, more than 60 cm of sandy soil);

• when pressure on land becomes acute, not only must soil be restored at all costs, but productive soil must be created wherever rocks can collect rainwater (as with the Dogon people of Mali).

LAND REHABILITATION TO PERMIT EXTENSIVE USE OF NATURAL RESOURCES

The procedure here is gentle intervention in order to encourage plant cover to regenerate, without major changes in the nature of the soil:

• improving water storage in sandy semi-arid zones by the use of an "imprinter" roller which pits the ground to trap runoff water, sand and wind-borne seeds (Dixon 1983); this method, called pitting, works well in sandy zones, but is almost useless on degraded vertisols;

• rough tillage or cries-cross ripping (subsoiling) with direct sowing of perennial grasses or fodder shrubs; this method has a temporary effect in Niger, but little effect in southern Mali on denuded bunds (700 mm rainfall) (Poel and Kaya 1989), north-western Burkina Faso (Roose, Dugue and Rodriguez 1992) on a gravel pediment (500 mm rainfall) and northern Cameroon on degraded vertisols;

• closing land off to prevent fire and grazing: very effective on soil still partially covered, at Gonsé in Burkina Faso (rainfall ~ 700 mm) (Roose and Piot 1984, Roose 1992a) and at Kaniko in Mali (Poel and Kaya 1989), but of no use if the soil is completely bare and crusted (Poel and Kaya 1989);

• spreading twigs and bark (Chase and Boudouresque 1989, in Niger), cotton stalks (Poel and Kaya 1989, in Mali - Table 1) or sorghum stalks (Roose and Rodriguez 1990, in Burkina Faso); this is the most effective way of trapping wind-borne seeds and sand and attracting mesofauna which will drill through the slaking crust, thereby restoring soil infiltration capacity;

• stone bunds to catch rain - or lines of straw, grass, pebbles and branches - act in the same way as above, but only foster vegetation regrowth three metres on either side of the permeable obstacle which slows down runoff and encourages sedimentation;

• Euphorbia balsamifera hedges have difficulty surviving on the deeply degraded and acid soils of Kaniko (Poel and Kaya in southern Mali), whereas Opuntia hedges in Algeria and zizyphus, acacia and various thornbushes in Burkina Faso have fixed the soil well... when protected from livestock (Roose et al. 1992);

• earth bunds, whether straight or semi-circular, are short-lived and allow only grass to grow at the points where runoff collects, in Burkina Faso (Roose, Dugue and Rodriguez 1992) and Mali (Poel and Kaya 1989);

• on degraded vertisols in Cameroon, only when earth bunds were used to isolate the small pits was there a slight improvement in infiltration - and in cereal production;

• if the area is used by livestock during the dry season, there will be less biomass and the surviving species will differ (Chase and Boudouresque 1989; Poel and Kaya 1989).

TABLE 1
Recovery of degraded soil protected from grazing (cf. Hijkoop, Poel and Kaya, 1991)

RESTORING THE PRODUCTIVITY OF FARMLAND

With deep, healthy soil which has been scoured by erosion or degraded by crops that have upset the balance of organic matter and nutrients, the use of a single approach, whether biological, physical or chemical, will rarely be successful. On the other hand, soil productivity can be restored very fast (1 to 4 years) in tropical semi-arid and especially semi-humid and humid zones as long as the following six rules are observed [see box] (Roose et al. 1992)

In a Sudano-Guinean climate, fallow periods of luxuriant tall grasses may improve the physical properties of rich soil which has not become too degraded through cropping (Morel and Quantin 1972), whereas in Sudano-Sahelian zones a short fallow period of natural grass-growth (2-6 years after 2-3 years under crops) cannot be expected to maintain much less, restore - the land's agricultural productivity (Pieri 1989). On tropical ferruginous sandy soil in the Tcholliré region of northern Cameroon, Roose (1992a) observed only very slight improvements in carbon content (0.3-0.6%), nitrogen content (0.01-0.06%) and pH (5.3-6) of grasslands after 30 years of burning and extensive grazing each year. The best results (C = 1%) were produced by earthworm casts, and termite mounds in old paddocks where livestock were kept overnight. Infiltration capacity, on the other hand, is much better on old fallow land, because of roots and the tunnelling of earthworms and termites.

An excellent example of the rapid restoration of productivity of degraded land is the traditional Mossi method known as zaï (Figure 5). During the dry season, farmers dig out pits 15 cm deep and 40 cm in diameter on degraded plots every 80 cm, tossing the earth downhill. The dry desert Harmattan wind blows various organic residues into them. These are quickly attacked by termites (Trinervitermes) which dig tunnels through the crusted surface, allowing the first rains to soak down deep, out of danger of direct evaporation.

Two weeks before the onset of the rains (15 May to 15 June), farmers spread one or two handfuls of dry dung (1-2.5 t/ha) in the bottom of the pits and cover it with earth to prevent runoff from carrying away dry organic matter on its surface.

Some sow a dozen seeds of millet if the soil is light, or sorghum if the soil is loamy-clayey (about 8 kg/ha of seed in seed holes before or after the first rains).

The first rains run copiously over the surface crust (2/3 of the land): the basins capture this runoff (enough to soak a pocket of soil up to a metre in depth). The seeds germinate together, break up the slaking crust and send roots down deep to where they find stores of both water and nutrients recycled by the termites. The concentration of water and nutrients around the seed holes can make yields as high as 800 kg/ha the first year and steadily increase for 30 years as the whole field improves.

At harvest time, stalks are cut at a height of one metre and left to reduce wind-speed and trap wind-borne organic matter.

In the second year, the farmer either finds time to dig new basins between the first ones and dress them with manure, or pulls up the stubble and resows in the old basin. Stubble-clumps laid between basins are in turn attacked by termites. After five years the whole area has been turned over and manured so that the soil is now pliant enough to be tilled normally. Some farmers say that land restored by the zaï technique can be cropped for over 30 years.

December to April

• Digging pits 40 cm in diameter, 15 cm deep, at 80 cm intervals, earth piled up in a crescent on the downhill side.

• The harmattan brings sand and organic matter.

April to June

• Application of 2 handfuls of powdered corral dung after the first rain (= 3 t/ha).

• Termites dig tunnels lined with excrement.

• Seeds planted in holes at the second rain.

• Water seeps down and is stored deep down, away from direct evaporation.

June-July

• Onset of the rainy season.

• Early emergence.

• Deep rooting.

• Weeding restricted to plant holes.

• Germination of forest seeds.

• Concentration of water and nutrients.

November

• Harvesting of panicles and forage.

• Stalks cut at 1 m, hiding forest shoots from livestock and inhibiting drying wind and erosion.

FIGURE 5 Zaï: traditional method of soil restoration (cf. Roose and Rodriguez 1990)

• Zaï means: hurrying to dig packed, crusted soil in the dry season.

• Abandoned land can be recovered and made to produce about 800 kg/ha of grain from the 1st year, maintaining soil fertility for more than 20 years.

• It concentrates water and fertility under the planting hole and allows combination with suitable forage trees (agroforestry).

• Limitations:

the starting date of work is decided by the village chief... after celebrations, and sometimes too late;

it takes 300 hours' very hard work, i.e. 3 months/person, to restore 1 ha; it needs 2 to 3 tonnes of organic matter, and carts to carry the corral dung and compost; to succeed, the field to be restored must be surrounded with a stone line to control runoff.

• Improvements:

cries-cross subsoiling with 1 tine up to 12 to 18 cm, after harvesting, every 80 cm (11 hours with well-fed oxen), then digging zaï for 150 hours;

supplementing organic manure with N and P which are lacking in dung exposed to the sun; introducing other forest species grown in nurseries (3 months gained).

A variation known as "forest zaï" is especially interesting. Dried, unfermented goats' dung contains a lot of seeds that have passed through the animals' digestive systems and are ready to germinate. Some astute farmers noticed that fodder shrubs, mostly pod legumes, were growing in the basins. During weeding they now leave two young forest plants every 3 metres, which then benefit from the water and manure meant for the cereals. At harvesting, sorghum stalks are left about 1 metre high, so that they protect the soil from wind erosion and keep the young forest shoots out of sight of the goats (see Figure 5). Cereals are sown and harvested every year, and every five years the forest transplants are cut for poles and firewood. In this way, without the use of wire fencing, an agroforestry intercropping system is set up that restocks Acacia albida and other legumes that can maintain cereal production while providing fodder, litter and wood. The forest zaï method can also be used to plant live fences.

There are three obstacles to the expansion of the zaï method: the work is very hard during the dry season (about 300 hours at 4 hours a day), runoff has to be controlled by a line of stones around the plot, and there are limits to supplies of manure. If the land is prepared in December, the oxen are still well-nourished on crop residues, morning temperatures are still cool, and the soil not too hard; and cries-cross subsoiling can cut tilling time by half. Lacking manure, termites can be attracted by twigs and other organic residues. Not all farmers are aware of the positive role of termites, but in fact fear them, preferring to recycle biomass by producing manure or compost rather than by mulching. However, it should be noted that it is even easier for termites to attack manure under the ground, although it is less visible. This method could be further improved with supplementary applications of nitrogen and phosphates, both of which are in short supply in the soil and in dung (due to gasification from exposure to sunlight) (Roose, Kaboré and Guénat, 1995). So here there is a traditional technique of soil restoration and reforestation oriented toward agroforestry and well-suited to slopes in Sudano-Sahelian zones that have become severely degraded after periods of drought (Roose and Rodriguez 1990).

MANAGEMENT OF LAND STRIPPED DOWN TO ROCK

• If rocks are hard and weather very slowly, and the storage capacity of the soil is very poor, the best practice in semi-arid zones (with at least one dry season) is to use these plots as a catchment area, collecting runoff in a gutter or a concrete track and leading it to a roughcast, waterproofed tank dug in the ground. After removing the sand, this water can be used for stock-watering and household purposes, as well as providing supplementary irrigation for small intensively cropped plots of legumes and other highly profitable crops out of the rainy season. Examples of this method of localized intensification can be found in Haiti (Smolikovski 1989) and Burkina Faso (Roose and Rodriguez 1990) (see Figures 37 and 38).

• If the rocks are soft (soft sandstone, schist, argillite, marl, etc.) or weather quickly (basalt and other volcanic or rough-grained rock), it is hard to cover the whole surface as the soil is too thin, too steeply sloping and scoured during the heaviest rains. However, it is possible to use the flower-pot technique of concentrating water, available nutrients and care on a few plants of primary local interest. Pits of at least 1 m³ are dug every 5 to 10 metres. The small amount of mineral earth available is then mixed in the bottom of these pits with two handfuls of complete fertilizer and a bucketful of well-decomposed manure/compost/peat. Banana or aby other suitable fruit-tree is planted in each pit, with a series of (preferably leguminous) creepers around it, which will steadily fill up the space between the patches under intensive cultivation. The only other task now is to divert surface flow to the pits, drain off any excess, and place ash and any vegetable residues that can turn into compost around the trees. A classic example of this technique can be seen in the Canary Islands, where grapevines are planted in pits dug in a lava field.

• Van der Poel and Kaya's system (1989). This method tries to combine regeneration of natural vegetation (grassing) with plantations of crops more profitable for farmers. Lines of stones or strips of cotton stalks are set along contour lines every 7 metres and a line of Anacardium occidentale planted between them - by sod seeding - after the second rainy season. This is a thrifty species well-suited to tropical ferruginous soils, and makes an excellent firebreak since its dense foliage smothers any other plants. The recovery rate and growth of the fruit trees can be improved by cutting the grass encroaching on the buffer strips and spreading it around the trees as a mulch. Clearly the plots have to be off limits to livestock for the system to work, and this can be reinforced by explaining the land-use plan to the village shepherds (plus coloured tags on the trunks of surrounding trees). The whole arrangement reduces runoff and the risk of erosion further down, and at the same time steadily restores land productivity (fodder, fruit and wood). This system is rather like "striped shrubland" in which runoff from a bare crusted area will irrigate an area of shrubland (grass + shrubs) which benefits from this additional water. There could be a whole series of variations adapted to each semiarid zone according to soil type, vegetation and the needs of those managing the soil.

• Agricultural rehabilitation of soils on hardened volcanic ash in Mexico. Quantin (1992) and a research team financed by the EC studied the rehabilitation of tepetate or hardened volcanic ash whose topsoil has been severely eroded. A variety of preliminary tilling techniques were used in order to restore the agricultural production capacity of this sterile material:

• cries-cross crawler-drawn subsoiling with teeth penetrating to 50 cm every 60 cm;
• levelling slope terraces isolated by banks and ditches;
• successive tilling and harrowing to reduce blocks of hardened ash to 3-5 cm in diameter.

As extensive farming is the practice (rangeland), farmers have very little manure available barely enough for 0.5-1 hectare. The cost of tilling the soil comes to 8000 FF/ha. If maize is planted the first year, yields are very poor, even with a dressing of NPK, either on its own or with a little dried paddock dung.

However, wheat can give 1500 kg/ha from the first year if there is a dressing of NPK, alone or with dung. From the third or fifth year, biological problems disappear, and yields normally amount to 6000 kg/ha/yr if weather conditions are favourable. If the farmer repays only the basic costs, this operation of rehabilitating degraded land will become profitable after eight years. The new manured soil has better infiltration and is more stable and less vulnerable to erosion than the original cultivated soil.

Contents - Previous - Next