1. INTRODUCTION

In many parts of the subhumid and semiarid tropics, crop yields are declining on response to inputs such as fertilizers, and droughts and shortages of irrigation water are increasingly evident.

Sub-Saharan Africa and Asia pose two different challenges in raising food production to meet their food needs:

• Much of the agriculture in sub-Saharan Africa and Asia is not irrigated but rainfed, with the associated uncertainty as to the onset, reliability and amounts of rainfall each year. Limitations to the expansion of the existing cultivated area are increasingly those of uncertain rainfall and of the topographic and chemical hazards associated with taking marginal lands into cultivation. In sub-Saharan Africa a growth rate in food production of about 4 percent per year is needed to keep up with population growth to 2030, whereas over past years only about 2 percent rate per year has been achieved (FAO, 1996b).

• The main crop in much of Asia is rice, both irrigated and rainfed in wet valley lands and also grown on rainfed uplands. Three major problems limit the expansion of irrigated rice production into the future: (i) increasing competition for irrigation water from non-agricultural uses; (ii) even with the development of high-yielding rice varieties and hybrids yields are rising more slowly than previously and appear to be approaching a plateau of productive potential; (iii) only about 5 percent of the original total of potentially irrigable land might still remain by 2030 for expansion of irrigated rice (except in India) (FAO, 2000a).

The problem of lower annual increases in yield per hectare is a not confined to irrigated rice. Wheat and maize are also apparently reaching similar plateaus. For the three major staple crops (paddy rice, wheat and maize) the average yield increases between 1963 and 1983 were respectively, 2.1, 3.6 and 2.9 percent per year; but in the 10 years 1983-1993 the rates of increase had fallen respectively to 1.5, 2.1 and 2.5 percent per year (FAO, 1996b).

Some 40 percent of all food is produced under irrigation, from about 18 percent of the world's area of arable land plus permanent crops, with 60 percent produced under rainfed agriculture. As populations have risen this arable land and permanent crops area per head has been falling, except in the case of Europe (Table 1).

TABLE 1
Arable land and permanent crops area (1 000 ha) per 1 000 capita by region (FAOSTAT Database, 2002)

There is evidence that yields per hectare of some unfertilized rainfed crops are declining - as indicated in Table 2.

Fertilizer trials with local maize varieties have shown that responses to fertilizers have also been declining for years (Table 3).

TABLE 2
Decline in average yields of unfertilized maize in kg/ha - local/traditional varieties, Malawi (Douglas, 1994)

TABLE 3
Decline in response of local maize to fertilizers in Malawi (Malawi Government, 1957 - 1985)

In the State of Paraná in southern Brazil, from the time of clearing the land from native forest decades ago, yields of crops under conventional tillage fell between 5 to 15 percent in 10 years. This was accompanied by severe losses of soil, associated organic matter and applied nutrients, resulting in downstream flooding, sedimentation and other damage (Plates 1 and 2).

Records from Lesotho show that mean yields of major crops (generally without added fertilizers) declined between 1978 and 1986, related to a combination of adverse weather conditions, decline in soil conditions, and recurrent erosion and runoff. A three-year running mean of yields has been used, in order to smooth the effects of between-year weather variations (Table 4).

TABLE 4
Three-year running means of five major crops' yields (kg/ha), Lesotho
(after Lesotho Government, 1987)

At the same time, it has been widely observed that ongoing land degradation across topographic catchments has resulted in increasingly irregular streamflow, with more floods of muddy water in the rainy season and declining volume and duration of streamflow during the dry season (Plates 3, 4 and 5).

Human-induced agricultural land degradation is widespread in irrigated and rainfed land and in both tropical and temperate zones. Land degradation represents a challenge to the sustainability of farming systems in all regions, even those of low population densities (after FAO, 2001a). In rainfed lands, compaction, erosion and runoff are significant problems. On irrigated lands, problems are often those of poor drainage control, salinization and compaction leading to nutrient deficiencies (Figure 1).

FIGURE 1. Human-induced soil degradation (FAO, 1996b)

Source: Map 12 - Technical atlas (part. no 15) - World food summit 1996 - Volume 3: technical background documents 12-15

Such problems are not confined to tropical areas. Clear-felling of trees (Plate 6), grazing on very steep slopes (Plate 7) and the compacting effects of farm machinery (Plate 8) result in excess water runoff and erosion in temperate zones. Plate 9 shows a soil that has been compacted at about 8-10 cm depth by repeated disking, which has the effect of reducing its effective depth. Under native vegetation, this soil is deep and water absorptive. The difference in the growth of the soybean plants between the top left and upper right of the photo can be related both to the effects of erosion and to induced soil moisture shortage in the root zone.

PLATE 1. Clear water, from stable absorptive land. Cerrado, Brazil

[T.F. Shaxson]

Estimates of damage caused by compaction and erosion in the Eurasian region suggest that about 327 million hectares of land in Eurasia have been severely affected by wind and water erosion. Approximately 170 million ha of land have been affected by soil compaction. Conservative estimates calculate a production loss of 15 million tons of grain, two million tons of sugar beet and 500 000 tons of maize. Others calculate a 16-27 percent decrease in production as a result of compaction, with a loss of 50 million tons of grain production alone (Karabayev et al., 2000).

PLATE 2. This floodwater is turbid with eroded soil: it did not enter the ground first to emerge clean into the river. Caledon River, Lesotho

[T.F. Shaxson]

PLATE 3. The Namadzi stream, arising in a poorly managed cultivated hilly catchment, is empty of water in the dry season. Namadzi, Malawi

[T.F. Shaxson]

PLATE 4. From the same viewpoint upstream of the road-bridge, in the rainy season just after a storm the Namadzi stream is a raging, soil-filled torrent. Namadzi, Malawi

[T.F. Shaxson]

PLATE 5. These mature trees must have grown up along and above the riverbank. But destruction of soil porosity and permeability by bad husbandry in the catchment has resulted in heightened flood peaks, which have eroded away the riverbanks and left the trees marooned in the middle of the river bed. Mikolongwe, Malawi

[T.F. Shaxson]

PLATE 6. Clear-felling of planted forest and destruction of ground cover on steep slopes bares the soil and encourages runoff and erosion. Palmerston North, New Zealand

[T.F. Shaxson]

PLATE 7. Reduction of ground cover due to severe grazing by sheep causes landslips, with loss of plants, water and soil. Palmerston North, New Zealand

[T.F. Shaxson]

PLATE 8. Use of heavy farm machinery can compact soil and encourage runoff even where rainfall is never very intense. Abbotsbury, England

[T.F. Shaxson]

A study of the effects of soil compaction on wheat production in New Zealand showed that as the soil becomes increasingly degraded, costs rise as yields fall, squeezing margins of profit per hectare (Shepherd, 1992). This indicates a wider problem that, as yields begin to decline, farmers may apply more fertilizer, masking the underlying decline into unsustainable and uneconomic production. A survey of small resource-poor farmers in central Paraguay showed that as erosion and runoff continued, yields of cotton, tobacco, maize and other crops declined. As a result net farm incomes fell and farmers could no longer afford to buy equipment or inputs that might help to reverse the downward trend. This has led to farms being abandoned and desperate families migrating to the cities in search of income that farming could no longer provide (Sorrenson et al., 1998).

PLATE 9. Effects of compaction on root habitat of soybean

[T.F. Shaxson]

Potential sources of growth in overall output are: (1) expansion of arable land area (2) increases in cropping intensities to give greater harvested area; (3) growth in yield per hectare (FAO, 2000a). Considering the present problems with production noted above, these expectations are optimistic if areas of already-damaged soils continue to be managed in the same way as in the past. Expansion of arable land will be limited because almost all lands of good and moderate quality have been settled. Expansion of the cultivated area will be onto land with increasing difficulties and hazards, which will reflect negatively on the yields and economics in crop production of both rainfed and irrigated crops. Increases in cropping intensity with shorter (or even no) regular recuperative periods during which damaged soils can recover their soil fertility will result in continued and worsening land degradation. Rates of yield increase are tending to fall and upper limits of potential yield, at least of key grain crops, are apparently being approached where high inputs are used on the best soils.

A way forward

Intensification is the only option to increase usable biomass and available water per unit area of land. The challenge is to achieve intensification without causing more damage to soils and to the quantity and reliability of water supplies. Unfortunately, past attempts to intensify production using conventional methods have often resulted in damage to the soil.

The sustainability of agriculture depends not only on the soil continuing to be a fit place for crops, pastures and trees, but also on young people being enthused by farming to provide a continuity from one generation to the next, developing and carrying forward up-to-date knowledge and relevant skills in the husbandry of plants, animals and land.

How can farming return to being a way of life which is satisfying to many and that encourages them to remain in the rural areas? How can sufficient water be ensured, both for plant growth and for the regular flow of rivers, when much recent experience shows increasingly severe effects of drought on crop plants and a decline in the regularity and volume of river flow? How to achieve not only greater total output, but also better quality and improved food security over the year? How to produce a greater variety of foods to improve nutrition and health and reduce poverty by generating income? Conventional approaches seem to be inadequate for the task, despite the efforts of many to date.

Key parts of any strategy to address these issues include:

• Recognizing the soil as a key and living component of the environment. To date it has received far less attention in comparison with the above-ground components, which are more readily perceived.

• Encouraging the inherent capacities of life itself. Particularly the ability of bacteria, fungi, soil fauna and plants to continually colonize and modify habitats.

• Prolonging the usefulness of rainwater and organic matter. By recycling through different biotic processes as many times as possible.

In 1971 D.A. Poole wrote:

"We must begin to regard our individual disciplines as part of a whole - an ecological whole - as one of the several moving parts that, depending on how applied, either catalyses or obstructs the working of the whole. We must recognize and promote the ecology of our individual disciplines, none of which can afford to act alone. The public cares less about the technical aspects of soil conservation, forestry, wildlife or any other discipline than it does about their environmental effects. People will support - with money and voices - those professionals whose programs truly assure them of environmental improvement. And they will resist - as they are doing so strongly today - those programs based more on textbook philosophy than on environmental acceptability."

This is as true now as when it was written more than 30 years ago.