Introduction

Since the Earth first appeared it has been shaped by erosion... and for over seven thousand years human beings have pitted themselves against erosion, trying to defend their lands against the assaults of rain and runoff (Lowdermilk 1953). One may therefore wonder whether there is anything left for research to discover, or anything that has not already been said.

The scientific study of erosion, however, did not start until the early 20th century, first in Germany (Wollny), then 40 years later in the United States of America at the time of the Great Depression. Under pressure from a public panic-stricken by duststorms that were darkening the midday sun (the Dust Bowl), the American Government commissioned Bennett to set up the famous Soil Conservation Service, with about ten field stations to measure runoff and sediment load. And it was not until the 1940s that a scientist, shut away in his laboratory while bombs rained down over Europe, discovered that the kinetic energy developed by falling raindrops was the source of soil surface degradation, runoff and a major part of the erosion observed on cultivated land (the splash effect) (Ellison 1944).

Only in the 1950s, following the Madison Congress of the International Soil Science Association, did American methods of measuring runoff and erosion on small plots spread to French-speaking (F. Fournier) and English-speaking (N.W. Hudson) Africa, then Latin America and, more recently, Asia and Europe.

The United States therefore had a 20-year start on the rest of the world in collecting data and developing the first empirical model, the Universal Soil Loss Equation (USLE), to predict soil loss at plot level. The sole claim made for this model is that of helping engineers to design soil conservation systems for specific soil, climatic, topographical and plant-cover conditions, and it has disappointed many scientists who have applied it inappropriately outside this compass. Although it has eventually been seen that the USLE is not universal but is confined to circumstances where erosive energy comes not only from rain (but also from runoff, as in upland areas and on soils rich in swelling clay, or from gravity, as in landslips), this somewhat dated model is still today - and will be for some time to come - the only one sufficiently balanced to be used in many countries where runoff is associated with soil surface degradation. It will take a further 12 years or so to perfect new physical models and adjust them for each region - nor is it certain that they will perform better than the latest versions of the USLE, as long as the latter are restricted to their intended sphere of sheet erosion.

Similarly, in the field of soil conservation, people have long been satisfied to apply American-developed methods throughout the world, without testing their suitability for local conditions. However, in the last ten years, the significance of climatic, social, demographic and economic elements has been recognized, and this fact, together with new trial results, has raised questions about the treatments prescribed in all the manuals since Bennett.

It is primarily a matter of the rising failure rate for erosion control projects in developing countries (Hudson 1992), for American methods do not translate successfully to tropical countries. Local farmers who are familiar with land husbandry strategies in their traditional agricultures have been disappointed by the modern soil conservation methods imposed by international experts and government authorities: they require a lot of hard work and upkeep, and provide no improvement in yields. Even if the soil cover is kept in place, tropical soils are usually so poor that their fertility has to be restored and their infiltration capacity improved if they are to produce significantly more than traditional systems.

Also, farmers will sometimes abandon such developed land or destroy trees donated under projects, suspecting the State of wanting to get its hands on their land - for land traditionally belongs to those who care for it, and trees mark its boundaries. Hence the spate of misunderstandings and failures throughout the Maghreb and West Africa.

Even in the United States, evaluation of 60 years of water and soil conservation - which have swallowed up billions of dollars - reveals only partial success. There are still major problems of pollution (linked to animal husbandry, chemical fertilizers and industry) and of sediment transport in rivers: 25% of tilled land loses over 12 t/ha/yr of sediment, the official tolerance level for deep soils. The situation today would of course be worse had nothing been done, but the need for a change in approach seems clear. Hitherto, soil protection has been carried out by volunteer farmers with State assistance, since everyone realized that the environment must be protected in order to ensure land productivity for future generations.

The American survey shows that erosion does not necessarily lead to a fall in yields, particularly on thick loess deposits. Today the State tends to introduce coercive clauses; for example, if farmers do not participate in a given programme to freeze fragile land, swamps and mountains, or do not abide by instructions for erosion control on tilled land, they will have no right to government subsidies intended to encourage them to diversify production.

Analysis of the effects of selective erosion on tropical land, especially forest areas where chemical and biological fertility is concentrated in the top 25 cm of the soil, shows that:

• it is not enough to improve degraded land (soil protection and rehabilitation) in order to address farmers' problems;

• even soil and water conservation (SWC) tends to be unwelcome, since it requires considerable work and brings little improvement in yields.

To meet the challenge of this century and feed a population that doubles every 20 years, not only must the serious processes of gullying and landslides that produce sediment load in rivers (the sphere of State concern) be halted, but also water and nutrients on good land must be correctly managed before degradation sets in, and degraded but potentially productive soils must be rehabilitated. Only farming communities can manage the rural environment, and if farmers' cooperation is to be assured, they must be shown, on their own land, that sound land management (including a range of technological packages) can quickly increase their output and returns, optimize their labour, and make their efforts more profitable, while effectively protecting their land capital.

It should be noted that it is not always necessary to resort to sophisticated techniques with expensive inputs, or to import machinery that is hard to maintain. Astonishing results can often be achieved simply by combining scientific knowledge of the phenomena to be corrected with traditional know-how. This is the case with zaï, a traditional method of rehabilitating degraded soil among the Mossi of Burkina Faso. With no other input than labour (350 hours/ha) and manure (3 t/ha/yr), 600 to 1000 kg of grain can be grown on the regenerated fields. And with a little supplementary mineral fertilizer (N and P) results considerably higher than the national average (600 kg/ha/yr) should be achieved (Roose, Dugue and Rodriguez 1992).

Certain favourable circumstances have led to a change in farmers' attitudes to soil conservation projects.

First, drought has brought much suffering to the people of the Sahel and reduced livestock by half. It has shown farmers that they must change their practice of extensive farming, balance their livestock holdings against availability of forage, and organize village-based land-use planning, as the boundaries of villages are now known. This crisis has revealed the importance not only of protecting land against erosion (expressed in t/ha/yr), but above all of managing the available water (reducing runoff) and nutrients (mulch, manure, compost, and mineral supplements), and halting the water and nutrient losses caused initially by erosion and then by drainage.

Secondly, and strangely enough, the "cost-pricing" operation for mineral fertilizers required by the World Bank in Africa has shown the validity of organic fertilizers and, most importantly, the low stocks of nutrients easily taken up by plants in most tropical soils (other than some volcanic soils, brown vertisols or alluvial soils). It is extremely dangerous to the nutritional status of both human beings and livestock to be reduced to simply recycling the biomass (dung, paddock litter, compost, mulch, and ever-shorter fallow) which inevitably translates into soil deficiencies (N, P, K and Ca + Mg in very acid soils). Mineral supplements incorporated in compost are essential for any intensification of farming, if only to allow the growth of atmospheric nitrogen-fixing legumes.

The third circumstance that has aroused interest in land husbandry projects is population growth (an increase of 2.5 to 3.7% per year, or a doubling every 20 years) as a result of improved hygiene and diet. In West Africa the boundaries of village lands used to be uncertain, if not indeed a bone of contention, but land was plentiful and traditional chiefs used to grant plots to anyone asking to farm them. Nowadays, land availability is frequently exhausted, and instead of expanding croplands with little thought to their degradation, people have to live exactly where they are, making the most of natural resources.

Three strategies are generally developed to cope with land pressure in African countries:

• emigration, either for the dry season or for good, of some of the children to the less arid zones where there are better returns from work;

• supplementing farm revenue with other activities - craft work, trading, teaching, etc.;

• improving land management, intensifying and diversifying production by choosing more profitable lines (specialized livestock, fodder crop production, vegetables, fruit, forestry products for fuelwood and poles, tree nurseries, etc.).

In Yatenga in north-western Burkina Faso, rural development project activities have enabled some young people to find enough resources locally to live decently. As a result of extension work, and of drought or pressure on land as the case may be, farmers today are much better disposed toward village-based land-use planning projects. Their aim is to protect their land resources, but especially to manage the scarce available water and nutrients in the biomass. Or they may simply want to own the land - for, after the various upheavals, no one knows for sure if the land belongs to the village community, the State, citizens with official documentation, or simply whomever develops and farms it.

Finally, research has also advanced in a number of fields. Experts have measured the relative effect of the various factors influencing erosion. They have shown that the slope gradient is more important than its length, whose effect is closely linked to the state of the soil surface, especially its roughness. Under certain conditions, the actual topographical position is extremely important, since the lower slopes quickly becomes waterlogged from hypodermic runoff from uphill or from a rising water table near rivers. Under certain conditions (e.g. the chalky, clayey soils of the Mediterranean region), sheet erosion on hillsides is less serious than regressive gullying, which starts from the streams, attacking rich alluvial soil and irrigated terraces before cutting into the slopes. This means that erosion control operations should not necessarily concentrate on steep slopes. Runoff from barely sloping broad pediments and slaking loamy soils can be more serious than from steep slopes that are well protected by leafy vegetation or a gravel pavement (Heusch 1970). A river can swell in a rainstorm without runoff from steep slopes (the theory of the partial contribution of a catchment basin to runoff; Cosandey 1983, Campbell 1983).

Soil is not necessarily a "non-renewable natural resource". While it is true that if the thin layer of a rendzine covering a chalky rock is lost, that land will be lost for thousands of years and runoff water will concentrate there, if the six rules for restoring soil fertility (page 36, Chapter 2) are respected, it will take one to five years to bring life and productivity back to totally degraded and abandoned soils (e.g. the tropical ferruginous soils rehabilitated by the zaï method in Burkina Faso).

Soil conservation has hitherto been seen as a long-term investment in order to protect future generations' land legacy, and this was in fact the theme of the fifth ISCO conference in Bangkok (Rimwanich 1988). The new strategy of land husbandry represents an attempt to solve the immediate problems facing farmers: ensuring a clear increase in biomass production and income by improving the management of surface water and nutrients on the best land, rehabilitating degraded land that has potential (a sufficiently deep profile), finding the least expensive way of stopping gullying, and collecting runoff water in order to establish core areas of agricultural intensification. Insofar as the farmers must be trained to protect their environment, matters must be viewed from their perspective; in other words, any effort must see a return - and very quickly.

Progress on the technological level is also being made today. For example, it has been realized that mechanical methods of erosion control (terracing, drainage ditches, diversion bunds) are not the main thing, but must be kept to a minimum, using the simplest and cheapest methods as back-up to more effective biological methods (Hudson 1992). Other methods of runoff control seem to be better suited to the African smallholder than the diversion works recommended by Bennett for large-scale mechanized farming in the United States.

Farmers are often more ready to accept water (and nutrient) management methods such as water harvesting in semi-arid zones, total infiltration (mulching) or dissipation of runoff energy through use of grass banks, hedges or stone bunds, for these are approaches that are closer to their traditional methods and enable them to improve security if not production levels.

Another major issue is that of tillage.

The validity of deep ploughing and heavy mechanization is being re-examined, for, in contrast to their success in allowing an immediate increase in infiltration, rooting and yields (more than 30% to 50% on soils capable of storing the extra infiltrated water), they also speed up mineralization of organic matter in the soil, destroy its stable macroporosity and structure, increase hydraulic differentiation in the soil profile, reduce its cohesiveness (and thus its resistance to runoff) and in the medium term (10 to 30 years) accelerate its degradation. Major efforts are at present being made in Africa and elsewhere (United States, Brazil, Europe) to develop cropping systems that use minimum tillage, confining the operation to breaking up the soil with a toothed implement along planting rows, which also receive fertilizer.

In the Sudanian zone of Cameroon, for instance, 10 to 15 years of annual ploughing + hoeing + ridging under intensive cropping of cotton + cereals are enough to induce degradation of tropical ferruginous soils, which are all the more fragile, being sandy, poor in organic matter (less than 1%) and exposed to violent rains (Boli, Bep and Roose 1991). Thirty years of fallow, burning and extensive grazing will not sufficiently improve soil fertility: carbon increases from 0.3% to 0.6%, nitrogen remains at around one-tenth the rate of carbon, and the pH goes up by one unit (5 to 6). Animals are the most effective means of improving soil quality: in earthworm casts, Trinervitermes termites' nests and former overnight cattle corrals, carbon can reach 1% and the pH exceed 6.5.

There must thus be a return to farming systems similar to forestry systems in which the soil is never completely bare, receiving regular mineral and organic inputs from the litter. As in traditional cropping systems, an attempt is now being made to reinstate spatial variation within the cropped zone, plant deep-rooting trees that will bring dispersed nutrients to the surface, rear animals that enhance the biomass and concentrate scattered nutrients and grow crops combined with an under-storey of plant cover (weeds or a carpeting of legumes).

Developed village land is no longer strictly divided into forest area (livestock help to control weeds, as do certain interplanted crops), rangeland area (forage shrubs play a major role in improving forage quality, especially during the dry season), inhabited area (the surrounding highly intensive multi-storey gardens are an important source of revenue) and cropped area. There are thus many positive interactions among trees, livestock and crops (see the ICRAF studies).

It was felt that it would be useful to present data gathered over the past 40 years by French-speaking scientists in Africa, Latin America and Europe, in order to provide a good overview of these new situations and a whole new approach to protecting agrosystems.

In Part One, after defining the terms to be used, the range of different situations in terms of processes, timescales and places, the aims of those directly involved, and the demographic, sociological and economic conditions of the farmers are indicated.

Part Two contains a brief study of the various processes and a more detailed one of "early" forms of erosion, i.e. sheet and rill erosion and dry mechanical erosion. A systematic analysis of factors governing erosion within the framework of Wischmeier and Smith's 1978 Universal Soil Loss Equation (USLE) for predicting soil loss leads naturally to proposals for a practical approach to defining erosion control.

Lastly, Part Three presents a series of case studies from densely populated tropical mountainous areas (Rwanda, Ecuador, Algeria, Cameroon), subequatorial areas (Côte d'Ivoire), semi-arid tropical areas (Burkina Faso, Mali) and temperate zones (northern France).

There is no intention here of denying the responsibility of the State in the spheres of land-use planning, rural infrastructure, mountain reforestation, torrent control, protection of rivers, dams and other engineering works such as the rehabilitation of mountainous terrain, teaching people to respect their environment, training specialized technical staff, and subsidies for upland agriculture to prevent emigration. However, it may be helpful to complement this hydraulic infrastructural approach with one from the perspective of rural agricultural development (farmers and herders) that enlists the solidarity of rural communities in the upkeep and improved management of the natural resources (water + soil + nutrients) that they have inherited and must responsibly bequeath to future generations.

This work has evolved from a course, "Land Husbandry as a Tool of Land Management," which has been given over the past seven years to agricultural engineers, foresters and water technicians at CNEARC and ENGREF in Montpellier in France, ETSHER in Ouagadougou, Burkina Faso, Chad and in Haiti. It is a working document which it is hoped will be improved as more information and trial results come in. It was produced with a view to offering constructive new ideas and encouragement to agricultural experts in NGOs and national and international institutions who have the task of working in the field to improve people's standard of living and the health of the land that feeds them.

Contents - Previous - Next