Contents - Previous - Next

Erosion control practices

This refers to cropping techniques used exclusively with a view to reducing runoff and erosion damage (Tables 31 and 33).


This simply entails making sure that cropping techniques follow contour lines. Soil roughness (clods and small hollows) must be laid perpendicular to the slope, so that the eventual runoff sheet is slowed as much as possible. The method is effective only on gentle slopes: the P (or erosion control practices) factor will be 0.5 on slopes between 1 and 8%, 0.6 on 8 to 12% slopes, 0.8 on 17 to 20% slopes, and close to 1 on slopes steeper than 25% (Wischmeier and Smith 1978). In other words, the steeper the slope, the less soil roughness can hold back water. A variant consists of alternating crops in contour strips, thus combining the above effect with that of rotating crops with varying degrees of sensitivity to erosion.

Cropping techniques and erosion control structures as a function of surface water management (cf. Roose, Ndayizigiyé and Sekayange 1992)

Water management methods


Cropping techniques

RUNOFF FARMING (water harvesting)
Arid to semi-arid zone

Water harvesting, cistern
Drain, bunds on wadis
Discontinuous terraces

Tillage, basins
Localized microcatchments

Semi-arid (R < 400 mm) or humid zone on highly permeable soil

Blind ditches
Radical terraces
Bench terraces

Tillage + tied ridges

Semi-humid climate, very high-rainfall months, soil fairly impervious

Diversion ditches
Algerian terraces
Radical draining terraces

Ridges oblique or parallel to the slope

All climates, semi-pervious soils, slopes not too steep

Stone lines or walls
Grass banks or lines

Cloddy ploughing
Alternating crops/pasture


It has been seen that tillage followed by ridging can increase erosion risks simply by increasing the slope. However, if the ridges are set perpendicular to the greatest slope, the furrows can hold a considerable amount of water containing suspended sandy or loamy solids. Contour ridging is twice as effective as simple contour tillage, reducing erosion to about 30% of that on the flat-tilled control plot for slopes of 1 to 8%. However, its effectiveness decreases as the slope increases, and on very steep slopes exceptionally heavy rain can cause breaks in ridges, thus giving rise to serious gullying or even landslides. This is all the more likely if the surface horizon is sandy and very permeable while the subsurface horizons are much less so. A first solution consists of setting the ridges at a slight slope, so that excess water can flow slowly - slowly enough to carry away very little solid matter - to a planned outlet (experiments by Hudson in Zimbabwe). Another solution is contour tied ridging, in which a series of pans and ties perpendicular to the ridges prevent the water behind the ridges from falling through a breach and creating a gully. Thirty to sixty millimetres of water will be trapped in the pans together with the heavier sediments, while the excess water can flow behind the ridges until it reaches designated outlets. To be effective, the ties must be 1 to 5 metres apart. The method has performed very well, reducing erosion to 10% of normal. Such methods are, however, suitable only for soils that are very permeable to a considerable depth.

On volcanic soil in Cameroon, the Bamiléké have developed an ingenious system of large, zigzag ridges set parallel to the slope and covered throughout the year by companion crops. This reduces the erosive force of runoff (Fotsing 1992b).

It is therefore difficult to advise on the orientation of ridges with a view to reducing erosion. The decision will depend on interactions between slope, cropping system and soil type. Only after local field trials can a decision be taken on the most effective and safest orientation for each cropping system.


On slopes of less than 8%, erosion is thus cut to 30% of the control plot (P = 0.3). However, the effectiveness of buffer strips varies according to their width, the crop mixture, and the amount of concentrated runoff. While such strips are strikingly effective in the case of light to medium rainstorms, they can quickly become waterlogged under exceptional rainfall. They act as filters, slowing down the runoff flow, causing a fall in its competence and hence the sedimentation of coarse sand and organic matter - and allowing its infiltration rate to rise. These filters are very effective when there is a mixture of pulses and grasses, and when the soil surface has a large number of stalks or roots per square metre (Roose and Bertrand [1971] in Côte d'Ivoire, and Delwaulle [1973] in Niger). In principle, ground-creeping plants with rhizomes and many scattered stalks are more effective than large tufts of grass. If the latter is used, a light mulch of cut tufts must be left on the soil-surface to prevent water from flowing between the tufts and digging channels. Live hedges staggered on alternate lines over a strip 50 to 100 cm wide act in a way similar to grassy strips, although they tend to be less effective, at least during the first years. In the semi-arid zones of Burkina Faso and even in southern Niger, when strips of Andropogon gayanus are sown on the edges of plots, or else about 20 metres from one another, a fair proportion of the sand carried off by wind erosion (Renard and Van den Beldt 1991) or water erosion (Roose and Rodriguez 1990) can be trapped. Erosion control strips have been tried out on erosion plots at Adiopodoumé and Bouaké in Côte d'Ivoire, and at Alokoto in Niger (Roose and Bertrand 1971, Delwaulle 1973), and it appears that once 0.5- to 4-metre strips of thick grass are established, they can reduce soil loss to one-tenth and runoff to about one-third compared to the control plot. The more aggressive the climate, the steeper the slope, the less crop cover there is and the more erosion-prone the soil, the wider the strips must be to ensure effectiveness. In any case, it is best to start with strips at least 5 metres wide, for they can always be narrowed later.

Any leafy plants provide good cover on erosion control strips, particularly natural fallow plants, but the presence of pulses with tap-roots and large, deep-rooting perennial grasses improves infiltration. In tropical areas, Andropogon gayanus, Pennisetum purpureum, Paspalum notatum, Tripsacum laxum, a mixture of various Stilosanthes, sugar cane and various forage plants can be used. Setaria sphacelata gives good results for the first two years, but is quickly exhausted on poor, acid soils.

However, plants whose seeds spread too easily into the fields should be avoided (unless the erosion control strips are mown before the plants flower). Spreading by suckers, runners or stolons (Synodon dactylon) is even worse. Plants with tightly packed roots and numerous stalks will slow runoff more effectively than free-standing trees.

Some experts have warmly recommended various vetivers, because they survive well in semi-arid regions where overgrazing is frequent. They produce siliceous, long-lasting mulch, but their forage quality is poor. The problem is that the strip has to be destroyed in order to extract the essence from their roots. Wherever possible, therefore, it is better to use forage plants and grasses that are suited to local conditions.

The buffer strip acts as a sponge, partially absorbing runoff waters, and also as a comb, slowing down runoff so it will deposit soil from the cropped field above. The runoff water infiltrates deeply or is at least slowed down, reducing its competence, so that it deposits the coarsest eroded sediments. This in turn maintains good porosity and leads to the formation of a small terrace at the rate of 5 to 20 cm per year, which as time passes transforms landscapes into a succession of gently sloping fields and banks protected by leafy growth.

This inexpensive method has been tested extensively and successfully on research stations, industrial plantations (rubber and pineapple) and modernized small-scale farms. It has some very decided advantages:

• it is easy and inexpensive for small farmers to launch;

• large areas can be treated fast without the costly, cumbersome intervention of surveying teams: after a one-day course in use of the water-tube level, most farmers are able to mark out the contours on their land;

• forage produced on the strips can be used to feed stock or to mulch the fields;

• this living network along contour lines can act as a reference for the orientation of cropping procedures;

• the land used to make buffer strips is not immobilized since they are also productive. Farmers who have no cows can be dissuaded from setting fire to the strips to destroy insects and other pests by planting trees - either fruit species, or trees that can produce kindling and posts - in the centre of the strips or on their lower side. The main problem with this method is to clearly and definitively separate the grass strips and the surrounding fields and fallow land. Particularly in arid zones, where it is hard to establish grass because of overgrazing, if rock debris is available this erosion control measure can be reinforced by arranging unbroken lines of stones inside the strips (Delwaulle 1973, Roose and Bertrand 1971, Roose and Rodriguez 1990). This combines contour cropping and buffer stripcropping, cutting the length of the slope and gradually reducing the gradient through the natural formation of grass banks. Such methods are already widely used in mountainous countries and are now being tested in semi-arid zones in Mali, Burkina Faso and Cameroon. They have been used for centuries in Europe, the Americas and Asia, where banks are protected by grass and bushes that can be as tall as 2 to 4 metres. The buffer strips develop spontaneously into banks, which act as boundaries between plots.

In the Sudano-Sahelian zone of southern Mali, the DRSPR suggested that grass strips 3 metres wide should be planted across cultivated fields at 50-metre intervals (thus covering 6% of the land). Six perennial species were compared in 1987/88. Brachiaria ruzizensis quickly covers the land even in the first year, but Stylosanthes hamata grows better in the second year. Andropogon gaianus is popular, although establishment need to be worked on. Macroptilium lathyroides and atropurpureum, Clitoria ternatea and Pennisetum pedicellatum proved disappointing. At present, the DRSPR is advising a mixture of Brachiaria and Stylosanthes. Some farmers toss the Brachiaria hay and mix it with molasses as feed for livestock.

At Yatenga (400 to 700 mm of rainfall), located north-east of this same zone but in Burkina Faso, Rodriguez has developed a method to be used by small farmers for harvesting Andropogon sp. and Pennisetum pedicellatum seeds in December. At the start of the rainy season in June, the seeds are pounded with damp sand to abrade them, then moistened for 12 hours. They are sown on a 50-cm-wide, shallow-tilled strip uphill of lines of stones, or between two plough furrows, every 20 to 25 metres. Farmers like these Andropogon hedges, not only because they help to control sheet runoff, but also because they produce the long straw needed for roofs and artisanal crafts; they also provide excellent forage, and livestock particularly appreciate their green shoots in the dry season: even where tufts of Andropogon grow in a cropped field, they are not hoed (Roose and Rodriguez 1990).


The aggressive rainfall plus the permeability and natural resistance of ferralitic soils to water erosion make the main problem in these high-rainfall tropical zones finding a way to cover the ground during the critical period of hard rains, thus preventing destruction of the structure of the surface horizon, the formation of slaking crusts and the start of runoff. Natural conditions are such that most food crops (particularly, cassava, yam, maize and groundnut) and certain industrial crops (banana, pineapple, etc.) are incapable of covering the soil sufficiently before the critical period. A light mulch, as a temporary supplement to plant cover, may be composed of crop residues or other inputs, or a soil conditioner such as Curasol may be used to create a flexible crust to protect the surface. A dead cover (straw mulch or a layer of pebbles) can be a satisfactory substitute for living cover in conserving water and protecting soil; for example, a plot covered with a few centimetres of straw (4 to 6 t/ha) protects the soil as well as a covered, 30 metre secondary forest, even in years with very heavy rainfall (Table 13). A method widespread among market gardeners, mulching is very effective in helping infiltration of rainwater, reducing runoff and evaporation, and protecting the soil against erosion. It deserves extension in traditionally farmed areas where fields are always surrounded by quantities of available brushwood.

The situation can be different under semi-arid conditions, particularly Sudano-Sahelian zones that are overgrazed during the dry season so that soils are practically bare at the start of the rainy season. In these regions, the problem is finding mulch. Although the mulching method is widely known, it tends to be confined to fertilizing the fields of the poorest farmers who have neither livestock nor manure.

In this case, the farmers go into the bushland, collecting the branches of shrubs (Bauhinia and Piliostigma) and pulses that the livestock tend not to eat, and spread them over their small fields, partly to reduce runoff and partly to encourage the activity of termites, which will open up infiltration passages into the soil and redistribute the fertile elements in the mulch. Collinet and Valentin (1984), using simulated rain, have also shown that mulching can slow down the reduction in infiltration capacity following cropping. However, when soils are fairly impervious, sandy or poor in organic matter, they can rapidly become degraded under mulching, simply through the wetting and drying out of the soil surface. Effectiveness is therefore dependent on soil texture and capacity to resist degradation through simple wetting or through clay dispersion when the cation exchange capacity is rich in sodium. In tropical mountains, notably in Rwanda and Burundi, coffee fields that have been mulched for 40 years have suffered no erosion. This shows how effective mulching can be, both in maintaining soil fertility and infiltration capacity, and also in protecting from erosion. The problem is that of collecting enough biomass throughout the year to keep several centimetres of mulch on the surface of the soil. In the beginning, mulching under coffee trees had two purposes: the soil was kept moist and fresh under the trees at the end of the rainy season by providing a covering of 10 to 15 cm of straw, and the soil surface was protected against erosion during the rainy season by a thin (2 to 5 cm) layer of straw. On small farms of about one hectare on steep slopes, it proved difficult in these mountainous areas to produce enough biomass to cover the whole surface, especially when this biomass is needed mainly to feed livestock in order to produce milk, meat and manure. It seems that if the existing banks are transformed into productive sloping banks, covered partly with leafy plants and partly with a double hedge of shrubby pulses (Leucaena leucocephala, Calliandra calothyrsus, etc.), and if trees are planted every 5 metres along the lower side of the bank, this will allow production of sufficient biomass to cover the soil surface, at least after preparation of the seed bed and after sowing, by cutting the hedges and spreading the prunings on the ground. The twigs can then be collected again some months later and used as fuel for cooking. This method is now being tested in Rwanda (Ndayizigiyé 1992) and Burundi. Crop residues can be another source of mulch (ISAR).

In the case of industrial crops, it may be hard to obtain enough green matter to make mulching economical. However, as many crop residues as possible can be left on the surface of the soil in order to protect it between two crops and even during the following cropping cycle. There are various versions of this stubble mulching technique, and although it is very popular in the United States, it does require special equipment to aerate the soil without turning it or disturbing the mulch.

Lal (1975) simply suggests pushing crop residues into the space between planting lines and restricting preparatory harrowing to the sowing line. On plots prepared in this way at the IITA Centre at Ibadan in Nigeria, he showed that infiltration speed remains maximal under crop residues laid on the surface - thanks to the activity of earthworms - and that runoff and erosion remain slight whatever the slope, whereas soil losses increase exponentially with the slope on neighbouring tilled plots. An ORSTOM and IRFA experiment at Adiopodoumé clearly shows the role of crop residues in pineapple plantations, and that of tillage in water management as a function of slope. During the first planting cycle, with about 2000 mm of rainfall, the average erosion on three slopes (4, 7 and 20%) was 197 t/ha on bare soil. Under pineapple, flat-planted in lines perpendicular to the slope, with the previous crop residues burned and turned in, erosion was under 25 t/ha. With similar treatment but with the residues simply dug in without burning, erosion was lower by half (12 t/ha). Lastly, when the residues were left on the surface, plant cover was total, and erosion negligible (0.4 t/ha, or 1% of that on bare soil, and less than 2% of that under pineapple when residues were burned and dug in). Similarly, runoff averaged 36% on bare soil, 6% under pineapple cover, 2% when residues were dug in, and 0.6% when residues were left on the surface. Moreover, no significant increase in runoff was observed under mulching when the slope increased from 4 to 22%. The main conclusion is that when crop residues are spread on the soil surface, the risk of erosion on the steeper slopes is reduced to the point where strict contour cropping could be abandoned, which would make it much easier to mechanize farming (Valentin and Roose 1982).

Lastly, ploughing organic matter into the soil can improve structural stability and resistance to rainfall impact. According to Wischmeier's nomograph, a 1% increase in organic matter in the surface horizon means that soil losses can be reduced by 5% for loamy soils due to improved structure and by 3% for clayey or sandy soils. However, this means digging considerable amounts of organic matter into the soil, for in humid tropical regions, most organic matter vanishes quickly, with less than 5% remaining in the soil as stabilized humus. On the other hand, if the same amount of matter is spread over the surface, it will act as a mulch and reduce soil loss by 60 to 99%. It therefore seems that managing the biomass on the soil surface not only considerably reduces losses through runoff and erosion, but also recycles nutrients through the gradual uptake by plants throughout the rainy season. Field observations under both humid tropical conditions Adiopodoumé and semi-arid conditions (Saria in Burkina Faso) show that crop residues can cover the soil surface for three to five months - the time needed for the crop to provide more than 80% cover - which is generally enough to reduce erosion to a tolerable level.


Mulching methods and their variants generally have technical or economic drawbacks (risks of plant disease, insect-damage and weed infestation in the first case, and 250 to 300 days' work to collect the mulch in the second) which tend to be unacceptable in large-scale industrial farming. Hence the idea of testing an artificial mulch easy to apply with spraying equipment already found on a good many mechanized farms. An acetate of polyvinyl (sold under the name Curasol by the Hoechst company) was tested. At Adiopodoumé this product was sprayed on immediately after tillage, levelling and planting, in a single dose of 60 grams of Curasol diluted in 1 litre of water per square metre of soil. After some hours of exposure to the sun, this sticky, milky product forms a flexible 1-mm-thick film which protects the soil against the kinetic energy of falling raindrops (Roose 1975; 1977a). The treatment was tested for four years on three pairs of plots:

• a 7% slope planted to Panicum maximum at 40 × 40 cm intervals;
• a 7% bare slope;
• a 20% bare slope.

Table 32 shows that Curasol considerably reduced soil loss (a reduction of 40 to 75%), and to a lesser extent runoff (a reduction of 20 to 55%). Although its protective effect decreased after three months of violent rain (1200 mm), it was still functional after one year. It had no significant effect on forage yields (Panicum) but was extremely effective against erosion under this plant cover.

It was not a foregone conclusion that use of this plastic glue might reduce runoff, for it could in fact have blocked the soil pores. On-site observation shows that when it is sprayed on to a well-aerated (recently tilled) soil it forms a flexible film which slightly increases runoff in comparison with the control plot for a number of rainstorms, but that after this the porosity of the untreated plot decreases faster than that of the protected plot, tipping the balance in favour of Curasol. The product does not form a unbroken, impermeable film, but coats soil surface aggregates, boosting their resistance to the onslaught of falling rain.

Curasol always allows a certain amount of erosion. Since protection is not uniform and unbroken, water discovers weak spots in the film, and raindrop energy digs holes into which runoff rushes, undermining the base of the microcliffs thus formed and broadening the areas attacked by headward erosion. This means that if there is a plant cover protecting the flexible plastic film against rainfall energy, the Curasol film lasts longer. It should also be noted that the plastic film will not stand up either to the abrasion of grains of sand carried in an active rill or to the passage of heavy machinery (tractors, etc.) and workers, for erosion sets in very fast at broken points.

Erosion (t/ha) and runoff control (KR %) of plastic mulch on bare soil under Panicum (cf. Roose 1975; 1977a)


EROSION (t/ha and % of control)

RUNOFF (mm, % and % of control)


Panicum, P = 7%

Bare soil, = 7%

Bare soil, P = 20%

Panicum, P = 7%

Bare soil, P = 7%

Bare soil, P = 20%



Control t/ha

Curasol % of control

Control t/ha

Curasol % of control

Control t/ha

Curasol % of control

Control mm and % of rain

Curasol % of control

Control mm and % of rain

Curasol % of control

Control mm and % of rain

Curasol % of control

5.70 to 3.71









368 mm


575 mm


423 mm





4.71 to 3.72









190 mm


562 mm


286 mm





4.72 to 3.73









106 mm


593 mm


363 mm





4.73 to 4.74









146 mm


942 mm


425 mm














203 mm


668 mm


374 mm


P = erosion control practices efficiency rate.

Runoff recorded on the control plots is given in two forms:

- depth of runoff sheet in mm
- runoff coefficient as % of depth of rainfall

Although very effective on these sandy soils, treatment with Curasol was unable to bring erosion on bare soil down to under 10 tonnes, the tolerance level on this type of soil. Its cost price (4000 FF/ha in 1973 for an average dose of 60 g/l/m²) and the large amount of water taken to apply it (10 m³/ha) are major drawbacks to its widespread use even in intensive farming. However, Curasol can play a very effective role in fixing road embankments, irrigation channels and scoured surfaces in urban or industrial areas if applied in combination with certain grass seeds and the fertilizer needed for the latter to grow.

By way of comparison, in Côte d'Ivoire it takes 250 days' work to collect 40 to 80 t/ha of brushwood in the bush and spread it on the fields, equal to 4000 FF in 1990. If a field of Guatemala grass (Tripsacum laxum) is available, it takes only 150 days to obtain a thick mulch. Tests have shown that 4 to 6 tonnes of dry straw are enough to obtain satisfactory protection against erosion (Lal 1975), so that the cost price of this technique could be cut still further. The protective value of the various forms of mulching has been shown time and time again and never disproven. If its use is still limited, this is because its applicability in tropical conditions on different soils and in various social contexts has yet to be proven (problems with herbicides and phytosanitary products). Implements capable of aerating the soil without disturbing the mulch also must still be developed, as must no-till cropping systems of proven long-term economic viability.


Since the main difficulty connected with mulching is that of producing and transporting biomass to the field once the soil has been prepared and planted, the obvious answer was to try to produce this biomass on the spot. Experiments with a crop of deep-rooting pulses sown as catch crops under maize or some other cereal have therefore been carried out, first in Brazil twenty years ago (Séguy et al. 1989), then in Nigeria (IITA Station, Ibadan). While the main crop grows and makes use of the top layers of soil, it slows down the growth of the pulse, which, while awaiting better light, sends tap-roots deep under the area, and gradually forms a carpet of leaves and stalks in varying states of decomposition. After the main crop has been harvested, space and light allow the pulse to grow fast during the few weeks when there is still enough water in the soil under the area used by the main crop. During the dry season, the soil is therefore covered by a carpet, fostering mesofauna activity - earthworms in humid tropical areas, and termites in semiarid areas or on too-sandy soil. At the start of the following cropping season, either the dry season has killed off all the pulses, leaving dead cover, or there has been enough water to keep them alive or to give rise to a new living cover from the seeds dropped by the pulses. In the latter case, either the pulses are killed off with a herbicide (31/ha of Gramoxone), taking advantage of this to kill off any other weeds that may have developed, or else they are left to grow, but then cut back before the following crop cycle begins. Instead of being ploughed in to prepare the ground for the following crop, this litter is separated by a toothed disc and the soil is broken up with a pronged implement, with the basic fertilizer and seeds being injected behind it. The soil is swept back over the sowing line, and a small roller presses it down to assure good contact between soil moisture and seeds. In this system, less than 10% of the soil moisture is bare, loosened and vulnerable to erosion. Experience shows that the soil surface is not degraded, the sand remains bound to the organic matter, and there is very little risk of erosion and runoff. The method also offers a certain number of other advantages: first and foremost, as in the original forest environment, it makes it possible to restore the organic matter and return a certain number of nutrients to the surface so that they can then be redistributed in the soil during the course of the rainy season. It is a neat technique, reducing inputs, fertilizers and cropping operations while protecting the soil against the onslaught of rain and erosion. Eliminating runoff can perhaps allow water to be supplied to both crops, and hence reduce the competition between them. The system can be compared to the old method of green manuring, which consisted of introducing a crop when the fallow time came, and ploughing it in before the end of the rainy season. Part of this crop could be used by leaving the cover plant on the soil surface in the form of litter. The use of cover plants, particularly pulses, is well known and very widespread under industrial tree crops such as palm, coconut, rubber, coffee and cacao. It seems that a form of this system could be developed for use under temporary root and tuber crops. Lastly, a more stable system could be found than the intensive systems which call for so many inputs in the form of mineral fertilizers and tillage.

The "erosion control practices factor" P in Wischmeier and Smith's equation for forecasting sheet erosion

United States (cf. Wischmeier and Smith 1978)

Maximum length


- Contour tillage

1 to 8%

122 to 61 metres


9 to 12%

36 m


13 to 16%

24 m


17 to 20%

18 m


21 to 25%



- Contour ridging = contour tillage × 0.5

idem × 0.5

- Contour tillage between grass strips

1 to 8%

40 to 30 m

0.25 (r) 0.50

9 to 16%

24 m

0.30 (r) 0.60

17 to 25%

15 m

0.40 (r) 0.90

West Africa (cf. Roose 1977b)

- tied contour ridging

0.2 to 0.1

- erosion control strips 2 to 4 m wide

0.3 to 0.1

- straw mulch, over 6 t/ha


- Curasol mulch, 60 g/l/m² (depending on slope and crop)

0.5 to 0.2

- temporary pasture or cover plant (depending on cover)

0.5 to 0.01

- low earth bunds protected by stones or rows of perennial grass or low dry stone walls every 80 cm + contour tillage + hoeing + fertilization

0.1 to 0.05

Sierra Leone

for rice

for beans

time needed for 100 m

E t/ha/yr


E t/ha/yr


- horizontal bench

808 hours





- stone bund






- row of stakes






- contour bund






- no method


41 to 55


11 to 55


Contents - Previous - Next