Contents - Previous - Next

Negative off-site effects of erosion

On-site erosion affects individuals and is often viewed fatalistically, whereas the downstream damage disturbs groups who have access to public opinion and the media, and who can organize protests against those responsible.

The negative effects of erosion on yields and the production potential of land vary widely (from nil to heavy). However, the cost of off-site damage in terms of eroded fields is generally much higher, and the effects much more spectacular, amply justifying the majority of large-scale erosion control schemes.

This observation is true of RML, which seeks to maintain communication links in the mountains and protect restructured valleys. It is also true of SPR, which seeks to protect soil, but especially to prevent dams from silting up too fast, and structural works, roads and villages from being destroyed.

Even soil conservation, officially designed to maintain land production capacity, also concentrates on protecting water quality, which is so essential for urban dwellers. This is why the State makes considerable efforts to provide technical and financial assistance to farmers to develop their land (a task undertaken with varying degrees of willingness or coercion in different regions). In the United States, nearly 50% of scientists in the Soil Conservation Service work on water quality and various types of pollution problems rather than on soil protection.

Off-site damage consists firstly of a deterioration in the quality of river water due to the suspended load that accompanies flood waters formed mostly by runoff. Suspended load includes organic matter (a threat to the oxygen essential to river fauna) resulting, for example, from intensive stock farming (liquid manure), as well as nitrogen and phosphorus (from mineral fertilizers used by farmers), which can lead to eutrophication of ponds (invasion by algae which will in turn asphyxiate the fish). While abundant runoff at certain times of year increases peak discharge into spillways, it also reduces supplies to groundwater and the rate of low-water flow. On the one hand, it causes downstream sediment on the river bed and banks to be recycled downstream - an erosion phenomenon often seen in small watercourses in Africa. On the other hand, the reduced low flow in the dry season can no longer carry away pollutants from industry, towns and intensive farming, resulting in the eutrophication of watercourses and the death of many tons of fish each year in Europe. Peak flood sediment loads also cause damage, leaving torrential mud flows at the bottom of fields, in ditches, on roads and in cellars. Once the peak flood is over, considerable amounts of sediment are deposited in lakes, rivers, canals and harbours.

FIGURE 12 Effect of cumulative erosion or scouring on production potential of soil

Advantages of rural infrastructure strategies (RML, SPR, SWC) and rural development (land husbandry) with regard to soil conditions.

This is why there are wide variations in the life-span of dams - an essential consideration in their economic viability - from one region to another, and even within the same region, depending on the respective size of the reservoir and catchment area, but also on climate, plant cover, and watershed management (basin gullies and river banks).

While the Kossou Dam in the tree savannah of central Côte d'Ivoire is unlikely to silt up in a thousand years, the main Maghreb reservoirs have a very short life-span (25 to 60 years) and the hill or check dams (small reservoirs very close to the source of silt) may well last less than two years and no more than ten.

Bearing in mind the cost of even the smallest dam, it is easy to appreciate why such huge efforts are made to reduce sediment load in the Mediterranean area, where the lithology is clay layers, marls, and soft sandstone or schist, alternating with hard limestone or sandstone strata, combined with steep slopes and plant cover often heavily degraded by overgrazing and fires.

In Algeria laudable efforts have been made since 1945 to reforest valley heads (50000 ha) or badlands, check further gullying, control wadis, and manage 300000 ha of cropland by putting in flat and graded channel terracing (built by the SPR service, then by the National Forestry Department). Since 1978 terrace construction has been suspended, following criticism by experts, rejection by farmers, and above all economic problems. Erosion control has been reduced to protection of structural works, reforestation, plant cover for gullies, and construction of major dams. Only RML is left to watch over water quality, irrigation schemes and the needs of urban populations. For small farmers, the only activity of the State today concerns land improvement (i.e. subsoiling of calcareous crusted soils to increase cereal productivity) and the building of small check dams to provide water for stock, household use and a few irrigated hectares at valley heads. Even this policy is questioned by hydrologists, who point out that the level of dam siltation remains the same after upper watershed management. The works of Heusch (1970 and 1982) and Demmak (1982) show that most of the sediment trapped in reservoirs comes from gullying, landslips, the collapse of river banks and streambed displacement. On the basis of erosion control projects intended to reduce downstream damage or preserve land resources in catchment areas, a compromise will be sought allowing work to be carried out in the valleys to trap silt and stabilize banks while managing the watershed to reduce and delay runoff (land improvement, grass banks, farming techniques to cover the soil in winter and replant overgrazed areas). There are methods of economic calculation that provide for selection of the most effective erosion control, balancing the costs of this against the expected damage in the absence of intervention (cf. the CEMAGREF courses in Grenoble).

The economic rationale for land husbandry

Erosion control has in the past been seen as a means of conserving long-term soil productivity, and it has been very difficult to justify the short-term economic viability of erosion control projects. In view of the size of the task, large sums (several thousand million dollars worldwide) have been spent on land protection programmes, but their effectiveness has been limited by the approach used and methods not really suited to social and economic conditions. Since available finance is limited, choices must be based on specific objectives. The most effective and economical methods must be identified; then, on this basis, the best sites for intervention can be worked out, depending on the objectives of each project. Figure 12 shows the different reactions of two types of soil to erosion. Curve 3 shows the rapid productivity loss of a forest soil in which fertility is concentrated near the surface. Curve 1 shows that productivity loss for a deep loess is slight even under heavy erosion, for there is a thick layer of fertile soil, and its water and nutrient storage capacity is barely affected (except by the loss of surface organic matter).

The conventional strategy applied by foresters and land-use planners (RML and SPR) entailed intervention wherever sediment transport is heaviest: steep slopes, gullies, scoured areas and sterile, exhausted soils. Such action has very little effect on the productivity of very degraded soils (yield gains (d y) 1 and 2 in Figure 12) - which is why farmers are reluctant to adopt the practices imposed, such as terraces, which do not improve yields, and why they also resist restricted grazing.

Major improvements in soil productivity (d y 3) will be seen only if work is carried out on soils that still have a good surface fertility, not on exhausted or deep soils (Curve 1). Curve 3 is much sharper in the case of slight erosion than when erosion is already high and soils too degraded. Land husbandry advocates this approach, in which a modest improvement in production systems results in improved rainwater storage, higher biomass production, improved soil cover, and therefore much less erosion.


1. On-site losses in eroded areas: affecting farmers

• Losses of water, fertilizers and pesticides
• Immediate production loss

- in regional terms: 2 to 10%: compensation possible through inputs
- in local terms: 2 to 50% = individual disaster = loss of profit margin

• Loss of arable land

- in world terms: 7 to 10 million ha per year
- in regional terms: 2 to 5%
- in individual farm terms: as much as 20 to 100%
It would take 200 years to destroy all arable land

• Long-term productivity loss = SOIL MEMORY

Reduced depth of topsoil
Reduced water and nutrient storage
Reduced effectiveness of rain and inputs
Reduced economic viability = soil depletion

2. Off-site - or downstream - damage: affecting townspeople

• Deterioration in water quality

pollution of rivers, death of fish, silting up of reservoirs, cost of dredging harbours

• Increase in suspended load (SL)

higher costs for drinking water

• Flooding of inhabited areas

mud flows, sanded up ditches

• Rise in peak flows of rivers

destruction of structural works, bridges, etc.

3. Major consequences for erosion control

Erosion on deep soils causes only off-site damage, but barely affects yields
Erosion on thin soils causes rapid falls in yields

Choice of an economically viable erosion control policy:

RML and SPR on degraded soil bring little improvement in yields, but do reduce sediment transport

SWC reduces soil loss from fields, conserves soil, but does not improve its fertility


is concerned primarily with productive land
improves yields
by restoring its fertility
reducing runoff
increasing plant cover and rooting depth

indirectly or in the long-term reduces peak flows and sediment transport by rivers


• Soil degradation creates serious problems only for some smallholders.

• Off-site damage is much more costly, and forces the State to act.

• Drainage is generally improved (concrete channels} to reduce damage.

• Land husbandry seeks to act on the cause and improve infiltration by modifying production systems.

If the money earmarked for soil protection is to be used to the best advantage, it is logical to invest it in the best productive land in order to prevent its degradation or else to restore production on soils just beginning to deteriorate, rather than pouring money into completely scoured land which requires large-scale investment and considerable time before recovery to an acceptable level of productivity. However, there are cases (e.g. Haiti) where the tops of very eroded hills (known as "finished land") are the source of heavy runoff which can destroy fertile soil on the lower slopes. These hill-tops must therefore be planted and protected from grazing, and the runoff water collected in tanks for supplementary irrigation of the deep soils of the lower slopes.

The conventional intervention on scoured soils rarely has positive effects on yields (d y1 and d y2). To enlist the farmers' involvement, it is best to choose land that is still viable and will respond fast and significantly to new production systems.

Nevertheless, from a social point of view, it is not possible to abandon all degraded and unprofitable land, for this would accelerate migration, with the attendant problems. There are also circumstances where erosion-degraded soils are the only kind available and can in fact be restored with a small financial outlay. This is the case for certain land on the Mossi plateau (Burkina Faso), which can at present be restored in one year using the zaï method (300 hours of digging and pick-work + 3 tonnes of manure and its transportation - Roose, Dughe and Rodriguez 1992).

On the other hand, if the aim is that of limiting sediment transport and the risk of siltation - or if the objectives are social (providing work for the poorest members of society in high emigration areas) - action must be taken in the most severely eroded areas and those closest to the river bed (gully, bank and torrent control).

In conclusion, the objectives must be clearly defined before erosion control methods can be advanced.

If the aim is that of reducing land degradation (on-site, small-farmer perspective), farming systems must be developed that can rapidly increase soil productivity (increased infiltration capacity, fertilizer input, selected seeds, plant health care, etc.): farmers will quickly become interested in participating in such projects.

If, on the other hand, the aim is that of reducing siltation risks (off-site perspective), the only way that will attract farmers in the high valleys is incentives or compensation for the loss of productive land, hours of work, and drawbacks of the structures (channels, banks, terraces, etc.). These farmers cannot really be relied on to maintain the structures, which are always dangerous if they overflow (gullying) (Roose 1991).

A compromise is often reached in practice: putting a rapid stop to the most active processes on degraded land and gullies, intercepting runoff on scoured gravel slopes and elsewhere, initiating a gradual change in cropping systems on good but deteriorating lands.

Criteria for the success of soil conservation projects

The 136 experts taking part in the 1987 Puerto Rico Seminar - research experts and developers, all of whom had practical experience of soil conservation in steeply sloping tropical areas - talked at length about the reasons for soil conservation project success (Sanders 1988, Hudson 1991), and here we offer our conclusions, including other experiences as well (Critchley, Reij and Seznec 1992).

No universal solution. While rules for water and soil-fertility conservation may be universally valid, environmental, social and economic conditions vary so widely that universal solutions must not be extended. The effectiveness, cost and limitations of each technique must first be studied, then the regions in each country that have more or less similar environmental conditions must be defined, and lastly an array of solutions must be proposed on the basis of local conditions - slope, land tenure system, economic possibilities, farmer training, and the availability of labour, equipment and supplies.

Taking account of farmers' immediate priorities. i.e. increased production, security, income and standard of living, getting the best return from labour. If soil conservation is planned exclusively on the basis of the off-site problems caused by siltation or flooding, high-valley farmers will not feel involved, and the State will have to step in and assume responsibility for rural hydraulic works. Soil and water conservation requires a major effort in terms of shaping the landscape, managing runoff, altering cropping techniques, and maintaining structures through the years. If farmers see that the soil continues to deteriorate (through mineralization of organic matter and rain splash) and crop yields keep dropping, they will quickly stop expanding - or even maintaining - the erosion control mechanisms that cost so much effort for nothing in return. Inexpensive, effective water management systems are therefore needed that can also be combined with a package of techniques to improve yields and net farmer income substantially, reduce risks or simplify work (new and more profitable crops, markets where farmers can sell at higher prices and buy selected seed, fertilizers, herbicides, pesticides). Since the soil already tends to be poor and poorly structured, measures must be taken to ensure the rapid recovery of soil fertility (turning in fermented organic matter to improve structure, restoring macroporosity, improving infiltration and water- and nutrient-storage capacity, correcting pH and soil deficiencies indicated by plants, providing nutrients directly to plants as and when necessary in cases of reduced storage capacity, encouraging deep rooting) while also minimizing water and nutrient loss through erosion and/or deep drainage.

Taking traditional methods as a starting point. These all-too-often scorned methods must be reviewed, and variations from one farmer to another assessed, together with their limitations, economic potential and possible improvements. Conclusions can also be drawn on the environment, the water balance and major risks (very dry years or exceptionally severe storms). Traditional farmers cannot allow themselves the luxury of harvest failure, and therefore take account of what happens on the land during exceptional phases (Roose 1992c).

Highly flexible long-term programmes. Since the aim is a radical change in behaviour (based on the knowledge that erosion is not inevitable, but the result of unsound management), it will take time to convince people, to finalize the techniques and to train the future leaders of rural communities. Some financiers already realize that they cannot demand the same immediate rates of return nor the same project durations when the intention is to improve the environment; however, evaluation teams still need to be persuaded that it is difficult to fix a time-frame for each operation when it is difficult to know a priori which technological package will be acceptable to the inhabitants. One poor rainy season can also hold up progress on a project. Financing should therefore be staggered, while one time-frame is needed for evaluation by farmers and local technical staff, and another by international experts.

Modest projects that are gradually expanded (replicability). Since it is vital for the rural community to take responsibility for its environment, it is best to start on a modest scale with simple operations to intensify production, and then move on - depending on the farmers' level of participation - to the various phases of finalizing, implementing, evaluating, maintaining and extending techniques to the whole of a slope, local area, hill or small watershed.

The need for land security. Farmers who rent their land are not sure of holding on to it once it has been developed, for the owner may be afraid that the improvements are an attempt at appropriation and therefore take it back, perhaps even renting it out to a competitor at a higher rate. This is a major problem in connection with agroforestry.

Clear examples of this were seen in a small watershed near Jacmel in Haiti, where farmers first chose to improve Plot A - land in good condition, well covered by a multi-storey garden surrounded by a hedge to protect the house and the produce of this fully-owned plot against theft. Only later would they turn their attention to the more degraded land, which was being farmed but not managed since it was rented from absentee landlords.

In other places (e.g. Yatenga in Burkina Faso) it has been observed that farmers view stone lines, grass strips or trees planted around their plots much more as confirmation of property rights and as aids to water and nutrient management than as protection against erosion. Without the landlord's agreement, a farmer will seldom feel inclined to improve rented land.

Making use of existing structures. When a totally new structure is set up, there is a danger that not enough attention will be paid to the views and customs of the local population, so that the erosion control mechanisms will be abandoned at the end of the project. It is better to choose NGOs and local organizations with care, and to bolster existing government structures (with vehicles, staff training and means for self-advancement); this is the price of ensuring sustainable project impact.

Taking account of local production systems and family constraints. Often the first question is to understand the economic, social and political organization of a farming community (village, district, etc.), and to grasp the constraints (availability of labour, energy, manure and inputs, and the possibility of marketing or processing surpluses through livestock production, crafts, trade).

Women account for over half the work force on SWC schemes, which means training must be planned for female groups. Traditional strategies for water management, soil fertility and erosion or acidification control must be reviewed, and representatives of farmers' groups chosen to communicate, train and gradually encourage farmers to introduce new techniques, taking care to avoid creating tension between "model progressive farmers" (who often receive too much aid to be truly representative) and the many conservative and mistrustful graduates of previous unfortunate experiences.

Initiating action simultaneously for agriculture, animal husbandry and tree production. Small farmers are generally concerned first and foremost with food crops (for food security), and next with animal husbandry - their "savings bank" or cash reserve in case of need. The only trees "grown" are fruit species: trees are traditionally considered a gift of nature to be used according to needs, and land, wood, trees and their fruit do not necessarily belong to the same owner. In some countries, the inland water and forestry department issues permits only to recognized woodcutters, who fell trees according to the market for fuel- or construction wood (posts and beams) in the local town, with no reference to the owner of the land. Landowners will quite obviously be reluctant to plant trees if they have no guarantee of profit from them. If "wood has no owner until it has been cut into logs," it is easier to understand the destruction of the tree cover in West Africa, since anybody can lop branches off a tree to provide forage for his/her flock.

Similarly, temporarily ruling a part of communal rangeland off-limits in order to allow regeneration of perennial plants and fodder shrubs is something of a challenge, given the risk that people from neighbouring villages will then send their hungry animals to graze it. However, when population density is high and land pressure intense, there comes a time when farmers realize that since the time when all the trees were cleared the microclimate has become drier, runoff damage more extensive (gullying) and water a rare commodity in the dry season. However, if trees are to be reintroduced in the form of hedges, contour lines or orchards, livestock cannot be left to wander at will, and a more intensive stock farming system must be developed (semi-confinement with grazing only on the way to the watering point, the clearing of tracks and forests, and tethering on fallow). Providing litter, supplementary fodder (crop residues) and mineral supplements for animals certainly entails more work, but it makes stock farming more profitable (less loss, improved health, better-quality meat), gives a better return from the dispersed biomass, and improves the quality and quantity of manure: up to 5 tonnes of composted manure/farmer/ha/yr in Rwanda and Burundi.

Traditional land-use planners tend to make a clear-cut distinction between land for crops, animal husbandry and forests, whereas the positive interaction between trees, crops and stock should be exploited. Stock draw as much advantage from crop residues as from pasture, particularly in forest areas. In the Mediterranean region, forest areas need herds to reduce fire risks by grazing the bushy undergrowth. Elsewhere, trees profit from association with crops, for they grow better on deep tilled and weeded soils than on wastelands that are too exhausted to make cropping economically viable (viz. village woods, which are often poorly managed since nobody knows who owns the wood). Crops need manure, and particularly nitrogen, phosphorus and other nutrients that are cropped over a large area and subsequently excreted by animals kept overnight in stables or corrals. Trees can help crops, providing litter, recycling nutrients from deep in the ground, and reducing wind speed and the risk of wind erosion. So, although each kind of land will have its main purpose, all positive interactions between these three sectors of agriculture must be fostered.

Subsidies, food aid, wages. It is now agreed that incentives, gifts of food, tools, wages, etc. (anything given in exchange for participation in a development project), should be limited, for what often happens is that participants lose interest when the assistance stops. Particularly in the case of private land under development schemes, aid must be kept to a minimum (fertilizer, trees and selected seed, etc.), and eliminated as soon as the positive effects are clear to the recipients. However, there are some especially harsh environments, e.g. the Sahel, with large landless families and young people in search of work, where some kind of wage must be paid if a sizeable labour force is to be on hand during the dry season: without this indispensable input for group survival, the most able-bodied adults emigrate to other countries to earn more from their work. Even in cases such as this, however, payments must be kept small to allow the participants to develop a sense of ownership vis-a-vis the improvements and to feel responsible for their upkeep and protection. On the other hand, it is a good idea to make the farmers' task easier by providing tools and other items at subsidized prices so that they have to spend less time on management activities (picks, shovels, pickaxes, sickles, fertilizers, wheelbarrows, carts to transport stones).

Training of men and women farmers in simple techniques. If the schemes are to continue to spread once the project itself is over, particular care must be taken to choose simple techniques accessible to all villagers once one villager has been trained, and needing no input that cannot be produced in the village. Each person must be able to work his or her own land as and when he/she wants.

Projects that introduce heavy machinery offer the best guarantee of rapid development of SPR in the field followed by failure once the project ends for lack of upkeep on the part of farmers. This approach short-circuits the dialogue phases and the preliminary tests to assess project feasibility, effectiveness and economic viability of the methods with the farmers.

Project design. At present it takes two or three hurried field missions to formulate a project, with too little time to talk with farmers about their problems and traditional methods. Each mission then draws up its report without much concern for the findings of its predecessor(s). Some people are now recommending that the three phases should be condensed into one, so that a single team has time to reach a deeper understanding of the country and gather first-hand information in the field.

Research and project monitoring-evaluation. There are still many technical aspects of erosion control to be clarified, but study of the interlinkage between the human environment and technical know-how (particularly the economic cost of erosion) is clearly needed. Research experts unfortunately seldom have the means to set up individual erosion control schemes. On the other hand, reviewing the history of earlier projects, and monitoring and frequent evaluation of new ones, could make it easier to grasp the technical and human constraints.

Contents - Previous - Next