Previous PageTable Of ContentsNext Page


Crop production and natural resource use

4.1 Introduction

This chapter discusses the main agronomic factors underlying the projections of crop production presented in Chapter 3. The focus is on crop production in developing countries, for which the projections were unfolded into land and yield projections under rainfed (five land classes) and irrigated conditions. Although the underlying analysis was carried out at the level of individual countries, the discussion here is limited to presenting the results at the level of major regions, which unavoidably masks wide intercountry differences. The parameters underlying the livestock production projections will be discussed in Chapter 5. Selected technology issues such as the scope for further yield increases, technologies in support of sustainable agriculture and the role of biotechnology are discussed in Chapter 11. Issues of environment and the possible impact of climate change on crop production are the subjects of Chapters 12 and 13.

4.2 Sources of growth in crop production

Aggregate crop production at the world level is projected to grow over the period to 2030 at 1.4 percent p.a., down from the annual growth of 2.1 percent of the past 30 years (Table 4.1). For the developing countries as a group, the corresponding growth rates are 1.6 and 3.1 percent p.a., respectively (or 1.8 and 2.7 percent p.a., excluding China). The reasons for this continuing deceleration in crop production growth have been explained in Chapter 3.

Table 4.1: Annual crop production growth

 

1969-99

1979-99

1989-99

1997/99
-2015

2015-30

1997/99
-2030

Percentage

All developing countries

3.1

3.1

3.2

1.7

1.4

1.6

   excl. China

2.7

2.7

2.5

2.0

1.6

1.8

   excl. China and India

2.7

2.6

2.5

2.0

1.7

1.9

   Sub-Saharan Africa

2.3

3.3

3.3

2.6

2.5

2.5

   Near East/North Africa

2.9

2.9

2.6

1.8

1.5

1.6

   Latin America and the Caribbean

2.6

2.3

2.6

1.8

1.6

1.7

   South Asia

2.8

3.0

2.4

2.1

1.5

1.8

   East Asia

3.6

3.5

3.7

1.3

1.1

1.2

Industrial countries

1.4

1.1

1.6

0.9

0.9

0.9

Transition countries

-0.6

-1.6

-3.7

0.7

0.7

0.7

World

2.1

2.0

2.1

1.5

1.3

1.4

The projected increase in world crop production over the period from 1997/99 to 2030 is 55 percent, against 126 percent over the past period of similar length. Similar increases for the developing countries as a group are 67 and 191 percent, respectively. The only region where the projected increase would be about the same as the historical one would be sub-Saharan Africa, namely 123 and 115 percent, respectively. The faster growth in the developing countries, as compared to the world average, means that by 2030 this group of countries will account for almost three-quarters (72 percent) of world crop production, up from two-thirds (67 percent) in 1997/99 and just over half (53 percent) 30 years earlier.

There are three sources of growth in crop production: arable land expansion which, together with increases in cropping intensities (i.e. increasing multiple cropping and shorter fallow periods), leads to an expansion in harvested area; and yield growth. About 80 percent of the projected growth in crop production in developing countries will come from intensification in the form of yield increases (67 percent) and higher cropping intensities (12 percent, Table 4.2). The share due to intensification will go up to 90 percent and higher in the land-scarce regions of the Near East/North Africa and South Asia. The results for East Asia are heavily influenced by China. Excluding the latter, intensification will account for just over 70 percent of crop production growth in East Asia. Arable land expansion will remain an important factor in crop production growth in many countries of sub-Saharan Africa, Latin America and some countries in East Asia, although much less so than in the past. The estimated contribution of yield increases is partly a result of the increasing share of irrigated agriculture in total crop production (see Section 4.4.1), and irrigated agriculture is normally more«intensive» than rainfed agriculture.

Table 4.2: Sources of growth in crop production (percentage)

 

Arable land expansion (1)

Increases in cropping intensity (2)

Harvested land expansion (1+2)

Yield increases

1961-1999

1997/99 -2030

1961-1999

1997/99-2030

1961-1999

1997/99-2030

1961-1999

1997/99-2030

All developing countries

23

21

6

12

29

33

71

67

   excl. China

23

24

13

13

36

37

64

63

   excl. China and India

29

28

16

16

45

44

55

56

   Sub-Saharan Africa

35

27

31

12

66

39

34

61

   Near East/North Africa

14

13

14

19

28

32

72

68

   Latin America and the Caribbean

46

33

-1

21

45

54

55

46

    South Asia

6

6

14

13

20

19

80

81

    East Asia

26

5

-5

14

21

19

79

81

World

15

 

7

 

22

 

78

 

All developing countries

    Crop production - rainfed

 

25

 

11

 

36

 

64

    Crop production - irrigated

 

28

 

15

 

43

 

57

The results shown in Table 4.2 should be taken as rough indications only. For example, yields here are weighted yields (1989/91 price weights) for 34 crops and historical data for arable land for many countries are particularly unreliable.1 Data on cropping intensities for most countries are non-existent and for this study were derived by comparing data on harvested land, aggregated over all crops, with data on arable land. The projections are the end result of a detailed investigation of present and future land/yield combinations for 34 crops under rainfed and irrigated cultivation conditions, for 93 developing countries.2 In the developed countries, the area of arable land in crop production has been stagnant since the early 1970s and recently declining. Hence growth in yields and more intensive use of land accounted for all of their growth in crop production and also compensated for losses in their arable land area.

Growth in wheat and rice production in the developing countries increasingly will have to come from gains in yield (more than four-fifths), while expansion of harvested land will continue to be a major contributor to production growth of maize, possibly even more so than in the past (Table 4.3). These differences are partly because the bulk of wheat and rice is produced in the land-scarce regions of Asia and the Near East/North Africa while maize is the major cereal crop in sub-Saharan Africa and Latin America, regions where many countries still have room for area expansion. As discussed in Chapter 3, an increasing share of the increment in the production of cereals, mainly coarse grains, will be used in livestock feed. As a result, maize production in the developing countries is projected to grow at 2.2 percent p.a. against«only» 1.3 percent for wheat and 1.0 percent for rice. Such contrasts are particularly marked in China where wheat and rice production is expected to grow only marginally over the projection period, while maize production is expected to nearly double. Hence there will be a corresponding decline in the wheat and rice areas but an increase of 36 percent in the maize area.

Table 4.3: Sources of growth for major cereals in developing countries (percentage)

 

Harvested land expansion

Yield increases

1961 - 1999

1997/99 - 2030

1961 - 1999

1997/99 - 2030

Wheat

22

17

78

83

Rice

23

14

77

88

Maize

30

49

70

51

The actual combination of the factors used in crop production (land, labour and capital) in the different countries will be determined by their relative prices. For example, taking the physical availability of land as a proxy for its relative scarcity and hence price, one would expect land to play a greater role in crop production the less scarce and cheaper it is. For the 60 countries out of the 93 developing countries studied in detail, which at present use less than 60 percent of their land estimated to have some rainfed crop production potential (see Section 4.3.1), arable land expansion is projected to account for one-third of their crop production growth. In the group of 33 land-scarce countries - defined here as countries with more than 60 percent of their suitable land already in use - the contribution of land expansion is estimated to be less than 10 percent.

For the developing countries, this study made an attempt to break down crop production by rainfed and irrigated land in order to analyse the contribution of irrigated crop production to total crop production. It is estimated that in the developing countries at present, irrigated agriculture, with about a fifth of all arable land, accounts for 40 percent of all crop production and almost 60 percent of cereal production (Table 4.4). It should be emphasized that, apart from some major crops in some countries, there are only very limited data on irrigated land by crops and the results presented in Table 4.4 are almost entirely based on expert judgement (see Appendix 2 for the approach followed in this study). Nevertheless, the results suggest an increasing importance of irrigated agriculture, which accounts for a third of the total increase in arable land and for over 70 percent of the projected increase in cereal production.

Table 4.4: Shares of irrigated production in total crop production of developing countries

 

All crops

Cereals

Shares (percentage)

Arable land

Harvested land

Production

Harvested land

Production

Share in 1997/99

21

29

40

39

59

Share in 2030

22

32

47

44

64

Share in increment 1997/99-2030

33

47

57

75

73

4.3 Agricultural land

At present some 11 percent (1.5 billion ha) of the globe's land surface (13.4 billion ha) is used in crop production (arable land and land under permanent crops). This area represents slightly over a third (36 percent) of the land estimated to be to some degree suitable for crop production. The fact that there remain some 2.7 billion ha with crop production potential suggests that there is still scope for further expansion of agricultural land. However, there is also a perception, at least in some quarters, that there is no more, or very little, land to bring under cultivation. In what follows, an attempt is made to shed some light on these contrasting views by first discussing the most recent estimates of land with crop production potential and some constraints to exploiting such suitable areas (Section 4.3.1). Then the projected expansion of the agricultural area during the next three decades (to 2030) is presented in Section 4.3.2, while Section 4.3.3 speculates about whether or not there will be an increasing scarcity of land for agriculture.

4.3.1 Land with crop production potential for rainfed agriculture

Notwithstanding the predominance of yield increases in the growth of agricultural production, land expansion will continue to be a significant factor in those developing countries and regions where the potential for expansion exists and the prevailing farming systems and more general demographic and socio-economic conditions favour it. One of the frequently asked questions in the debate on world food futures and sustainability is: how much land is there that could be used to produce food to meet the needs of the growing population? Since the late 1970s, FAO has conducted a series of studies to determine the suitability of land for growing various crops. Recently, a new study was undertaken together with the International Institute for Applied Systems Analysis (IIASA) to refine the methods, update databases and extend the coverage to all countries in the world by including also countries in temperate and boreal climates, which were previously not covered. A summary description of the method is given in Box 4.1 and a full description and presentation of results can be found in Fischer, van Velthuizen and Nachtergaele (2000).

Table 4.5 gives some results for selected crops and input levels. At a high input level (commercial farm operations, see Box 4.1), over 1.1 billion ha would be suitable for growing wheat at an average maximum attainable yield level of 6.3 tonnes/ha, i.e. taking into account all climate, soil and terrain constraints. At the low technology level (subsistence farming), 1.5 billion ha would be suitable, but at an average maximum yield level of only 2.3 tonnes/ha. The suitable area at this lower input level is greater because, for example, tractors (high input level) cannot be used on steep slopes. For developing countries alone, the estimates are 314 million ha and 5.3 tonnes/ha under the high technology level because most of the suitable area for wheat at this input level is in the developed countries. For the other crops shown, the bulk of the suitable area is in the developing countries.

Table 4.5: Land with rainfed crop production potential for selected crop and input levels

 

Actual 1997/99

Total suitable

Very suitable

Suitable

Moderately suitable

Marginally suitable

A

Y

% of land

A

Y

A

Y

A

Y

A

Y

A

Y

Wheat - high input

World

226

2.6

8.5

1139

6.3

160

9.4

397

7.8

361

5.2

221

3.0

All developing

111

2.5

4.1

314

5.3

38

8.2

97

6.7

105

4.7

74

2.7

Transition countries

51

2.0

15.6

359

6.3

44

9.6

107

8.2

130

5.3

78

3.4

Industrial countries

65

3.3

14.3

466

6.9

78

9.9

193

8.1

125

5.4

69

2.9

Wheat - low input

World

226

2.6

11.3

1510

2.3

175

4.1

403

2.9

487

2.0

445

1.2

All developing

111

2.5

6.1

467

1.7

31

3.1

101

2.4

152

1.7

183

1.0

Transition countries

51

2.0

18.5

425

2.6

47

4.4

127

3.2

151

2.3

100

1.3

Industrial countries

65

3.3

19.0

617

2.5

97

4.2

175

3.1

183

2.1

161

1.2

Rice - high input

World

161

3.6

12.5

1678

4.3

348

6.2

555

4.9

439

3.6

337

2.2

All developing

157

3.6

21.4

1634

4.3

347

6.2

549

4.9

423

3.6

315

2.2

Transition countries

0.5

2.5

0.0

1

3.2

0

0.0

0

7.6

0

4.3

1

2.6

Industrial countries

4

6.5

1.3

44

4.1

1

9.0

6

6.4

16

4.6

21

2.6

Maize - high input

World

144

4.2

11.6

1557

8.2

246

13.2

439

10.3

393

7.4

479

4.3

All developing

99

2.8

18.2

1382

8.0

221

13.2

359

10.3

339

7.3

463

4.3

Transition countries

9

3.9

0.5

11

6.9

1

13.4

8

7.3

2

5.3

1

3.4

Industrial countries

38

7.7

5.0

163

9.6

24

13.9

73

10.7

52

7.7

15

4.6

Soybean- high input

World

72

2.1

10.3

1385

2.4

183

4.0

353

3.1

415

2.2

434

1.3

All developing

41

1.8

16.8

1277

2.4

173

4.0

324

3.1

372

2.2

407

1.3

Transition countries

0.7

1.3

0.1

3

3.0

1

4.2

1

3.2

1

2.3

0

1.5

Industrial countries

30

2.6

3.2

105

2.6

10

4.1

27

3.3

42

2.4

26

1.5

Notes: A=area in million ha; Y=average attainable yield in tonnes/ha. The 1997/99 data are not distinguished by input level as information does not exist. The area data for 1997/99 refer to harvested area and elsewhere in the table to arable area.

Summing over all crops and technology levels considered (see Box 4.1), it is estimated that about 30 percent of the world's land surface, or 4.2 billion ha, is suitable for rainfed agriculture (Table 4.6). Of this area, the developing countries have some .8 billion ha of land of varying qualities that have potential for growing rainfed crops at yields above an«acceptable» minimum level. Of this land, nearly 960 million ha are already in cultivation. The remaining 1.8 billion ha would therefore seem to provide significant scope for further expansion of agriculture in developing countries. However, this favourable impression must be much qualified if a number of considerations and constraints are taken into account.

Box 4.1 Summary methodology of estimating land potential for rainfed agriculture

For each country an evaluation was made of the suitability of land for growing 30 crops1 under rainfed conditions and various levels of technology. The basic data for the evaluation consist of several georeferenced data sets: the inventory of soil characteristics from the digital FAO-UNESCO Soil Map of the World (SMW; FAO, 1995a), an inventory of terrain characteristics contained in a digital elevation model (DEM; EROS Data Center, 1998), and an inventory of climate regimes (New, Hulme and Jones, 1999). The data on temperature, rainfall, relative humidity, wind speed and radiation are used, together with information on evapotranspiration, to define the length of growing periods (LGPs), i.e. the number of days in a year when moisture availability in the soil and temperature permit crop growth.

The suitability estimates were carried out for grid cells at the 5 arc minute level (9.3 by 9.3 km at the equator), by interfacing the soil, terrain and LGP characteristics for each grid cell with specific growth requirements (temperature profile, moisture, nutrients, etc.) for each of the 30 crops under three levels of technology. These levels of technology are: low, using no fertilizers, pesticides or improved seeds, equivalent to subsistence farming; intermediate, with some use of fertilizers, pesticides, improved seeds and mechanical tools; and high, with full use of all required inputs and management practices as in advanced commercial farming. The resulting average attainable yields for each cell, crop and technology alternative were then compared with those obtainable under the same climate and technology on land without soil and terrain constraints, termed here the maximum constraint-free yield (MCFY). The land in each grid cell is for each crop (and technology level) subdivided into five suitability classes on the basis of the average attainable yield as a percentage of the MCFY, as follows - very suitable (VS): at least 80 percent; suitable (S): 60 to 80 percent; moderately suitable (MS): 40 to 60 percent and marginally suitable (mS): 20 to 40 percent. Not suitable (NS) land is that for which attainable yields are below 20 percent of the MCFY.

The result of this procedure is an inventory of land suitability by grid cell for each crop and technology level. To make statements of the overall suitability for rainfed agriculture, one has to aggregate the suitability estimates for all crops and technology levels.2 There are various ways of doing this (e.g. one could add up over crops by applying value (prices) or energy (calories) weights to arrive at an«average» crop). Here the method applied in Fischer, van Velthuizen and Nachtergaele (2000) was followed: For each grid cell, first the largest (i.e. out of all the crops considered) extent of very suitable and suitable area under the high technology level was taken. Then the part of the largest very suitable, suitable and moderately suitable area under the intermediate technology, exceeding this first area, was added. Finally the part of the largest very suitable, suitable, moderately suitable and marginally suitable area under the low technology, exceeding this second area, was added. The rationale for this methodology is that it is unlikely to make economic sense to cultivate moderately and marginally suitable areas under the high technology level, or to cultivate marginally suitable areas under the intermediate technology level. The result of this is the maximum suitable area in each grid cell under what was dubbed the«mixed» input level. Table 4.6 shows the results of aggregating over all grid cells in each country.

It is noted, however, that some of the land classified as not suitable on the basis of this evaluation is used for rainfed agriculture in some countries, e.g. where steep land has been terraced or where yields less than the MCFY are acceptable under the local economic and social conditions (see also Box 4.2). For these reasons, land reported as being in agricultural use in some countries exceeds the areas deemed here as having rainfed crop production potential.

1 These crops are: wheat (2 types), rice (3 types), maize, barley (2 types), sorghum, millet (2 types), rye (2 types), potato, cassava, sweet potato, phaseolus bean, chickpea, cowpea, soybean, rapeseed (2 types), groundnut, sunflower, oil palm, olive, cotton, sugar cane, sugar beet and banana.
2 For a full explanation of the methodology, see Fischer, van Velthuizen and Nachtergaele (2000). The estimate of total potential land of 2603 million ha in developing countries, excluding China, is about 3 percent higher than the 2537 million ha estimated for the 1995 edition of this study (Alexandratos, 1995). This is due to the more refined methodology followed in the present study, new climate and terrain data sets, the increase in the number of crops for which suitability was tested (from 21 in 1995 to 30 in the present study), and a different method of aggregation.

First, the method of deriving the land suitability estimates: it is enough for a piece of land to support a single crop at a minimum yield level for it to be deemed suitable. For example, large tracts of land in North Africa permit cultivation of only olive trees. These lands therefore are counted as«suitable» although one might have little use for them in practice (see also Box 4.2 for further similar qualifications).

Box 4.2 Estimating the land potential for rainfed agriculture: some observations 1

The evaluation of land potential undertaken in the global agro-ecological zones (GAEZ) study starts by taking stock of (i) the biophysical characteristics of the resource (soil, terrain, climate); and (ii) the growing requirements of crops (solar radiation, temperature, humidity, etc.). The data in the former set are interfaced with those in the second set and conclusions are drawn on the amount of land that may be classified as suitable for producing each one of the crops tested (see Fischer, van Velthuizen and Nachtergaele, 2000).

The two data sets mentioned above can change over time. Climate change, land degradation or, conversely, land improvements, together with the permanent conversion of land to non-agricultural uses, all contribute to change the extent and characteristics of the resource. This fact is of particular importance if the purpose of the study is to draw inferences about the adequacy of land resources in the longer term.

In parallel, the growth of scientific knowledge and the development of technology modify the growing requirements of the different crops for achieving any given yield level. For example, in the present round of GAEZ work the maximum attainable yield for rainfed wheat in subtropical and temperate environments is put at about 12 tonnes/ha in high input farming and about 4.8 tonnes/ha in low input farming. Some 25 years ago, when the first FAO agro-ecological zone study was carried out (FAO, 1981b), these yields were put at only 4.9 and 1.2 tonnes/ha, respectively. Likewise, land suitable for growing wheat at, say, 5 tonnes/ha in 30 years time may be quite different from that prevailing today, if scientific advances make it possible to obtain such yields where only 2 tonnes/ha can be achieved today. A likely possibility would be through the development of varieties better able to withstand stresses such as drought, soil toxicity and pest attack. Scientific knowledge and its application will obviously have an impact on whether or not any given piece of land will be classified as suitable for producing a given crop.

Land suitability is crop-specific. To take an extreme example, more than 50 percent of the land area in the Democratic Republic of the Congo is suitable for growing cassava but less than 3 percent is suitable for growing wheat. Therefore, before statements can be made about the adequacy or otherwise of land resources to grow food for an increasing population, the information about land suitability needs to be interfaced with information about expected demand patterns - volume and commodity composition of both domestic and foreign demand. For example, the Democratic Republic of the Congo's ample land resources suitable for growing cassava will be of little value unless there is sufficient domestic or foreign demand for the country's cassava, now or in the future.

Declaring a piece of land as suitable for producing a certain crop implicitly assumes that people find it worthwhile to exploit the land for this purpose. In other words, land must not only possess minimum biophysical attributes in relation to the requirements of the crops for which there is, or will be, demand, but it must also be in a socio-economic environment in which people consider it an economic asset. For example, in low-income countries, people will exploit land even if the yields or, more precisely, the returns to their work, are low relative to the urgency to secure their access to food. This means that the price of food is high relative to their income and that the opportunities of earning higher returns from other activities are limited as well. Thus, what qualifies as land with an acceptable production potential in a poor country may not be so in a high-income one. An exception would be if poor quality of land were compensated by a larger area per person with access to mechanization2 so that returns to work in farming would generate income not far below earnings from other work. Obviously, the socio-economic context within which a piece of land exists and assumes a given value or utility, changes over time: what qualifies today as land suitable for farming may not be so tomorrow.

It is no easy task to account fully for all these factors in arriving at conclusions concerning how much land with crop production potential there is. For example, if food became scarce and its real price rose, more land would be worth exploiting and hence be classified as agricultural than would otherwise be the case. Therefore, depending on how such information is to be used, one may want to adopt different criteria and hence generate alternative estimates.

1 Adapted from Alexandratos and Bruinsma (1999).
2 Relatively low-yield rainfed but internationally competitive agriculture (wheat yields of 2.0-2.5 tonnes/ha compared with double that in western Europe) is practised in such high-income countries as the United States, Canada and Australia. But this is in large and fully mechanized farms permitting the exploitation of extensive areas that generate sufficient income per holding even if earnings per ha are low.

Second, the land balance (land with crop production potential not in agricultural use) is very unevenly distributed among regions and countries. Some 90 percent of the remaining 1.8 billion ha is in Latin America and sub-Saharan Africa, and more than half of the total is concentrated in just seven countries (Brazil, the Democratic Republic of the Congo, the Sudan, Angola, Argentina, Colombia and Bolivia). At the other extreme, there is virtually no spare land available for agricultural expansion in South Asia and the Near East/North Africa. In fact, in a few countries in these two latter regions, the land balance is negative, i.e. land classified as not suitable is made productive through human intervention such as terracing of sloping land, irrigation of arid and hyperarid land, etc. and is in agricultural use. Even within the relatively land-abundant regions, there is great diversity of land availability, in terms of both quantity and quality, among countries and subregions.

Table 4.6: Land with rainfed crop production potential

 

Total land surface

Share of land suitable (%)

Total land suitable

Very suitable

Suitable

Moderately suitable

Marginally suitable

Not suitable

Million ha

Developing countries

7302

38

2782

1109

1001

400

273

4520

Sub-Saharan Africa

2287

45

1031

421

352

156

103

1256

Near East/North Africa

1158

9

99

4

22

41

32

1059

Latin America and the Caribbean

2035

52

1066

421

431

133

80

969

South Asia

421

52

220

116

77

17

10

202

East Asia

1401

26

366

146

119

53

48

1035

Industrial countries

3248

27

874

155

313

232

174

2374

Transition countries

2305

22

497

67

182

159

88

1808

World*

13400

31

4188

1348

1509

794

537

9211

* Including some countries not covered in this study.

Third, much of the land also suffers from constraints such as ecological fragility, low fertility, toxicity, high incidence of disease or lack of infrastructure. These reduce its productivity, require high input use and management skills to permit its sustainable use, or require prohibitively high investments to be made accessible or disease-free. Alexandratos (1995, Table 4.2) shows that over 70 percent of the land with rainfed crop production potential in sub-Saharan Africa and Latin America suffers from one or more soil and terrain constraints. Natural causes as well as human intervention can also lead to deterioration of the productive potential of the resource, for example through soil erosion or salinization of irrigated areas. Hence this evaluation of suitability may contain elements of overestimation (see also Bot, Nachtergaele and Young, 2000) and much of the land balance cannot be considered to be a resource that is readily usable for food production on demand.

There is another cause for the land balance to be overestimated: it ignores land uses other than for growing the crops for which it was evaluated. Thus, forest cover, protected areas and land used for human settlements and economic infrastructure are not taken into account. Alexandratos (1995) estimated that forests cover at least 45 percent, protected areas some 12 percent and human settlements some 3 percent of the land balance, with wide regional differences. For example, in the land-scarce region of South Asia, some 45 percent of the land with crop production potential but not yet in agricultural use is estimated to be occupied by human settlements. This leaves little doubt that population growth and further urbanization will be a significant factor in reducing land availability for agricultural use in this region.

These considerations underline the need to interpret estimates of land balances with caution when assessing land availability for agricultural use. Cohen (1995) summarizes and evaluates all estimates made of available cultivable land, together with their underlying methods, and shows their extremely wide range. Young (1999) offers a critique of the more recent estimates of available cultivable land, including those given in Alexandratos (1995), and states that«an order-of-magnitude estimate reaches the conclusion that in a representative area with an estimated land balance of 50 percent, the realistic area is some 3 to 25 percent of the cultivable land».

4.3.2 Expansion of land in crop production

There is a widespread perception that there is no more, or very little, new land to bring under cultivation. Some of this perception may be well grounded in the specific situations of land-scarce countries and regions such as Japan, South Asia and the Near East/North Africa. Yet this perception may not apply, or may apply with much less force, to other parts of the world. As discussed above, there are large tracts of land with varying degrees of agricultural potential in several countries, most of them in sub-Saharan Africa and Latin America, with some in East Asia. However, this land may lack infrastructure, be partly under forest cover or be in wetlands that have to be protected for environmental reasons, or the people who would exploit it for agriculture lack access to appropriate technological packages or the economic incentives to adopt them.

In reality, expansion of land in agricultural use takes place all the time. It does so mainly in countries that combine growing needs for food and employment with limited access to technology packages that could increase intensification of cultivation on land already in agricultural use. The data show that expansion of arable land continues to be an important source of agricultural growth in sub-Saharan Africa, South America and East Asia, excluding China (Table 4.7).3

The projected expansion of arable land in crop production shown in Tables 4.7, 4.8 and 4.9, has been derived for the rainfed and irrigated land classes. In each country the following factors have been taken into account: (i) actual data or, in many cases, estimates for the base year 1997/99 on harvested land and yield by crop in each of the two classes; (ii) total arable land and cropping intensity in each class; (iii) production projections for each crop; (iv) likely increases in yield by crop and land class; (v) increases in the irrigated area; (vi) likely increases in cropping intensities; and (vii) the land balances for rainfed agriculture described in the preceding section, and for irrigated land discussed in the following section. This method was used only for the 93 developing countries covered in this study (see Appendix 1). For the developed countries or country groups, only projections of crop production have been made, which were then translated into projections for total harvested land and yield by crop.

The overall result for developing countries is a projected net increase in the arable area of 120 million ha (from 956 in the base year to 1076 in 2030), an increase of 12.6 percent (see Table 4.7).4 The increase for the period 1961/63 to 1997/99 was 172 million ha, an increase of 25 percent. Not surprisingly, the bulk of this projected expansion is expected to take place in sub-Saharan Africa (60 million), Latin America (41 million) and East Asia, excluding China (14 million), with almost no land expansion in the Near East/North Africa and South Asia regions and even a decline in the arable land area in China. The slowdown in the expansion of arable land is mainly a consequence of the projected slowdown in the growth of crop production and is common to all regions.

Table 4.7: Total arable land: past and projected

 

Arable land in use

Annual growth

Land in use asm % of potential

Balance

1961/63

1979/81

1997/99

1997/99 adj.

2015

2030

1961-1999

1997/99-2030

1997/99

2030

1997/99

2030

(million ha)

(% p.a.)

(%)

(million ha)

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

Sub-Saharan Africa

119

138

156

228

262

288

0.77

0.72

22

28

803

743

Near East/ North Africa

86

91

100

86

89

93

0.42

0.23

87

94

13

6

Latin America and the Caribbean

104

138

159

203

223

244

1.22

0.57

19

23

863

822

South Asia

191

202

205

207

210

216

0.17

0.13

94

98

13

4

    excl. India

29

34

35

37

38

39

0.37

0.12

162

168

-14

-16

East Asia

176

182

227

232

233

237

0.89

0.06

63

65

134

129

   excl. China

72

82

93

98

105

112

0.82

0.43

52

60

89

75

Developing countries

676

751

848

956

1017

1076

0.68

0.37

34

39

1826

1706

   excl. China

572

652

713

822

889

951

0.63

0.46

32

37

1781

1652

   excl. China and India

410

483

543

652

717

774

0.81

0.54

27

32

1755

1633

Industrial countries

379

395

387

     

0.07

 

44

 

487

 

Transition countries

291

280

265

     

-0.19

 

53

 

232

 

World

1351

1432

1506

     

0.34

 

36

 

2682

 

Source: Column (1)-(3): FAOSTAT, November 2001.
Note:«World» includes a few countries not included in the other country groups shown.

The projected increase of arable land in agricultural use is a small proportion (6.6 percent) of the total unused land with rainfed crop production potential. What does the empirical evidence show concerning the rate and process of land expansion for agricultural use in the developing countries? Microlevel analyses have generally established that under the socio-economic and institutional conditions (land tenure, etc.) prevailing in many developing countries, increases in output are obtained mainly through land expansion, where the physical potential for doing so exists. For example, in a careful analysis of the experience of Côte d'Ivoire, Lopez (1998) concludes that«the main response of annual crops to price incentives is to increase the area cultivated». Similar findings, such as the rate of deforestation being positively related to the price of maize, are reported for Mexico by Deininger and Minten (1999). Some of this land expansion is taking place at the expense of long rotation periods and fallows, a practice still common to many countries in sub-Saharan Africa, with the result that the natural fertility of the soil is reduced. Since fertilizer use is often uneconomic, the end result is soil mining and stagnation or outright reduction of yields.

The projected average annual increase in the developing countries' arable area of 3.75 million ha (120/32), compared with 4.8 million (172/36) in the historical period, is a net increase. It is the total of gross land expansion minus land taken out of production for various reasons, for example because of degradation or loss of economic viability. An unknown part of the new land to be brought into agriculture will come from land currently under forests. If all the additional land came from forested areas, this would imply an annual deforestation rate of 0.2 percent, compared with the 0.8 percent (or 15.4 million ha p.a.) for the 1980s and 0.6 percent (or 12.0 million ha p.a.) for the 1990s (FAO, 2001c). The latter estimates, of course, include deforestation from all causes, such as informal non-recorded agriculture, grazing, logging, gathering of fuelwood, etc.

The arable area in the world as a whole expanded between 1961/63 and 1997/99 by 155 million ha (or 11 percent), the result of two opposite trends: an increase of 172 million ha in the developing countries and a decline of 18 million ha in the developed ones. This decline in the arable area in the latter group has been accelerating over time (-0.3 percent p.a. in the industrial countries and -0.6 percent p.a. in transition countries during 1989-99). The longer-term forces determining such declines are sustained yield growth combined with a continuing slowdown in the growth of demand for their agricultural products. In addition, there are more temporary phenomena such as policy changes in the industrial countries and political and economic transition problems in the former centrally planned countries. No projections were made for arable land in the developed countries but, assuming a continuation of these trends, one would expect a further decline in the developed countries' arable area. However, this decline in arable area could in part be offset by the emerging trend towards a de-intensification of agriculture in these countries through increasing demand for organic products and for environmentally benign cultivation practices, and a possible minor shift of agriculture to temperate zones towards the end of the projection period because of climate change. The net effect of these countervailing forces could be a roughly constant or only marginally declining arable area in the developed countries. Arable area expansion for the world as a whole therefore would more or less equal that of the developing countries.

Although the developing countries' arable area is projected to expand by 120 million ha over the projection period, the harvested area will expand by 178 million ha or 20 percent, because of increases in cropping intensities (Table 4.8). The increase of harvested land over the historical period (1961/63 to 1997/99) was 221 million ha or 38 percent. Sub-Saharan Africa alone accounts for 63 million ha, or 35 percent, of the projected increase in harvested land, the highest among all regions. This is a consequence of the high and sustained growth in crop production projected for this region (see Table 4.1) combined with the region's scope for further land expansion. The other region for which a considerable expansion of the harvested area is foreseen, albeit at a slower pace than in the past, is Latin America with an increase of 45 million ha.

Table 4.8: Arable land in use, cropping intensities and harvested land

 

Total land in use

Rainfed use

Irrigated use

A

CI

H

A

CI

H

A

CI

H

Sub-Saharan Africa

1997/99

228

68

154

223

67

150

5

86

4.5

2030

288

76

217

281

75

210

7

102

7

Near East/North Africa

1997/99

86

81

70

60

72

43

26

102

27

2030

93

90

83

60

78

46

33

112

37

Latin America and the Caribbean

1997/99

203

63

127

185

60

111

18

86

16

2030

244

71

172

222

68

150

22

100

22

South Asia

1997/99

207

111

230

126

103

131

81

124

100

2030

216

121

262

121

109

131

95

137

131

East Asia

1997/99

232

130

303

161

120

193

71

154

110

2030

237

139

328

151

122

184

85

169

144

All above

1997/99

956

93

885

754

83

628

202

127

257

2030

1076

99

1063

834

87

722

242

141

341

    excl. China

1997/99

822

83

679

672

76

508

150

114

171

2030

951

90

853

769

81

622

182

127

230

   excl. China/India

1997/99

652

75

489

559

70

392

93

105

97

2030

774

83

641

662

77

507

112

119

134

Note: A=arable land in million ha; CI=cropping intensity in percentage; H=harvested land in million ha.

As mentioned before, the quality of the data for arable land use leaves much to be desired (see also Young, 1998). The data of harvested or sown areas for the major crops are more reliable. They show that expansion of harvested area continues to be an important source of agricultural growth, mainly in sub-Saharan Africa, but also in Southeast Asia and, to a lesser extent, in Latin America. Overall, for the developing countries, excluding China, the harvested area under the major crops (cereals, oilseeds, pulses, roots/tubers, cotton, sugar cane/ beet, rubber and tobacco) grew by 10 percent during the ten years from 1987/89 to 1997/99, or about 1 percent p.a. This is only slightly higher than the growth rate of 0.9 percent p.a. projected for all crops for the 21-year period 1988/90-2010 in Alexandratos, 1995 (p. 165).

The overall cropping intensity for developing countries will rise by about 6 percentage points over the projection period (from 93 to 99 percent). Cropping intensities continue to rise through shorter fallow periods and more multiple cropping. An increasing share of irrigated land in total agricultural land contributes to more multiple cropping. About one-third of the arable land in South and East Asia is irrigated, a share which is projected to rise to 40 percent in 2030. The high irrigation share is one of the reasons why the average cropping intensities in these regions are considerably higher than in the other regions. Average cropping intensities in developing countries, excluding China and India which together account for more than half of the irrigated area in the developing countries, are and will continue to be much lower.

The rise in cropping intensities has been one of the factors responsible for increasing the risk of land degradation and threatening sustainability, when it is not accompanied by technological change to conserve the land, including adequate and balanced use of fertilizers to compensate for soil nutrient removal by crops. It is expected that this risk will continue to exist because in many cases the socio-economic conditions will not favour the promotion of the technological changes required to ensure the sustainable intensification of land use (see Chapter 12 for a further discussion of this issue).


continued


1 See Alexandratos (1995, p. 161, 168) for a discussion on problems with land use data.
2 Unfortunately, revised data for harvested land and yields by crop for China (mainland) are not available until the results of the 1997 Chinese Agricultural Census have been processed and published. Therefore, ad hoc adjustments had to be made to base year data based on fragmentary non-official information on harvested land and yield by crop.
3 Historical data for China have been drastically revised upwards from 1985 onwards, which distorts the historical growth rates in Table 4.7 for East Asia (and for the total of developing countries).
4As mentioned in Section 4.2, data on arable land are unreliable for many countries. Therefore base year data were adjusted and are shown in column (4) as«1997/99 adjusted».


Previous PageTop Of PageNext Page