Cover Image
close this bookSoil Degradation - A Threat to Developing-Country Food Security by 2020? 2020 Vision for Food, Agriculture, and the Environment Discussion Paper 27 (IFPRI, 1999, 70 p.)
View the document(introduction...)
View the documentForeword
View the documentAcknowledgments
View the documentAcronyms and Abbreviations
View the document1. Introduction
View the document2. Evaluating the Impact of Soil Degradation on Food Security
View the document3. Past and Present Effects of Soil Degradation
View the document4. Future Effects of Soil Degradation and Threats to Developing-Country Food Security
View the document5. Policy and Research Priorities
View the documentAppendix - Types of Soil Degradation: A Glossary19
View the documentBibliography
View the documentRecent Food, Agriculture, and the Environment Discussion Papers

3. Past and Present Effects of Soil Degradation

The past half century has been a period of unprecedented agricultural change in developing countries in response to large population increases, integration of rural areas into national and international agricultural and other markets, new technologies, and infrastructure development. Major increases in aggregate agricultural production in this period have been associated with different kinds of soil degradation. This chapter reviews the available evidence on the economic effects of degradation - at the global level and for three developing regions: South and Southeast Asia, Sub-Saharan Africa, and Mexico and Central America - and assesses the importance of soil degradation to policy concerns.

Land Use and Management in Developing Countries Since the 1950s

It is useful to examine the overall patterns of agricultural change in developing countries first. Rural population increase, expansion of cultivated area, and intensification of production have all affected soil quality. Although the rural population growth rate in developing countries declined from 2.2 percent in 1960-65 to 1 percent in 1990-95, the absolute number of rural dwellers grew almost 40 percent, from 2.0 to 2.8 billion over the same period (UN 1995). Rural population was fairly stable in Latin America, but it increased 37 percent in Asia (outside Japan) and a remarkable 68 percent in Africa. Total growth rates in 1970-88 for agricultural production in developing countries (4.1 percent per year in East Asia; 3.1 percent in South Asia, Near East, and North Africa; 2.6 percent in Latin America and the Caribbean; and 1.8 percent in Sub-Saharan Africa) have rivaled or surpassed growth rates in the industrialized countries (1.2 percent per year in the same period), though not on a per capita basis.

This growth came in part from extensive clearing of new agricultural lands. Yet even with this expansion, arable land per capita declined from just under 0.5 hectare in 1950 to just under 0.3 hectare in 1990 (FAO 1993). Yield increases on land already in production thus contributed far more to total production. For example, more than 90 percent of the growth in developing-country cereal production between 1961 and 1990 came from yield growth (World Bank 1992b).

It should not be surprising that agricultural expansion and yield growth on such a scale would be associated with some degradation of soil resources. Yet the patterns of degradation vary in the different pathways leading to agricultural intensification and reflect the level of resource endowments in each pathway (Scherr 1997b). The five main pathways are summarized below (also see Table 3).

Irrigated Lands

Area under irrigation in developing countries in 1995 totaled 190 million hectares - an increase of 60 percent since the early 1960s (Pinstrup-Andersen, Pandya-Lorch, and Rosegrant 1997). Irrigated land now accounts for about 7.5 percent of all arable and grazing lands (Nelson et al. 1997). In low- and middle-income countries, the proportion averages 20 percent of arable and perennial cropland, reaching 34 percent in East and South Asia. Only in Africa has irrigation, at only 4 percent of arable land, been unimportant (World Bank 1997). In 12 developing countries, including populous states like Egypt, China, Iran, Iraq, North Korea, and South Korea, and Pakistan, more than 40 percent of all arable land was under irrigation in 1994.

Table 3 - Major pathways of change in agricultural land use in developing countries and associated degradation problems

Land type

Main changes

On-site soil degradation

Other resource degradation

Irrigated lands

60 percent increase in irrigated area, 1961-90; increased multiple cropping

* Salinization and waterlogging
* Nutrient constraints under multiple cropping
* Biological degradation (agrochemicals)

* Nutrient pollution in ground/ surface water
* Pesticide pollution
* Water-borne disease
* Water conflicts
* Pesticide pollution

High-quality rainfed lands

Transition from short fallow to continuous cropping, HYVs, mechanization

* Nutrient depletion
* Soil compaction and physical degradation from overcultivation, machinery
* Acidification
* Removal of natural vegetation, perennials
* Soil erosion
* Biological degradation (agrochemicals)

* Deforestation of commons

Densely populated marginal lands

Transition from long to short fallows or continuous cropping; Cropping in new landscape niches

* Soil erosion
* Soil fertility depletion
* Removal of natural vegetation, perennials from landscape
* Soil compaction, physical degradation from overcultivation
* Acidification

* Loss of biodiversity
* Watershed degradation

Extensively managed marginal lands

Immigration and land-clearing for low input agriculture

* Soil erosion from land-clearing
* Soil erosion from crop/livestock production
* Soil nutrient depletion
* Weed infestation
* Biological degradation from topsoil removal

* Deforestation
* Loss of biodiversity
* Watershed degradation

Urban and peri-urban agricultural lands

Rapid urbanization; diversification of urban food markets; rise in urban provery

* Soil erosion from poort agricultural practices
* Soil contamination from urban polutants
* Overgrazing and compaction

* Water pollution
* Air pollution
* Human disease vectors

Irrigation brought myriad changes in land husbandry practices, increased multiple cropping, new purchased inputs (hybrid varieties, chemical inputs), land leveling, and new forms of local organization. While yields and production have increased markedly, some soils have also degraded, particularly due to poor management of water causing salinization and waterlogging, but also more subtle nutrient management problems which have slowed down yield increases in recent years. High fertilizer and pesticide applications have often contaminated water supplies.

High-Quality Rainfed Lands

“High-quality” rainfed lands are located mainly in areas with naturally deep, fertile, and less-weathered soils: temperate zones (for example, Argentina, southern Brazil, Chile, South Africa); volcanic regions in the tropical highlands (for example, the East African highlands, Java); and tropical regions with vertisols and alfisols (for example, South Asia, West African savannahs). These lands account for about 605 million hectares, or 23 percent of arable and grazing lands in developing countries, and, with irrigated lands, for about 35 percent of the rural population (Nelson et al. 1997).

In these prime rainfed lands, farmers have greatly increased cropping intensity, even where permanent agriculture had already been the norm. The Green Revolution - which brought increased use of hybrids, increased chemical use, mechanization, and a trend toward monocropping - also played a pivotal role in these areas. In some cases, inappropriate use of machinery has led to soil compaction; poor vegetation management has exposed soils to erosion; and substitution of organic inputs with chemical fertilizers has led to declining organic matter and acidification of vulnerable soils.

Densely Populated Marginal Lands

Most agricultural land area in developing countries falls outside the category of high-quality, relatively resilient irrigated and rainfed land. These “marginal” lands, which have lower-quality or degradation-prone soils, and are subject to harsher and more variable climates, account for about 69 percent of arable and grazing lands and 65 percent of rural population6 (Nelson et al. 1997). They are associated with two contrasting pathways, distinguished by rural population density.

6“Densely populated” is used here as a relative term indicating population pressure on arable lands from smallholder farming using short-fallow or permanent cultivation systems. Actual densities may be as low as 30 per square kilometer in areas with lots of unusable and low-quality land, or as high as 1,000 per square kilometer. The term “marginal lands” is conventionally used. However, readers should remember that lands are typically defined as “marginal” for the purpose of plow-based grain cultivation. These lands may be superior production sites for other types of products or production systems.

Large areas of long-settled marginal lands are now under intensive crop production as a result of high and rapidly growing rural population and development of agricultural markets. Cultivation has spread into landscape niches, such as steep slopes, with poorer and more vulnerable soils. Human settlements compete for use of agricultural lands. External inputs are often less available, more costly, and less profitable than in the high-potential areas, and intensive farming practices (typically borrowed from high-potential areas) are often not adapted to marginal environmental conditions. Overexploitation for subsistence and commercial uses has led to loss of vegetation for soil cover. Soil erosion and nutrient depletion are common, though there is evidence that intensification has sometimes led to greater use of soil-protecting practices (Tiffen, Mortimore, and Gichuki 1994; Clay, Reardon, and Kangasniemi 1998; Turner, Hyden, and Kates 1993; Reij, Scoones, and Toulmin 1996; and Templeton and Scherr 1997).

Extensive Agriculture in Marginal Lands

Other marginal lands - commonly considered the “agricultural frontier” - have much lower populations. As land is relatively abundant, it is managed using labor-extensive practices such as long crop fallows or extensive grazing.

In the 1950s, it was estimated that around 200 million people on 14 million square miles (10 percent of the world’s population on 30 percent of exploitable soils) were practicing shifting cultivation (Nye and Greenland 1960). Between the early 1960s and the mid-1990s, land area under annual crops increased by 19 million hectares in Asia, 28 million in South America, and 31 million in Africa. Area under permanent pastures expanded even more in aggregate terms, while total forest and woodland area declined, especially in Asia and South America (FAO 1995). It is estimated that 200 million landless people have migrated to tropical forests since the 1960s, and that as many as 500 million people - most of them poor - now use shifting cultivation systems (ASB undated).

Most of the land claimed from these frontier areas has lower intrinsic soil quality or poses higher production risks due to factors such as steep slopes and very high or very low rainfall. Population densities in these areas are relatively low, infrastructure limited, and market development weak. Soils are degraded by the land-clearing process itself, by decreasing fallow periods that deplete nutrients, and by widespread burning to control weeds and pests and provide ash for plant nutrition. Large areas have been abandoned due to nutrient and organic matter depletion and weeds. In Southeast Asia, Im-perata grass now covers 40 million hectares; in the Amazon there are an estimated 20 million hectares of degraded pastures (ASB undated). There are few economic incentives for investing in land improvement, because land is still relatively abundant, of low market value, and often available without secure land rights.

Urban and Peri-Urban Agricultural Land

During the 1980s, the importance of urban agriculture accelerated dramatically throughout the world. The Urban Agriculture Network has estimated that by the early 1990s, approximately 800 million people globally were actively engaged in urban agriculture, of whom 200 million were farmers producing for sale on the market (many part time). Evidence from eight African and three Asian countries showed 33-80 percent of urban families engaged in food, horticultural, or livestock production. Low-income urban residents, who would otherwise spend a very high proportion of their income on food, typically engage in agriculture to increase their food security, income levels, and sometimes the nutritional quality of their food. Middle- and high-income urban farmers grow food mainly to improve diet quality or supplement incomes with high-value crops (Cheema et al. 1996, Tables 2.1 and 3.1).

Contrary to popular belief, a high proportion of urban land is available for agriculture.7 In Beria, Mozambique, 88 percent of the city’s “green spaces” are used for family agriculture. Large areas of many cities are so used: Beijing (28 percent of the city); Zaria, Nigeria (66 percent); Hong Kong (10 percent); Bangkok (60 percent of the metropolitan area); and San Jose, Costa Rica (60 percent of the metropolitan area). Farmers may borrow, rent, or squat on the land they farm.

7Horticulture takes place in homesites, parks, rights-of-way, rooftops, containers, wetlands, and greenhouses. Livestock are produced in zero-grazing systems, rights-of-way, hillsides, coops, peri-urban areas, and open spaces. Agroforestry is practiced using street trees, homesites, steep slopes, within vineyards, greenbelts, wetlands, orchards, forest parks, and hedgerows. Aquaculture is practiced in ponds, streams, cages, estuaries, sewage tanks, lagoons, and wetlands. Food crops are grown in homesites, vacant building lots, rights-of-way for electric lines, schoolyards, church yards, and the unbuilt land around factories, ports, airports, and hospitals (Cheema et al. 1996, Table 5.1).

Global Effects of Soil Degradation

The land surface of the earth totals 13.0 billion hectares, of which 1.5 billion are unused wasteland and 2.8 billion are unused but largely inaccessible (Oldeman 1994). Of the 8.7 billion hectares under use, most is suitable only for forest, woodland, grassland, or permanent vegetation. Only 3.2 billion hectares are potentially arable. About half of this potentially arable land is currently cropped and 41 percent is considered moderately to highly productive (Table 4).

Table 4 - Global supply and use of land

Type of land

Area

(billion hectares)

Total ice-free land area in the world

13.4

Total land area without water bodies

13.0

Land used

8.7

Potentially arable land

3.2

Moderately to highly productive

1.3

Low productive land

1.9

Current use of potentially arable land

3.2

Cropland

1.5

Permanent pasture, forest, and woodland

1.7

Source: Buringh and Dudal 1987. Data for total land area without water bodies and land used are from Oldeman 1994.

Note: Potentially arable land is defined as land that can be cultivated or maintained in permanent pasture or both.

The 16 studies reviewed below assess the global extent, rate, and effects of soil degradation. Their data suggest that soil quality on three-quarters of the world’s agricultural land has been relatively stable since the middle of the twentieth century. On the rest, however, soil degradation is widespread and the pace of degradation has accelerated in the past 50 years. Productivity has declined substantially on approximately 16 percent of agricultural land in developing countries, especially on cropland in Africa and Central America, pasture in Africa, and forests in Central America. Large land areas of 5 to 8 million hectares have gone out of production each year. Increased land in production and under irrigation, increased productivity through new varieties and inputs, and improved marketing systems have compensated for some productivity losses caused by degradation. But in the specific regions, countries, and subregions where it is widespread, the economic and welfare effects of soil degradation pose pressing policy challenges.

Historical Soil Degradation

There is historical evidence of large-scale soil degradation in many parts of the world in the past 5,000 years (Hillel 1991; Hyams 1952). UNEP (1986) calculated that 2 billion hectares of land that was once biologically productive has been irreversibly degraded in the past 1,000 years. Rozanov, Targulian, and Orlov’s (1990) analysis of global changes in the humusphere found that there has been a loss of humus at a rate of 25.3 million tons per year on average ever since agriculture began 10,000 years ago. This loss accelerated to 300 million tons per year in the past 300 years and 760 million tons per year in the past 50 years. Nearly 16 percent of the original stock of organic soil carbon may have been lost. Within the past 300 years, 100 million hectares of irrigated land alone apparently have been destroyed and another 110 million hectares have come to suffer from diminished productivity due to secondary salinization. The amount of land thus affected is nearly equivalent to the 220 million hectares of global irrigated area in 1984. Rozanov, Targulian, and Orlov conclude that more productive land may have been irreversibly lost in the past 10,000 years than is currently under agricultural production.

Extent of Degradation

During the past half century, soil degradation concerns have focused principally on soil erosion. The earliest reports typically were cast in terms of tons of soil lost,8 a measure difficult to use for policy assessment. The Global Assessment of Soil Degradation (GLASOD), based on a formal survey of regional experts, was the first worldwide comparative analysis to focus specifically on soil degradation (Oldeman 1994). GLASOD was designed to provide continental estimates of the extent and severity of degradation from World War II to 1990.9 The study concluded that 1.97 billion hectares - 23 percent of globally used land - had been degraded.

8For example, Judson (1968) estimated that 14.7 billion tons of soil were lost annually due to human-induced soil erosion, in addition to 9.3 billion tons due to natural processes. Brown (1984) extrapolated from data for the United States, USSR, China, and India to conclude that one-third to one-half of global cropland had “excess” soil loss from erosion beyond a sustainable level.

9The objective of GLASOD was to create awareness about the status of soil degradation. Over 250 soil and environmental scientists cooperated in preparing 21 regional maps of human-induced soil degradation, using a common methodology. Following delineation of physiographic units with homogeneous topography, climate, soils, vegetation, and land use, each unit was evaluated for its degree, relative extent, and recent past rate of degradation, as well as for the forms of human intervention causing degradation. Types of degradation were ranked in importance. Map segments were compiled and reduced to the final 1:10 million scale of the GLASOD map. The map units were digitized and linked to a GLASOD database to calculate the areal extent of degradation. Since the maps rely on expert evaluation, they may reflect unsubstantiated biases and assumptions.

Thirty-eight percent of all agricultural land had become degraded, along with 21 percent of permanent pasture and 18 percent of forests and woodland (Table 5). Nine percent of all cropland, pasture, and woodland was lightly degraded in 1990; 10 percent was moderately degraded, implying a large decline in productivity; and 4 percent was strongly degraded, implying a virtual loss in productive potential. Water erosion caused the most degradation, followed by wind erosion, soil nutrient depletion, and salinization (Tables 6 and 7). Overgrazing was the leading proximate cause, followed by deforestation and agricultural activity.

Of all degraded soils, 58 percent were in drylands and 42 percent in humid areas.10 For the tropics alone, 915 million hectares had been degraded by water erosion, 474 million by wind erosion, 239 million by chemical degradation and 50 million by physical degradation (Lal 1994, using GLASOD data). Estimates show that nearly 20 percent of 1.1 billion hectares of global dryland soils have been degraded. This is well below estimates from Dregne and Chou’s (1992) comprehensive review of literature on dryland degradation (including degradation of soil as well as vegetation and nonagricultural soil functions). They found that more than 70 percent of drylands in Africa, Asia, and South America are degraded - 30 percent of irrigated drylands, 47 percent of rainfed drylands, and 73 percent of rangelands.

10“Dryland” was defined as climatic regions with annual precipitation/evapo-transpiration ratio of £ 0.65; “humid” are those regions with less than 0.65.

Table 5 - Global estimates of soil degradation, by region and land use


Agricultural land

Permanent pasture

Forests and woodland

All used land

Region

Total

Degraded

Percent

Total

Degraded

Percent

Total

Degraded

Percent

Total

Degraded

Percent

Seriously degraded

Percent


(million hectares)


(million hectares)


(million hectares)


(million hectares)


(million hectares)


Africa

187

121

65

793

243

31

683

130

19

1,663

494

30

321

19

Asia

536

206

38

978

197

20

1,273

344

27

2,787

747

27

453

16

South America

142

64

45

478

68

14

896

112

13

1,516

244

16

139

9

Central America

38

28

74

94

10

11

66

25

38

198

63

32

61

31

North America

236

63

26

274

29

11

621

4

1

1,131

96

9

79

7

Europe

287

72

25

156

54

35

353

92

26

796

218

27

158

20

Oceania

49

8

16

439

84

19

156

12

8

644

104

17

6

1

World

1,475

562

38

3,212

685

21

4,048

719

18

8,735

1,966

23

1,216

14

Sources: For all totals, FAO 1990, and for others, Oldeman, Hakkeling, and Sombroek 1991.

Notes: The last two columns refer only to land that is moderately, strongly, or extremely degraded. In the GLASOD study “lightly degraded soil” is defined as having somewhat reduced agricultural suitability, but is suitable in local farming systems. Original biotic functions are still largely intact, and restoration to full productivity is possible through modifications in farm management. “Moderately degraded soil” is soil that offers greatly reduced productivity, but is still suitable for use in local farming systems. Major improvements are needed that are typically beyond the means of local farmers; the original biotic functions are partially destroyed. In “strongly degraded soil” productivity is virtually lost and soil is not suitable for use in local farming systems; the original biotic functions are largely destroyed. Major investments and/or engineering works would be needed to restore land to full productivity. “Extremely degraded soil” is defined as a “human-induced wasteland,” unreclaimable, beyond restoration, and with biotic functions that are fully destroyed. Data for permanent pasture and forests and woodland include arable and nonarable land.

Table 6 - Global extent of chemical and physical soil degradation, by region

Chemically degraded area

Physically degraded area



Region

Loss of nutrients

Salinization

Pollution

Acidification

Compaction sealing, and crusting

Water logging

Loss of organic matter

Total degraded land

Total degraded land as per cent of total land used

(million hectares)

Africa

45

15

+

1

18

1

-

81

4.8

Asia

15

53

2

4

10

+

2

86

3.0

South America

68

2

-

-

4

4

-

78

5.1

Central America

4

2

+

-

+

5

-

12

6.0

North America

-

+

+

+

1

-

-

1

+

Europe

3

4

19

+

33

1

2

62

7.7

Australia

+

1

-

-

2

-

-

3

World

136

77

21

6

68

11

4

323

3.7

Source: Oldeman, Hakkeling, and Sombroek 1991.

Note: Degradation figures include data for slightly, moderately, strongly, and extremely degraded lands. Plus sign means negligible; minus sign means none reported.

Table 7 - Global extent of soil degradation due to erosion, by region


Area eroded by water erosion

Area eroded by wind erosion




Region

Light

Moderate

Strong and extreme

Total

Light

Moderate

Strong and extreme

Total

Total area eroded

Total area seriously eroded

Total area seriously eroded as a percent of total land used

(million hectares)

Africa

58

67

102

227

88

89

9

186

413

267

16

Asia

124

242

73

441

132

75

15

222

663

405

15

South America

46

65

12

123

26

16

...

42

165

93

6

Central America

1

22

23

46

246

4

1

5

51

50

25

North America

14

46

...

60

3

31

1

35

95

78

7

Europe

21

81

12

114

3

38

1

42

156

132

17

Oceania

79

3

222

83

16

...

27

16

99

3

3

World

343

526

223 1,

094

269

254

26

548

1,642

1,029

12

Source: Oldeman, Hakkeling, and Sombroek 1991.

Notes: The last two columns refer only to land that is moderately, strongly, or extremely degraded. In the GLASOD study “lightly degraded soil” is defined as having somewhat reduced agricultural suitability, but is suitable in local farming systems. Original biotic functions are still largely intact, and restoration to full productivity is possible through modifications in farm management. “Moderately degraded soil” is soil that offers greatly reduced productivity, but is still suitable for use in local farming systems. Major improvements are needed that are typically beyond the means of local farmers; the original biotic functions are partially destroyed. In “strongly degraded soil” productivity is virtually lost and not suitable for use in local farming systems; the original biotic functions are largely destroyed. Major investments and/or engineering works would be needed to restore land to full productivity. “Extremely degraded soil” is defined as a “human-induced wasteland,” unreclaimable, beyond restoration, and with biotic functions that are fully destroyed. Ellipses indicate negligible amounts.

Effects on Productivity

Data on the effects of degradation on global productivity are necessarily very rough. Pimentel, Alien, and Beers (1993) estimate, based on available secondary data, that global production is 15-30 percent lower as a result of all the various effects of soil erosion. Buringh and Dudal’s (1987) estimates are even higher. Using an International Institute for Advanced Systems Analysis (IIASA) model that assumes no soil conservation, they predicted that just between 1984 and 2000, 22 percent of the more productive crop, pasture, and forest land - including 14 percent of the most productive soils - would be degraded. Erosion-induced soil nutrient depletion would result in a 29 percent decline in rainfed crop production and a 19 percent loss in total potential production (South America would lose 10 percent, Africa 17 percent, Southwest Asia 20 percent, Central America 30 percent, and Southeast Asia 36 percent).

Other figures for the effects on global productivity, based more on empirical evidence, are much lower. Dregne and Chou (1992) estimate that more than a third of irrigated land in Asia and more than half of rainfed land in Africa and Asia had experienced a 10 percent loss in productive potential, while 8 percent of irrigated and 10 percent of rain-fed land in Asia had experienced at least a 25 percent loss in potential productivity, with lower incidence elsewhere. They estimated that over half the rangelands had experienced more than 50 percent loss in potential productivity. Using GLASOD data, Crosson (1995b) estimated an aggregate global loss of 11.9-13.4 percent of agricultural supply, assuming a 15 percent, 35 percent, and 75 percent yield decline, respectively, for light, moderate, and strongly degraded cropland soils, and a 5 percent, 18 percent, and 50 percent decline for pasture soils. Global production would be 12-13 percent higher if the 15 percent of strongly and extremely degraded lands were restored to full productivity. Oldeman (1998) used Crosson’s coefficients to calculate that global cropland production was 12.7 percent lower and pasture production 3.8 percent lower than they would have been without degradation, for a total agricultural loss of 4.8 percent. With higher estimates of pasture yield decline, global loss increases to 8.9 percent (Table 8).

Economic Effects

While environmental economists have used resource valuation techniques to estimate the global value of other natural resources, no such studies are available for soil (Costanza et al. 1997). Early crude estimates of the annual cost of soil erosion hovered around U.S.$26 billion, about half the cost borne by developing countries (UNEP 1980). A decade later, Dregne and Chou (1992) proposed $28 billion per year as the cost of dryland degradation. Pimentel, Allen, and Beers (1993) valued the plant nutrients lost annually just through sediment loss and nitrogen in water runoff at $5 billion, or 0.4 percent of the annual global value added in agriculture.

Effects on Consumption by Poor Farmers

There has been no global mapping of the relationship between poverty and soil quality or soil degradation. However, a number of factors suggest that soil degradation affects the rural poor in a particularly negative way. Studies in Asia and West Africa in the 1980s (reviewed in Malik 1998) found that the rural poor were more dependent on agriculture than the nonpoor. The poor depended more on annual crops, which typically degrade soils more than other crops. They also relied more on common property lands, which tend to suffer greater degradation than privately managed land. When the principal assets of the poor comprise low-productivity or degrading lands, and their ability to seek more remunerative livelihood options is restricted by economic, political, or social conditions, they may fall into a poverty “trap,” in which they lack sufficient assets to undertake the land husbandry and investment necessary to maintain or increase productivity (Malik 1998). The poor tend to be “pushed” to marginal lands by political forces, expulsion of squatters from higher-quality lands during modernization, or the inability to compete for higher-quality land. Because the poor use fewer inputs, they rely more on intrinsic soil quality.

Poverty may also exacerbate degradation when poor people can meet subsistence food, feed, and fuel needs only through overexploitation of natural vegetation and consumption of organic residues from fanning and livestock-keeping that would otherwise help replenish the soil. The poor play a significant role in expansion of farming into marginal lands, especially when nonfarm employment opportunities decline. Thus a negatively reinforcing relation between poverty and soil degradation can develop. There is also evidence, however, that poor farmers may respond effectively to soil degradation, both to reverse degradation and to cushion its effects on their livelihoods (Scherr 1999).11

11Wealthier farmers, agricultural investors, and multinational corporations typically control more total land area than the poor, and play a prominent role in large-scale clearing of natural vegetation, overuse of agrochemicals, large-scale degradation of grazing lands, and overexploitation of soils for commercial production.

Table 8 - Average cumulative loss of productivity during the post-Second World War period as a result of human-induced soil degradation, worldwide and by region

Region

Cropland

Pasture land

Crops and pastures (low estimates of impact)

Crops and pastures (high estimates of impact)

(percent)

Africa

25.0

6.6

8.1

14.2

Asia

12.8

3.6

4.7

8.9

South America

13.9

2.2

4.1

6.7

Central America

36.8

3.3

8.7

14.5

North America

8.8

1.8

3.0

5.8

Europe

7.9

5.6

4.6

9.0

Oceania

3.2

1.1

1.2

3.2

World

12.7

3.8

4.8

8.9

Source: Oldeman 1998, 4, Table 1.

Notes: These figures were calculated by multiplying the area by a coefficient of yield loss for each soil degradation category. In the case of cropland, the coefficients were 15 percent loss for “light” soil degradation, 35 percent for “moderate,” 75 percent for “strong,” and 100 percent for “extreme” degradation. In the ease of pasture land, the corresponding coefficients were 5 percent for light, 18 percent moderate, and 50 percent strong. For combined crop and pasture land, two different sets of coefficients were used: 5 percent for light, 18 percent for moderate, 50 percent for strong, for pastures; and 15, 35, and 75 percent, respectively, for cropland.

Agricultural Land Loss

Estimates of the annual rate of loss of agricultural land due to degradation range from 5 to 12 million hectares, or about 0.3 to 1.0 percent of the world’s arable land. On the higher end are Lal and Stewart (1990), who estimated that 12 million hectares were being destroyed and abandoned annually. UNEP (1986) estimated that 6 million hectares were being lost each year through desertification processes. GLASOD calculated that since the mid-1940s 5-6 million hectares per year had been permanently lost to agriculture through human-induced soil degradation, a rate (0.3-0.5 percent of the world’s arable land area) comparable to earlier estimates by Dudal (1982). Rozanov, Targulian, and Orlov (1990) estimated that 6-7 million hectares per year are being irreversibly lost.

Effects of Soil Degradation in South and Southeast Asia

Regional studies and studies for Bangladesh, China, India, Indonesia, and Pakistan show that soil degradation - mainly from nutrient depletion and salinization - has a significant effect on national agricultural supply in South and Southeast Asia. Estimates of the total annual economic loss from soil degradation range from under 1 to 7 percent of agricultural gross domestic product (AGDP). Given that more than half of all land is not affected by degradation, the economic effects in the degrading areas would appear to be quite serious.

Extent of Degradation

The extent of soil degradation in Asia was evaluated in five major studies in the 1980s and 1990s. A literature review by FAO (1986) found that 31 percent of the total land area in 13 Asian-Pacific countries was degraded, with the highest incidence (³30 percent) in China, India, Laos, Thailand, and Viet Nam, and the lowest incidence (<10 percent) in Tonga, Bangladesh, and Myanmar (Table 9). The main hazard was soil nutrient depletion, though waterlogging and salinity also posed significant problems.

Dregne and Chou’s (1992) literature review of dryland degradation concluded that 71 percent of Asian drylands are degraded, and 39 percent “severely” so. They estimated that degradation affected 35 percent of irrigated lands, 56 percent of dry rainfed lands, and 76 percent of rangelands.

Young (1993) and national soil experts in eight South Asian countries12 revised the continental GLASOD figures, incorporating the best available national data. GLASOD data indicated that a total of 43 percent of the agricultural land in these eight countries was affected by some type of degradation. Most nondegraded land was either in rainfed lands of the humid zone or irrigated alluvial areas in both humid and dry zones. The revised figures showed that 25 percent of the region’s agricultural land had been degraded by water erosion, of which 60 percent was moderately or strongly degraded (that is, costly or nearly impossible to reverse). Another 18 percent had been degraded by wind erosion (of which 77 percent was moderately or strongly degraded), and 13 percent by soil fertility decline (less than 10 percent was moderate or severe). Two percent was degraded by waterlogging (three quarters was moderate or severe), 9 percent by salinization (72 percent moderate or severe), and 6 percent by lowering of the water table (40 percent was moderate).

12Afghanistan, Bangladesh, Bhutan, India, Iran, Nepal, Pakistan, and Sri Lanka.

Using a more detailed and nationally representative GLASOD-type methodology, however, the Assessment of Human-Induced Soil Degradation in South and Southeast Asia (ASSOD), found 20 times greater decline in soil fertility and organic matter, triple the extent of salinization, and nearly 100 times the extent of waterlogging than in the GLASOD study (van Lynden and Oldeman 1997; also see Table 10). Agricultural activity had led to degradation on 27 percent of all land and deforestation on 11 percent; overgrazing played a minor role.

Table 9 - Degraded cropland in selected countries in the Asian-Pacific region

Country

Total land area

Arable and permanent cropland

Arable and permanent cropland as a percent of total land area

Total degraded land

Total degraded land as a percent of total land area

(thousand hectares)

(thousand hectares)

Bangladesh

13,017

9,292

71

989

7.4

China

932,641

96,115

10

280,000

30.0

India

297,319

168,990

57

148,100

49.8

Indonesia

181,157

21,260

12

43,000

24.0

Laos

23,080

901

4

8,100

35.0

Myanmar

65,754

10,034

15

210

3.2

Pakistan

77,088

20,730

27

15,500

17.3

Philippines

29,817

7,970

27

5,000

16.8

Samoa, Western

283

122

43

32

11.3

Sri Lanka

6,463

1,901

29

700

10.8

Thailand

51,089

22,126

43

17,200

33.7

Tonga

72

48

67

3

4.5

Viet Nam

32,549

6,600

20

15,900

48.9

Asian-Pacific region

1,710,329

366,089

21

534,734

31.3

Source: Watershed Management in Asia and the Pacific: Needs and Opportunities for Action, Technical Report FO:RAS/85/017, FAO, Rome, 1986 (cited in FAO 1992). Arable and permanent cropland data are for 1989.

ASSOD collaborators collected data on type of farm management for nearly half of the degraded land. They found little association between land management and degradation: 38 percent of degraded lands were under a high level of management, 36 percent under medium management, and 25 percent under low management (defined as “traditional” systems existing for more than 25 years). In recent years, however, degradation increased more often under low and medium management.

Two unique historical data sets based on soil surveys dating from the 1940s recently became available for China and Indonesia. These data suggest that nutrient depletion may not have been as severe since the 1940s as commonly assumed (Lin-dert forthcoming a and b; Lindert, Lu, and Wanli 1996a and 1996b). The researchers found declines in organic matter and nitrogen in Java and North China, with a rise in total phosphorus and potassium in Java. There was little overall change in nutrient status in South China over the period. Cropping intensity correlated with nutrient depletion; erosion appeared to have had a minor effect on soil degradation.

Agricultural Supply

Dregne (1992) concluded in a literature review that well-confirmed instances existed of permanent soil productivity loss of at least 20 percent due to human-induced water erosion in significant areas of China, India, Iran, Israel, Jordan, Lebanon, Nepal, and Pakistan. Strong presumptive evidence of such effects existed in Indonesia, the Philippines, Syria, Thailand, and the Caucasus region. He concluded that wind erosion, while widespread in dry areas, had not had much effect on long-term soil productivity.

Using the GLASOD data, Oldeman (1998) calculated that since World War II soil degradation in Asia had led to a cumulative loss of productivity in cropland of 12.8 percent, and 4.7-8.9 percent loss in cropland and pastures together (Table 8). ASSOD data showed major, irreversible productivity loss13 (that is, strong or extreme degradation) only in small areas. However, moderate degradation was found on a tenth of all lands, and serious fertility decline or salinization on more than 15 percent of arable land (van Lynden and Oldeman 1997, Table 4.5). In South and Southeast Asia, 11 percent of agricultural land had badly degraded soils. In terms of proportion of land area, degradation was reported to be most serious (more than 20 percent of land badly degraded) in India, Pakistan, the Philippines, and Thailand.

13Changes in productivity were expressed in relative terms, that is, the current average productivity compared to the average productivity in the nondegraded situation, assuming a given input use. For instance, if yield averaged 2 tons of rice per hectare previously, but only 1.5 tons at present, despite high inputs (and all other factors being equal), “strong” degradation was present (van Lynden and Oldeman 1997, 8-9). “Moderate” degradation indicates either that no change in production had occurred despite high management levels, that a small decrease had occurred despite medium management, or that a large decrease had occurred under low management.

Studies in China found that degradation had reduced grain yields. One calculated that for the period 1983-89, total grain production would have been 60 percent higher in the absence of a deteriorating environment. Increased floods and drought caused 30 percent of this yield loss, erosion 19 percent, salinity 0.2 percent, and increased multiple-cropping intensity 11 percent. Environmental degradation during the period cost the country as much as 5.6 million metric tons of grain per year - a figure equivalent to nearly 30 percent of China’s yearly grain imports in the early 1990s. Without the effects of a deteriorating environment, mostly erosion, rice yields would have grown 12 percent faster in the late 1980s and early 1990s. Erosion affected maize, wheat, and cash crops in North China the most, reducing production by up to 20 percent in the 1980s and 1990s (Huang and Rozelle 1994 and 1996; Huang, Rosegrant, and Rozelle 1996). A grain-yield function estimated for 1975-90, pooling data for 23 provinces, found yield to be significantly influenced by degradation, with elasticities of grain yield of-0.146 for soil erosion, -0.003 for salinization, and -0.276 for multiple cropping intensity. The latter elasticity was probably due to nutrient depletion (Huang and Rozelle 1994 and 1996).

Table 10 - ASSOD estimates of the area and effect of soil degradation in South and Southeast Asia

Type of degradation

Land degraded by degree of degradation


Nondegraded or negligible

Light

Moderate

Strong or extreme

Degraded land as a percent of total land


(percent)

Loss of topsoil from water erosion

84.3

9.5

5.3

0.9

15.7

Terrain deformation from water erosion

95.1

1.2

0.9

1.8

4.9

Off-site effects in uplands from water erosion

99.7

0.2

...

0.6

0.3

Topsoil loss from wind erosion

94.6

4.0

0.9

0.4

5.4

Terrain deformation from wind erosion

95.8

0.4

0.6

3.2

4.2

Off-site effects from wind erosion

99.2

0.1

0.5

0.2

0.8

Fertility decline


Total land

93.8

3.7

2.4

0.1

6.2


(Arable land)

(69.6)

(18.0)

(11.9)

(0.5)

(30.4)

Salinization


Total land

97.9

1.1

0.8

0.2

2.1


(Arable land)

(89.8)

(5.5)

(3.8)

(0.9)

(10.2)

Dystrification


(Arable land)

(99.3)

(0.5)

(0.2)

...

(0.7)

Aridification


Total land

98.7

1.3

...

...

1.3


(Arable land)

(93.3)

(6.3)

...

(0.4)

(6.7)

Compaction


Total land

99.9

0.1

0.1


(Arable land)

(98.7)

(0.8)

(0.4)

(1.3)

Waterlogging


Total land

99.6

1.0

0.3

0.1

1.4


(Arable land)

(92.9)

(5.0)

(1.4)

(0.7)

(7.1)

Source: van Lynden and Oldeman 1997.

Notes: Estimates of arable land degradation were calculated by the author using FAO data on total arable land area, and assuming that all land reported by ASSOD with these types of degradation were arable lands. This is generally but not always true, and thus these figures may overestimate soil degradation on arable lands. ASSOD stands for Assessment of Human-Induced Soil Degradation in South and Southeast Asia. The total area surveyed was 1,843.4 million hectares. The total area of arable land reported by the Food and Agriculture Organization of the United Nations was 380 million hectares (20.6 percent). “Light” degradation implies little impact on productivity. “Moderate” implies major impact and a need to compensate for degradation with high management. Medium management does not compensate and low management leads to significant productivity decline. “Strong” or “extreme” implies a major impact on productivity that cannot be compensated for even with high levels of management and is unproductive under low management. Ellipses indicate negligible amounts.

For India, Sehgal and Abrol (1994) synthesized the results of national soil surveys, a survey of national soil experts, and crop experimental data to estimate the scale and productivity effects of soil degradation. They concluded that although no significant degradation affects 36 percent of the land area in India, 5 percent of the land is suffering from low degradation (less than 15 percent loss in yield), 11 percent from moderate degradation (15-33 percent loss), 43 percent from high degradation (33-67 percent loss), and 5 percent had become so degraded that soils were unusable.

A 1985-86 household- and plot-level study in four villages in Uttar Pradesh, India, found significant effects of salinization and waterlogging on productivity over the preceding 10-year period (Joshi and Jha 1991). Paddy yield declined by 61 percent and wheat yield by 68 percent on salt-affected soils. The average yield of high-yielding paddy varieties on alkaline plots decreased by 51 percent and local varieties by 46 percent. Under waterlogged conditions, the corresponding figures were 41 percent and 26 percent. Alkalinity accounted for as much as 72 percent of the difference in gross income between normal and salt-affected plots; the other 23-28 percent could be attributed to reduced input use on degraded soils (Joshi and Jha 1991).

Ali and Byerlee (1998) used district-level data for 33 crops, 8 livestock products, and 17 input categories to estimate changes in total factor productivity from 1971 to 1994 in 4 irrigated production systems of Punjab province, Pakistan. Average annual growth in total factor productivity was moderately high, at 1.25 percent for both crops and livestock, but wide regional variation in productivity growth was observed, with negative growth in the wheat-rice system. A second disaggregated data set on soil and water quality was then used to analyze underlying effects of resource degradation through application of a cost function. Ali and Byerlee found that continuous and widespread resource degradation lowered productivity growth in the province by about 58 percent on average. The largest effect was in the wheat-rice system, where resource degradation more than offset the productivity effects of technological change.

More subtle types of degradation in Asia’s intensive, irrigated agricultural systems are a growing concern in the scientific community (Olk et al. 1996; Cassman and Harwood 1995). Long-term experiments on plots representing the major farming systems in India found mixed evidence. There were negative trends in soil productivity without the use of farmyard manure, and flat trends with manure, in an annual, double-crop, irrigated rice system in the warm, subhumid tropics of Orissa. In the warm, subhumid subtropics of Uttar Pradesh, an irrigated rice-wheat system showed negative soil productivity trends for rice and flat trends for wheat. In the warm semiarid subtropics of the Punjab, a maize-wheat cropping system showed flat productivity trends for maize and positive for wheat (Cassman, Steiner, and Johnson 1995).

Agricultural Income and Economic Growth

Young (1993) estimated the annual cost of soil degradation in South Asia at $9.8-$ 11.0 billion, the equivalent of 7 percent of AGDP. Water and wind erosion accounted for more than two thirds of the loss, salinization and waterlogging for about a fifth, and soil fertility decline the rest. In Pakistan, the value of reduced wheat production due to waterlogging and salinization in 1993 equaled about 5 percent of AGDP, while in India, annual cereal production loss amounted to about 5 percent of AGDP. Pagiola (1995) concluded that total factor productivity in Bangladesh had declined between 1975 and 1985 due to deteriorating nutrient balance and loss of organic matter. Significant negative trends over time were found for both farmer production and experimental plots.

The densely populated and intensively cultivated island of Java appears to have experienced high soil degradation (De Graaffand Wiersum 1992; Diemont, Smiet, and Nurdin 1991). Magrath and Arens (1989) calculated that agricultural productivity was declining by a rate of 2-5 percent a year due to soil erosion, creating annual economic losses of nearly 1 percent of the gross national product (GNP) (or approximately 3 percent of AGDP). Repetto et al. (1989) found that for two crop groups on 25 soil types, the one-year costs of erosion in Java in 1984 equaled 4 percent of the annual value of rainfed farm output - the same order of magnitude as the annual recorded growth in agricultural production in the uplands. The capitalized losses in future productivity equaled 40 percent of the total value of annual production.

Huang and Rozelle (1994 and 1996) and Huang, Rosegrant, and Rozelle (1996) calculated that the economic loss from soil degradation in China in the late 1980s reached $700 million (1990 prices) - an amount equal to China’s budget for rural infrastructure investment, though less than 1 percent of AGDP. But Lindert’s (forthcoming b) study showed that despite some nutrient depletion in China, the economic value of topsoil rose by nearly 8 percent between the 1950s and 1980s (4 percent in the north, 16 percent in the south). The shifts to soil-preserving products and practices largely accounted for this gain.

Consumption by Poor Farmers

None of the Asian studies analyzed the impact of soil degradation on food consumption by the poor. However, an econometric analysis of the effects of policy on soil erosion and salinization, using district-level data in China, showed that degradation had a much greater effect on poor and densely populated areas than other areas, and that general agricultural policies had a greater impact on this outcome than specific land management policies (Rozelle, Huang, and Zhang 1997).

Effects of Soil Degradation in Sub-Saharan Africa

Soil degradation is widespread in Sub-Saharan Africa. Agricultural lands are especially prone to erosion and nutrient depletion. Reported yield losses range from modest levels (2 percent decline over several decades) to catastrophic (>50 percent), depending on crop, soil type, climate, and production systems, with most studies reporting significant losses. Direct economic losses due to declining yields and lost nutrients are large in terms of the national economy, even in recent studies using more conservative methods of estimation. Several studies assessed the effects of degradation on rural poverty, but results were not consistent.

More subnational studies of the economic effects of degradation exist for Sub-Saharan Africa than for other regions. These studies are concerned mostly with marginal lands that are experiencing rapid population growth and a shift from short-fallow systems to permanent cropping, and with high-quality rainfed lands that have high population densities.

Extent of Degradation

Five continental-scale studies have assessed the extent of soil degradation in Africa. A literature review by Dregne (1990) of 33 countries found compelling evidence of serious land degradation in subregions of 13 countries: Algeria, Ethiopia, Ghana, Kenya, Lesotho, Mali, Morocco, Nigeria, Swaziland, Tanzania, Tunisia, Uganda, and Zimbabwe. In another literature review, focused on drylands only, Dregne and Chou (1992) estimated that 73 percent of drylands were degraded and 51 percent severely degraded. They concluded that 18 percent of irrigated lands, 61 percent of rainfed lands, and 74 percent of rangelands located in drylands are degraded.

The GLASOD expert survey found that 65 percent of soils on agricultural lands in Africa had become degraded since the middle of this century, as had 31 percent of permanent pastures, and 19 percent of woodlands and forests (Oldeman, Hakkeling, and Sombroek 1991). Serious degradation affected 19 percent of agricultural land. A high proportion (72 percent) of degraded land was in drylands. The most widespread cause of degradation was water erosion, followed by wind erosion, chemical degradation (three-quarters from nutrient loss, the rest from salinization), and physical degradation. Overgrazing accounted for half of all degradation, followed by agricultural activities, deforestation, and overexploitation.

Lal (1995) calculated continent-wide soil erosion rates from water using data from the mid to late 1980s, and then used these rates to compute cumulative soil erosion for 1970-90. The highest erosion rates occurred in the Maghreb region of Northwestern Africa, the East African highlands, eastern Madagascar, and parts of Southern Africa. Excluding the 42.5 percent of arid lands and deserts with no measurable water erosion, Lal found that land area affected by erosion fell into the following six classes of erosion hazard: none, 8 percent; slight, 49 percent; low, 17 percent; moderate, 7 percent; high, 13 percent; and severe, 6 percent.

Stoorvogel, Smaling, and Janssen (1993) undertook a continental-scale study of soil nutrient depletion in the early 1990s. They calculated that average annual nutrient loss on arable lands in 1982-84 amounted to 22 kilograms per hectare of nitrogen, 2.5 kilograms of phosphorus, and 15 kilograms of potassium. The main loss of nutrients occurred through the harvest and removal of the crops and inadequate use of organic and inorganic inputs. The authors extrapolated that the average nutrient loss over the past 30 years equaled 1.4 tons per hectare of urea fertilizer, 375 kilograms of triple superphosphate, and 896 kilograms of potassium chloride. Rates of nutrient depletion were especially high in densely populated and erosion-prone countries in East and Southern Africa - Ethiopia, Kenya, Malawi, and Rwanda in particular. Countries in semiarid environments, Botswana and Mali, for example, experienced low or zero depletion rates.

Subnational studies of nutrient depletion found annual losses of 112 kilograms per hectare of nitrogen, 2.5 kilograms of phosphorus, and 70 kilograms of potassium in the western Kisii highlands of Kenya; and significantly lower losses in southern Mali (Smaling 1993; Smaling, Nandwa, and Jans-sen 1997). Farm monitoring and modeling of nutrient cycles for the western highlands of Kenya found that more nitrogen (63 kilograms per hectare) was being lost through leaching, nitrification, and volatilization than through removal of crop harvests (43 kilograms per hectare). Depending on type of farm management practice, net nitrogen balances on cropped land varied between -39 and 110 kilograms per hectare per year, and net phosphorus balances between -7 and 31 kilograms per hectare per year (Shepherd and Soule 1998).

Agricultural Supply

Using GLASOD data, the productivity loss in Africa from soil degradation since World War II has been estimated at 25 percent for cropland and 8-14 percent for cropland and pasture together (Oldeman 1998; also see Table 8). These figures are consistent with Dregne’s (1990) estimates that irreversible soil productivity losses of at least 20 percent due to erosion had occurred over the past century in large parts of Algeria, Ethiopia, Ghana, Kenya, Lesotho, Morocco, Nigeria, Southern Africa, Swaziland, Tunisia, and Uganda. More dramatic productivity declines under agricultural intensification are suggested by a review of African farm-survey and experimental data, which shows that in originally fertile lands, under continuous cropping without nutrient inputs, cereal grain yields declined from 2-4 tons per hectare to under 1 ton per hectare (Sanchez et. al. 1997).

The effects of erosion on crop productivity may be smaller, though still important. Crop yield losses in 1989 due to past erosion ranged from 2 to 40 percent, with a mean of 6.2 percent for Sub-Saharan Africa (8.2 percent for all Africa). In the absence of erosion, 3.6 million tons more of cereal (8.2 million for the continent), 6.5 million tons more of roots and tubers (9.2 million), and 0.4 million tons more of pulses (0.6 million) would have been produced in 1989 (Lal 1995).

Country-level data on productivity effects are quite varied. A study of the effects of soil erosion in Malawi (World Bank 1992a) found that annual yield loss for specific crops grown in Malawi varied from 4 to 11 percent. National- and district-level estimates for Lesotho showed negative but statistically insignificant yield declines for maize and sorghum that were associated with degradation (B9910). Grohs 1994 (reported in B996) evaluated the effects of erosion on yield across eight provinces in Zimbabwe and found no statistically significant influence of erosion on the yield trend for maize, possibly due to the overriding importance of rainfall variability in these areas. A crop growth simulation model for the Chaouia Plains in Morocco showed that erosion had a significant impact on yields only on slopes with a gradient of over 15 percent. Yields declined 20-30 percent over 50 years, but returns to wheat declined 40-50 percent (Pagiola 1994).

Field studies in three ecoregions of Tanzania that included experimental trials and field surveys of crop growth under different erosion and acidity conditions were used to construct models of soil erosion-productivity relationships. For every millimeter of topsoil depth reduction, maize yields declined by less than 1 percent to 5 percent for different soil types. Highland maize yields in four farming systems in different ecozones declined significantly, although with application of fertilizer, the decline was only half as much. Fertilizer-induced soil acidification reduced highland maize yield to zero in 20 years; with application of lime, yields dropped to half in 30 years. Cotton yields could be maintained with adequate chemical inputs; coffee yields were also stable (Aune et al. 1997; Aune 1995). A large field survey in Tanzania found that yields were 30 percent higher in the least eroded soil classes than in the most eroded classes (Kilasara et al.1995).

Agricultural Income and Economic Growth

Perhaps due to the centrality of agriculture in African economies, the economic effects of soil degradation are relatively high. B1996) evaluated evidence on the economic losses due to soil erosion from 12 studies completed in 8 countries in Sub-Saharan Africa (Table 11). The gross annual immediate loss (the lost value of that year’s production) ranged from under 1 percent of AGDP in Ethiopia, Madagascar, Mali, and South Africa, to 2-5 percent of AGDP in Ethiopia14 and Ghana, and exceeded 8 percent in Zimbabwe. The gross discounted future loss (the value of the stream of losses due to a particular year’s soil degradation) ranged from <1 percent in Ethiopia and Zimbabwe to 18 percent in Malawi. The gross discounted cumulative loss (which assumes a continued process of degradation over time), calculated for five countries, ranged from under 1 percent of AGDP to a high of 36-44 percent in Ethiopia. Except in Zimbabwe, most erosion effects were less than 5 percent of AGDP.

14Estimates for effects of erosion in Ethiopia vary with the methodology used. Sutcliffe found the economic value of damage from soil erosion to be only a tenth of that of the 1986 Ethiopian Highlands Reclamation Study; however, he reported much higher costs from nutrient loss (Sutcliffe 1993; B996). Bnd Cassells (1995) reassessed Sutcliffe’s data, concluding that net yield loss from cropland was half or less of the gross figures used in the Sutcliffe analysis, though they also emphasized nutrient depletion problems.

For Zimbabwe, using experimental data from the 1950s and 1960s on four soil types and numerous crops to derive the cost of fertilizer replacement for soil nutrients lost through depletion and erosion, Stocking (1986) concluded that nitrogen and phosphorus losses on arable lands were equal to three times the level of total fertilizer applications in 1984/85 (not including nutrients in run-off water). The total annual loss from arable land amounted to US$ 150 million ($5-20 per hectare), and to US$ 1.5 billion for all land.

Estimates of the effect of soil degradation on the broader economy in Ghana show productivity losses due to soil degradation of 2.1 percent per year in cocoa and 2.9 percent per year in all agriculture. As a result, economic growth declines by 1 percent, even with increased fertilizer use. In some scenarios, real economic growth declines up to 4.8 percent over the course of 8 years (Alfsen et al. 1997).

Household and field survey data from Rwanda illustrate farm income effects of erosion. Farm fields with higher erosion have lower marginal value product (MVP) of land - 30 percent lower on the more eroded soils. The MVP for labor is 15 percent lower on high erosion farms than on those with low erosion. Conservation investments on less degraded farms increased MVP by 27 percent. For moderately and very degraded farms, the increments were 28-34 percent and 42 percent, respectively (Clay, Reardon, and Kangasniemi 1998; Byringiro and Reardon 1996). Data from monthly farm monitoring in three districts in Kenya found that the average cost of replacing depleted soil nutrients was equivalent to 32 percent of average net farm income (Jager et al. 1998).

Table 11 - Comparative analysis of national-level annual economic effects of soil erosion in Africa

Study

Country

Gross annual immediate loss

Gross discounted future loss

Gross discounted cumulative loss

(US$ million)

FAO (1986)

Ethiopia

14.8 (<1)

-

2,993 (44)

Sutcliffe (1993)

Ethiopia

155 (5)

15 (<1)

-

Bnd Cassells (1995)

Ethiopia

130 (4)

22 (<1)

2,431 (36)

Convery and Tutu (1990)

Ghana

166.4 (5)

-

-

B1991b)

Lesotho

0.3 (<1)

3.2 (5)

31.2(5)

World Bank (1988)

Madagascar

4.9-7.6 (<1)

-

-

World Bank (1992a)

Malawi

6.6-19.0 (3)

48- 136 (18)

-

Bishop and Allen (1989)

Mali

2.9-11.6 (<1)

19.3-76.6(4)

-

McKenzie (1994)

South Africa

18 (<1)

173(4)

503(<1)

Stocking (1986)

Zimbabwe

117 (9)

-

-

Norse and Saigal (1992)

Zimbabwe

99.5 (8)

-

-

Grohs(1994)

Zimbabwe

0.6 (<1)

6.7 (<1)

44.7 (<1)

Source: Calculations are based on B994 and 1996 (methods detailed by author from original studies). Costs of erosion include yield losses and value of nutrients lost through erosion.

Notes: Data in parentheses indicate percent of agricultural gross domestic product (AGDP). AGDP data are based on figures for 1992 from the World Bank (1994), inflated by 3.9 percent per year to 1994. Percentages are based on midpoint values of economic losses, where an interval is used. A dash indicates that the figure could not he calculated because data were not available.

Consumption by Poor Farmers

Geographic information systems have been used to examine the correlation of key poverty indicators for West Africa with the GLASOD data on soil degradation and agroclimatic zones. The proportion of children who died before the age of five was highest (more than 30 percent of children) in areas with high soil degradation. A little over half of all mortality occurred in areas of high or very high degradation. Other variables, such as adult female literacy, rate of primary school enrollment, and incidence of children with stunted growth do not show a clear relation with degradation as measured by GLASOD. Poverty in dexes are correlated more with agroclimatic zones. The incidence of child mortality declines moving from arid to moist subhumid climates (that is, from north to south) and adult female literacy rates and primary school enrollment rise strikingly (UNEP/GRID-Arendal 1998).

A bioeconomic linear programming model using nutrient budgets from a sample of farms in the western highlands of Kenya found that farmers with low and medium resource endowments had only 7-13 percent of the farm income of farmers with high resource endowments, due to lack of resources for soil nutrient replenishment (Shepherd and Soule 1998). However, a study based on monthly monitoring of 26 representative farms in three other districts of Kenya found no relation between net farm income and soil nutrient balance (Jager et al. 1998).

Effects of Soil Degradation in Mexico and Central America

Summarized here are two regional studies and ten national studies, all published in English. Several of these studies concern degradation in densely populated marginal lands, particularly hillsides. Agricultural supply and income effects in these areas appear to be very significant; large rural consumption effects due to degradation are implied but not documented. Policy issues include erosion and off-site effects of agrochemical use in some high quality lands in Costa Rica and Mexico. Salinization of irrigated lands in Mexico also was cited as a problem, but its effects were not documented.

Extent of Degradation

Dregne and Chou (1992) estimated that about 430 of the 570 million hectares of drylands in South America, Central America, and the Caribbean had been moderately to very severely degraded. A quarter of irrigated lands had been degraded through salinization and waterlogging, 38 percent of rainfed cropland through water erosion, and 80 percent of rangeland through degradation of natural vegetation.

The GLASOD study found that nearly a third of land in Central America (excluding Mexico) was degraded, including 74 percent of agricultural land and 38 percent of forest land, largely due to water erosion. Half of the degraded soils were moderately affected and half were strongly or severely affected (Tables 6-8).

Agricultural Supply

Using the GLASOD data, Oldeman (1998) calculated that agricultural productivity in Central America was 37 percent lower than what it would otherwise have been without soil degradation - the largest loss of any region. The cumulative loss for South America was 13.9 percent, only a little more than Asia (Table 8).

Lutz, Pagiola, and Reiche (1994) examined the potential profitability of soil conservation measures in Central America and the Caribbean. Without conservation measures, over a 10-year period, peanut yields would remain stable in the Dominican Republic; maize yields would decline by 20-25 percent in the subhumid hillsides of Honduras; bean yields also would decline by 20-25 percent in the Dominican Republic; coffee yields would decline by 10 percent in the Costa Rican highlands; maize and sorghum yields would decline by 60 percent in the hillsides of Haiti; and cocoyam yields would drop to zero in the humid lowlands of Costa Rica.

White and Jickling (1994) evaluated soil erosion effects in the humid, bimodal Central Plateau of Haiti, finding an annual yield decline in corn and sorghum of 6 percent in the first 10 years without conservation, with smaller declines thereafter. Net financial returns declined to zero after 24 years.

Cuesta (1994) compared the effect of uncontrolled soil erosion on crop production in three sites in different ecozones of Costa Rica. Highland coffee yields declined by half in 3 years and to zero in 20 years. Highland potato yields declined more slowly, by 40 percent after 50 years. Lowland cocoyam yields declined by more than half the first year and to zero in the fourth year.

These studies probably underestimate the effectiveness of farmers’ soil protection practices, particularly on more erosion-resistant soils and in permanent crop fields. Pagiola and Dixon (1997) assessed the qualitative effect of soil erosion in El Salvador through a household and plot survey. Farmers reported that erosion was causing significant problems on 36 percent of fields on mild slopes, 70 percent of fields on moderate slopes, and 82 percent of fields on steep slopes. However, severe long-term productivity declines were only expected on 16 percent of fields on steep slopes and 5 percent of fields on moderate slopes.

Agricultural Income and Economic Growth

Solno et al. (1991) evaluated the economic effects of soil erosion in Costa Rica by measuring the cost of replacing lost nutrients. Annual replacement costs were found to equal 5.3-13.3 percent of annual value-added in agriculture in the same year.

McIntire (1994) examined typical farming situations in five tropical and eight highland or semiarid states of Mexico. Erosion led to an average estimated loss in maize production valued at 2.7 percent of AGDP, reaching 12.3 percent in some states. Economic losses were nine times higher in the highlands and semiarid regions than in the lowland tropics. Generally, losses were four times higher without than with soil conservation measures. At a 5 percent discount rate, losses were 3-4 times higher than those calculated at a 10 percent rate.

Alfsen et al. (1996) constructed a national CGE model integrating soil erosion effects for Nicaragua based on local expert assessment of productivity losses. Annual productivity loss due to erosion in 1991 was assumed to be less than 1 percent per year for bananas, rice, sugar, and vegetables; 1-2 percent per year for coffee, cotton, and sorghum; and more than 2 percent per year for sesame, maize, beans, and pasture. This level of erosion would have major national economic effects: GDP, imports, exports, and consumption in the year 2000 are projected to decrease by 4-7 percent from the baseline scenario, while total investment is projected to decrease by 9 percent.

Consumption by Poor Farmers

Few of these studies addressed the effects on farmer welfare. In a study of four villages in central Honduras, Casey and Paolisso (1996), using household surveys, soil sampling, and group interviews, found that declining household income due to soil degradation had led to reduced male labor in farming and increased off-farm labor. An increase in women’s labor in maize production occurred in poor households despite declining returns to labor.

The Alfsen et al. (1996) CGE model calculated that erosion in 1991-2000 in Nicaragua would lead to a rise in the producer and domestic price indexes by 1.7 percent and 2.1 percent, respectively, compared to the 2000 baseline, while consumer price indexes would increase by 4.0-5.8 percent for different social groups. A large part of the cost of erosion is passed on from smallholders to other social classes through price effects. Assuming no unemployment in the rural sector, net urban migration would increase from 3.5 percent of urban labor supply to almost 4 percent per year.