Rehabilitating watersheds: Reforestation, though necessary, is only part of a broader development strategy to preserve watershed environments

The key to arresting current degradation of upland catchment areas lies in enabling farmers to establish appropriate farming practices combined with physical measures to check erosion and flooding. Reforestation is, in many situations, only a partial solution. A broader approach is needed.


The key to arresting current degradation of upland catchment areas lies in enabling farmers to establish appropriate farming practices combined with physical measures to check erosion and flooding. Reforestation is, in many situations, only a partial solution. A broader approach is needed.

John Spears

Half of the world’s population lives in or adjacent to mountain watershed environments and is affected by the way these are farmed. Rising human and livestock populations in developing countries, and the mounting pressure on scarce upland and forest resources, have led to major environmental degradation in the present century. They are the cause of much of the catastrophic flooding, loss of human life, drying up of perennial rivers, increased sedimentation of dams, and disruption of downstream agriculture frequently reported in the international press.

Studies have shown, for instance, that the denudation of water catchment areas has caused the flooding in the Indus River system in Pakistan to be far higher in the last 25 years than during the previous 60 years, and has led to serious silting of the dams and canals of Pakistan’s irrigation system. The cost of repairing flood damage below the Himalayan catchments in India has been, on average, US$250 million a year in recent years, in addition to the loss of production and livelihood suffered by millions. While part of this damage is the result of natural geological erosion, much of it is attributable to excessive population pressure and misuse of land. Severe soil erosion has occurred in the Ethiopian Highlands, and in Java and the Philippines, where five million hectares of denuded, formerly forested watersheds are a source of increasing downstream flooding and disruption of agriculture.

The litany of examples could go on. They illustrate one fact: that as populations increase in developing countries, more and more people are forced onto land that becomes less and less productive. These marginal areas inevitably include the mountainous regions with their thin soils and fragile ecological systems that are particularly susceptible to abuse. Ironically, it is generally the poor who are pushed into these areas, attempting to survive on increasingly unproductive and eroded land without being able—or even knowing how—to prevent the degradation of their source of survival.

The only long-term answer to this problem is to reduce the number of people living off these marginal watersheds. And this could happen, eventually, if economic development could raise the productivity of land and provide more job opportunities in the lowlands. Once population pressure is lifted from the poorer mountainous areas, they can be maintained through sustainable agriculture, forestry, and small-scale industry—as hilly watersheds are being maintained successfully in Austria, Japan, and Switzerland. But the prospect of this happening in time to save the critically overused watersheds in most developing countries is dim. The degradation could be checked if the authorities were willing to force people to move out of endangered areas or were ready (and able) to spend the money needed to tackle the large-scale rehabilitation that is needed. The former is a politically difficult option. As for the latter, expenditure on watershed rehabilitation is usually shelved in favor of shorter-term development priorities.

Given a clearer perception of the negative consequences of ecological damage in watersheds and a political commitment to tackle the issue, there is, nevertheless, a lot that could be done now, at least to prevent the situation from becoming worse. Enough is known about appropriate farming systems, about soil conservation, irrigation, flood control techniques, and reforestation to justify immediate action to contain the current degradation. This article draws on the World Bank’s experience to discuss a multifaceted approach to the rehabilitation and maintenance of watersheds. The key to containment is to improve the practices and productivity of upland farmers; these improvements must be accompanied by physical measures to minimize erosion and flooding and to rehabilitate the land.

Farming practices

Watershed projects deal with people. Investment in flood control structures, reforestation, or soil conservation work will, in themselves, achieve little unless the farmers and herders residing in the upland catchment area are given the means to move away from ecologically destructive shifting cultivation. This practice involves felling forests over a very wide area and then cultivating for short periods, frequently without regard to topography or fragility of soil. Once settled in more stable communities, the people can adopt more intensive, sustainable farming systems. So investments in conservation have to be matched—and often exceeded—by substantial investments in improvements in farming practices and inputs, social services, and marketing systems. Probably one of the most difficult aspects of designing watershed projects is how to achieve in practice this balance between giving small farmers the maximum opportunity to improve their incomes while arresting high rates of soil erosion.

In establishing appropriate farming systems for these areas, one of the main problems is frequently that excessive numbers of livestock are kept on the land. A strategy to reduce the density of livestock would aim primarily to improve the draught power and milk production of fewer and better quality animals. This can be achieved in several ways: by improving the quality of fodder produced, by encouraging stall feeding, by introducing improved animals, and by providing effective marketing systems.

Fodder development, for instance, may involve a variety of activities. Communal grazing lands, grasslands, and degraded forest lands may need to be closed temporarily to allow natural grasses to regenerate and to permit planting of fodder trees, such as acacias to feed goats. Annual fodder species (like winter oats or legumes) may be introduced as a second rotational crop in permanent cropping areas. Fodder trees, pasture grasses, and legumes may be planted on “bunds” (earthern levees), on risers of terraced farmlands, near houses, and in other marginal land spots. Livestock exchange programs could be established to encourage farmers to improve the productivity of their cattle by trading surplus bullocks and low-grade cows for improved breeds. Such measures are an integral component of the Kandi Watershed project in India, which builds on successful small-scale programs already initiated by the Government.

Where a high proportion of the population in the catchment area is dependent on crop-based agriculture, the productivity of the land can be increased by improved cultivation practices. These may include plowing on the contour, preparing smoother seedbeds, sowing seeds in time on the contours, increasing plant cover, and using improved seed, fertilizers, and pesticides. On steep slopes where sheet erosion has to be arrested, farmers can be encouraged to plant fruit trees on the contour, to mulch them, and to interplant trees with legumes and pulses. In some places, fodder tree farming, bamboo plantations, or grasses, such as bhabhar grass, might be more appropriate. (Such activities can, in many situations, be undertaken by paid, landless labor.) This will reduce the likelihood of extending food crop production on steep, erosion-prone slopes.

The key to securing people’s participation in such programs lies in designing broad strategies based on a better understanding of their perceived needs and priorities and in particular of local land tenure and rights regarding the use of land. This implies that enough time has to be spent on sociological studies in order to define the type of incentives needed to elicit farmers’ cooperation. The Hazara project in Pakistan, for example, includes a provision for detailed studies of the sociological issues involved in livestock and ownership. Preliminary results indicate a far more complex pattern of land tenure and use than was initially anticipated, and the implementation of the project will probably be delayed.

Experience has shown that other factors are also important. These may include compensatory payments to farmers excluded from upland grazing areas; the timely provision of inputs such as improved seeds, fertilizer, and credit; the construction of improved access roads and feeder tracks; and the provision of social services—such as improved water supplies, schools, and health clinics. The Bank-financed Northern Agriculture Development Project in Thailand provides a good illustration of this approach. While 25 per cent of the total project cost was allocated to soil conservation and reforestation, 44 per cent of the total investment went to agricultural improvements. By intensifying crop production on a smaller area of land, the project has reduced farmers’ dependence on shifting cultivation, which should lead to more effective protection of the remaining forests.

The regulation of upland grazing and voluntary closure of rangelands, in particular, may frequently run up against sociological and religious attitudes, especially in Africa and Asia, which result in excess livestock being maintained relative to the carrying capacity of the land. Time is needed to change these. Experience with improving ranching practices also shows that herders need a clear demonstration that it is more profitable to keep fewer but better livestock. This may mean establishing more effective marketing systems or providing attractive savings mechanisms. Improved marketing systems introduced in Kenya in the 1960s successfully reduced stock pressure on marginal land in the Baringo and Isiolo districts of the northern region. Other elements contributing to the success of the scheme were the existence of stock routes (which minimize the risk of spreading disease), holding grounds to fatten the cattle, well-organized slaughterhouses, and meat marketing outlets.


It has been reasonably well established that in steep mountain areas of high rainfall, undisturbed natural forest provides the optimum catchment cover for protecting water and soil resources. The forest tree roots act as a sponge, releasing water gradually throughout the year, thereby ensuring perennial stream flow for farmers downstream, while providing optimum retention of nutrients and soils.

However, evidence from East Africa and elsewhere, where comparative studies have been carried out in upland catchment areas over relatively long periods of time, suggests that perennial tree crops such as tea, oil palm, rubber, or coconut can be almost as effective as natural forest—provided that their cultivation is combined with soil conservation measures such as terracing. The same is true for fast-growing forestry plantations, although care is needed not to crowd certain tree species, such as eucalyptus and wattle, which could shade out all ground vegetation cover.

The conventional wisdom that reforestation is needed to restore eroded catchment is not borne out by experience, at least in the developing world. Reforestation can be extremely expensive (upward of $1,000 per hectare, in 1981 U.S. dollars). If the main objective is to restore the vegetation to protect the soil and to regularize stream flow, a much cheaper solution may be to leave an area to regenerate naturally or to augment the natural regeneration with seeding of grasses, legumes, or low shrubs. Grass cover or low shrubs are particularly appropriate for conserving soils in semiarid areas and for ensuring maximum water yield, because evapotranspiration from tall trees is far higher than from grassland or low-canopy shrubs. Reforestation could be confined to gulleys where there is serious risk of deepening erosion unless early remedial measures are taken.

A strategy to use reforestation to prevent erosion would normally involve three concurrent activities:

  • the re-establishment of vegetation cover in denuded areas;

  • the provision of fuelwood, fodder, and timber (for agricultural implements, fencing, house construction, and so on) of desirable species in sufficient quantity planted around homesteads or at convenient distances from villages where they will be utilized;

  • the rehabilitation of existing natural forests (which have protective and productive functions) and the development of appropriate management plans and harvesting rules.

These objectives could be met by establishing nurseries to supply the trees, fodder grasses, and so on for planting. (The cheapest solution for supplying seedlings is likely to be to establish a number of temporary nurseries rather than a few permanent ones.) Plantations will be needed for soil conservation purposes and to supply fuelwood, fodder, and commercial timber. Improved stoves could help to reduce fuelwood consumption; fencing should be set up around protected forest areas; and tracks and access roads should be upgraded. The “socially” oriented forestry projects being funded by the Bank in upland watershed regions of countries such as Nepal and the Philippines incorporate such features. The project in the Philippines is noteworthy in that it introduced tree farming as a cash crop.

There are three main constraints on reforestation: local people are often unwilling to support long-term forestry investments (incentives and subsidies are a feature of many forestry programs); sufficient land is frequently not available for tree planting because it is urgently needed for agriculture (implying that trees may have to be planted along farm boundaries or around homesteads); and it is difficult to protect young trees from grazing and fire. In Niger, for instance, trees planted as part of a village woodlot program were rapidly destroyed because of uncontrolled grazing, inadequate consultation with local people, and an inadequate incentive program.

Further investments are also needed in most catchment areas to prevent erosion and flooding. A variety of cheap, technical measures has been developed, including construction of levees, bench terracing, run-off disposal drains, retaining walls, and farm ponds (to encourage improved water harvesting practices and stall feeding).

Principal cost components for five Bank-financed agricultural or rural development projects located in upland watersheds

(In millions of U.S. dollars)

article image
Source: World Bank data.—Indicates no component.

Flood control measures include the construction of cheap silt traps and headwater dams. Silt traps usually consist of suitably sited retaining structures not more than five meters high. Headwater dams are usually concrete or rockfill structures sited where the reservoirs have sufficient capacity to control short duration flash floods. Capacities of at least 150,000 cubic meters to 200,000 cubic meters are required for mi-crowatersheds of about 1,000 hectares to provide significant flood control. A key issue is likely to be that of appropriate dam design. Dams for catchment areas of 1,000 hectares could be 15 meters or more in height and are major structures requiring specific civil engineering and hydrological expertise. In some situations, it is difficult to find such engineering capability in the project area.


Organization and management probably represent two of the most complex aspects of any watershed project—mainly because the number of institutions involved in an integrated program is so large and also because the socioeconomic issues of how to ensure farmers’ receptivity and involvement in the design and implementation of the project are so intricate. Many of the failures of past rehabilitation programs have been the result of technocratic approaches to planning that were implemented without adequate consultation with local people.

The Erosion Control Program in India, located in the Himalayan region, offers a good example of an attempt to ensure local participation in designing and implementing a watershed rehabilitation program. The program aims to establish, under existing panchayat legislation (panchayat being a village institution consisting of 9 wards), joint catchment area committees at the field level. Membership would include local panchayat leaders and representatives of the implementing technical departments. The functions of the joint committee would include coordination, communication, and assistance with the planning and implementation of project works. In practice this is proving difficult to implement because of the reluctance of officials working in separate technical agencies to coordinate their own programs and to delegate budgetary authority to the panchayat level. By contrast, in the Rasuwa-Nawakot Rural Development project in Nepal, good leadership and coordination of the technical aspects at the village administrative level have been a factor in the ready adoption of improved practices by the farmers. Significantly, response has been better in remoter areas where there has been more reliance on indigenous institutions and less central government direction.

Economic justification

Every catchment area is different, with individual determinants of costs, but some generalization is relevant. The cost structure of five projects already financed by the Bank illustrates the fact that agriculture, irrigation, farm improvement, and extension and training inputs often account for a significantly higher percentage of total project cost than do investments in soil conservation and reforestation (see the table). There is a tendency in much of the past literature dealing with watershed conservation to lay too great a stress on the role of soil conservation and forestry investments as an antidote to further ecological deterioration. An associated issue is the likely unit cost of effective catchment rehabilitation programs and their replicability. Very little research has been carried out in this area. Experience with five Bank projects suggests that initial unit cost per hectare of a catchment area under rehabilitation has ranged from $500 to around $1,000 per hectare (in 1981 U.S. dollars).

The benefits of watershed rehabilitation projects are both direct and indirect. They include general improvements in the lives of the people in the area, such as greater agricultural production and farm income; increased output of salable products from reforestation; and the restoration of existing forests. Rehabilitation indirectly increases the life span of irrigation and power works (mainly outside the watershed) by reducing silting. By reducing peak floods chiefly outside the treated watersheds, it can lead to diminished loss and damage to human life and property and to increased productivity of downstream agriculture. Finally, rehabilitation means conservation of the natural resources—mainly soil and water.

This article has concentrated on the importance of combining improved land practices with investment in physical measures to protect and rehabilitate deteriorating upland soils. Other factors outside the scope of this discussion also affect the success of these efforts. Strong central government commitment, for instance, must exist on the issues of how to tackle watershed rehabilitation and allocate budgetary resources to this end. Improved roads, water and electricity supplies, and the provision of social services such as schools and health clinics can all be key factors in improving the quality of life for farmers living in upland watershed areas and in encouraging a transition from ecologically destructive shifting agriculture to settlement in more stable communities. To ensure that the lessons gained from the experience with watershed management benefit current work, effective monitoring and evaluation of the effects on incomes as well as on physical reparations to the watershed and downstream are needed.

Perhaps the most important lesson learned from experience is that it may take up to ten years of more institutional strengthening, education training, research, and small-scale projects before the practices of improved agriculture, forestry, and soil conservation lead to a discernible improvement in watershed land use, in erosion, and in perennial stream flow. A classic example is the rural development and reforestation program carried out with Bank assistance in the upland watersheds of Korea, where it took a decade of patient demonstration and infrastructure development before spontaneous self-help planting of trees and other rural improvement works took off on a significant scale.

Sustainable development and resource management

In the Fairfield Osborn Memorial Lecture delivered by the Bank’s President A. W. Clausen in Washington, D.C. on November 12, 1981, Mr. Clausen spoke about the importance of resource management as a basis for sustainable development efforts throughout the world. Following are excerpts from his speech:

Environmental spoliation is an international cancer. It respects no boundaries. It erodes hard-won economic gains and thus the hopes of the poor…,

The Bank is well-known for the thoroughness of its project preparation. And, for a decade now, the Bank has required, as part of project evaluation, that every project it finances be reviewed by a special environmental unit.

As a matter of policy, we won’t finance a project that seriously compromises public health or safety; that causes severe or irreversible environmental deterioration; that displaces people without adequate provision for resettlement; or that has transnational environmental implications which are importantly negative.

Our environmental experts have reviewed more than 2,000 projects and programs in developing countries since 1970.

We’ve found that the cost of paying attention to environmental concerns has been much lower than many people expected when this procedure was first established. Nearly two thirds of the projects reviewed have raised no serious health or environmental questions, and I’m pleased to say that it has been possible to incorporate adequate protective measures in all the projects we have financed during the past decade.

The cost of these environmental and health measures has proved not to place an unacceptable burden on our borrowing countries. And we’ve learned, as have many private corporations, that the cost tends to be lower the earlier that environmental problems are identified and handled.

We’re convinced that it’s almost always less expensive to incorporate the environmental dimensions into project planning than to ignore them and pay the penalties at some future time.

It costs a lot less, for instance, to protect the forested watershed above a new dam than to deal with a silted reservoir after it’s been built. Similarly, the benefits of an irrigation project can be diminished if, for lack of proper planning, it leads to an increase of schistosomiasis.

We’re financing an increasing number of projects to rectify undesirable side effects of past development. We are working with the Mexican authorities, for instance, on various environmental problems, including air and water pollution in Mexico City. In Kenya, we’re financing a wildlife management project. We’re also involved in helping Nepal deal with its dramatic problems of deforestation and soil erosion.

In the Sahel region, we’ve financed a series of projects designed to halt the expansion of the desert and, more recently, to also reclaim desertified land by providing shelter belts of restored forest cover.

On the basis of the Bank’s experience in trying to anticipate environmental problems associated with development, we’ve published extensive checklists, handbooks, and guides for different types of projects—to help Bank staff and planners around the world think through all the identifiable effects of development projects.

Finance & Development, March 1982
Author: International Monetary Fund. External Relations Dept.