4.3.3 Is land for agriculture becoming scarcer?⁵

As noted in the preceding section, land in agricultural use (arable land and land under permanent crops) in the world as a whole has increased by only 155 million ha or 11 percent to about 1.5 billion ha between the early 1960s and the late 1990s. Nevertheless there were very significant changes in some regions. For example, the increase was over 50 percent in Latin America, which accounted for over one-third of the global increase. During the same period, the world population nearly doubled from 3.1 billion to over 5.9 billion. By implication, arable land per person declined by 40 percent, from 0.43 ha in 1961/63 to 0.26 ha in 1997/99. In parallel, there is growing preoccupation that agricultural land is being lost to non-agricultural uses. In addition, the ever more intensive use of land in production through multiple cropping, reduced fallow periods, excessive use of agrochemicals, spread of monocultures, etc. is perceived as leading to land degradation (soil erosion, etc.) and the undermining of its long-term productive potential.

These developments are seen by many as having put humanity on a path of growing scarcity of land as a factor in food production, with the implication that it is, or it will be in the near future, becoming increasingly difficult to produce the food required to feed the ever-growing human population. Are these concerns well founded? Any discourse about the future should be as precise as possible concerning the magnitudes involved: how much land there is (quantity, quality, location) and how much more food, what type of food and where it is required, now or at any given point of time in the future. The brief discussion of historical developments and, in particular, of future prospects in world food and agriculture presented in Chapters 2 and 3, provides a rough quantitative framework for assessing such concerns.

The evidence presented above about historical developments does not support the notion that it has been getting increasingly difficult for the world to extract from the land an additional unit of food. Rather the contrary has been happening, as shown by the secular decline in the real price of food. This secular decline indicates that it has been getting easier for humanity to produce an additional unit of food relative to the effort required to produce an additional unit of an«average» non-food product. This statement applies to the world as a whole, not necessarily to particular locations, and is valid only under particular conditions which are, essentially, the absence of market failures and ethical acceptability of the resulting distribution of access to food by different population groups.

The notion that resources for producing food, in which land is an important constituent, have been getting more abundant rather than scarcer in relative terms, i.e. in relation to the aggregate stocks of resources of the global economy, appears counter-intuitive. How can it be reconciled with the stark fact that the world population nearly doubled while land in agricultural use increased by only 11 percent, meaning that land per capita declined by some 40 percent? The answer is to be found in the fact that over the same period yields per ha of cropped area increased, as did the cropping intensity in the areas where a combination of irrigation and agro-ecological conditions permitted it and the growth of the demand for food justified it economically. For example, during the 36-year period when world average grain yields more than doubled from 1.4 tonnes/ha in 1961/63 to 3.05 tonnes/ha in 1997/99 and the overall cropping intensity probably increased by some 5 percentage points, the amount of arable land required to produce any given amount of grain declined by some 56 percent. This decline exceeded the above-mentioned 40 percent fall in the arable land per person which occurred during the same period.

In this comparison of physical quantities, land for food production is seen to have become less scarce, not scarcer. The economic evidence, a declining real price of food, corroborates in a general sense the conclusion that it has also become less scarce relative to the evolution of the demand for food and relative to what has been happening in the other sectors of the economy. However, as noted, such economic evidence properly refers to the decreasing relative scarcity of the aggregate resource base for food production in which land is only one component together with capital, labour, technology, etc. rather than to land alone.⁶ In practice, what we call land today is a composite of land in its natural form and capital investments embodied in it such as irrigation infrastructure, levelling, fencing and soil amendments. It follows that any further discussion of the prospective role of land in meeting future food needs has to view it as just one component, indeed one of changing and probably declining relative weight, in the total package of factors that constitute the resource base of agriculture which, as the historical record shows, is flexible and adaptable.

Concerning the future, a number of projection studies have addressed and largely answered in the positive the issue as to whether the resource base of world agriculture, including its land component, can continue to evolve in a flexible and adaptable manner as it did in the past, and also whether it can continue to exert downward pressure on the real price of food (see, for example, Pinstrup-Andersen, Pandya-Lorch and Rosegrant, 1999). The largely positive answers mean essentially that for the world as a whole there is enough, or more than enough, food production potential to meet the growth of effective demand, i.e. the demand for food of those who can afford to pay farmers to produce it.

The preceding discussion refers to the evidence about land scarcities that can be deduced from the evolution of global magnitudes, whether aggregates such as world population, averages such as world per capita values of key variables, or food price trends observable in world markets. However, observing, interpreting and projecting the evolution of global aggregates can go only part of the way towards addressing the issues often raised in connection with the role of land in food production, essentially those issues pertaining to the broader nexus of food security and the environment. A more complete consideration of the issue, which goes beyond the scope of this report, will require an analysis at a more disaggregated level and going beyond the use of conventional economic indicators of scarcity or abundance. It should also address the following issues. First, whether land availability for food production is likely to become, or has been already, a significant constraint to solving problems of food insecurity at the local level. Second, whether the market signals which tell us that the resources for producing food, land among them, have been getting relatively less scarce, are seriously flawed because they fail to account for the environmental costs and eventual future risks associated with the expansion and intensification of agriculture.

4.4 Irrigation and water use

4.4.1 Expansion of irrigated land

The projections of irrigation presented below reflect a composite of information on existing irrigation expansion plans in the different countries, potentials for expansion and need to increase crop production. The projections include some expansion in informal (community-managed) irrigation, which is important in sub-Saharan Africa. Estimates of«land with irrigation potential» are notoriously difficult to make for various reasons (see Alexandratos, 1995, p. 160-61) and should be taken as only rough orders of magnitude.⁷

The aggregate result for the group of developing countries shows that the area equipped for irrigation in this group of countries will expand by 40 million ha (20 percent) over the projection period (Table 4.9). This means that some 20 percent of the land with irrigation potential not yet equipped at present will be brought under irrigation, and that 60 percent of all land with irrigation potential (403 million ha) would be in use by 2030.

The expansion of irrigation will be strongest (in absolute terms) in the more land-scarce regions hard-pressed to raise crop production through more intensive cultivation practices, such as South Asia (+14 million ha), East Asia (+14 million ha) and the Near East/North Africa. Only small additions will be made in the more land-abundant regions of sub-Saharan Africa and Latin America, although they may represent an important increase in relative terms. The importance of irrigated agriculture has already been discussed in Section 4.2. Because of a continuing increase in cropping intensity on both existing and newly irrigated areas, the harvested irrigated area will expand by 84 million ha and will account for almost half of the increase in all harvested land (Table 4.8).

The projected expansion of irrigated land by 40 million ha is an increase in net terms. It assumes that losses of existing irrigated land resulting from, for example, water shortages or degradation because of salinization, will be compensated through rehabilitation or substitution by new areas for those lost. The few existing historical data on such losses are too uncertain and anecdotal to provide a reliable basis for drawing inferences about the future. However, if it is assumed that 2.5 percent of existing irrigation must be rehabilitated or substituted by new irrigation each year, that is, if the average life of irrigation schemes were 40 years, then the total irrigation investment activity over the projection period in the developing countries must encompass some 200 million ha, of which four-fifths would be for rehabilitation or substitution and the balance for net expansion.

The projected net increase in arable irrigated land of 40 million ha is less than half of the increase over the preceding 36 years (100 million ha). In terms of annual growth it would be«only» 0.6 percent, well below the 1.9 percent for the historical period. The projected slowdown reflects the projected lower growth rate of crop production combined with the increasing scarcity of suitable areas for irrigation and of water resources in some countries, as well as the rising costs of irrigation investment.

Most of the expansion of irrigated land is achieved by converting land in use in rainfed agriculture or land with rainfed production potential but not yet in use, into irrigated land. Part of the irrigation, however, takes place on arid and hyperarid land which is not suitable for rainfed agriculture. It is estimated that of the 202 million ha irrigated at present, 42 million ha are on arid and hyperarid land and of the projected increase of 40 million ha, about 2 million ha will be on such land. In some regions and countries, irrigated arid and hyperarid land form an important part of the total irrigated land at present in use: 18 out of 26 million ha in the Near East/North Africa, and 17 out of 81 million ha in South Asia.

The developed countries account for a quarter of the world's irrigated area, 67 out of 271 million ha (Table 4.9). Their annual growth of irrigated area reached a peak of 3.0 percent in the 1970s, dropping to 1.1 percent in the 1980s and to only 0.3 percent in 1990-99. This evolution pulled down the annual growth rate for global irrigation from 2.4 percent in the 1970s to 1.3 percent in the 1980s and 1990-99. Perhaps it is this sharp deceleration in growth which led some analysts to believe that there is only limited scope for further irrigation expansion. As already said, no projections by land class (rainfed, irrigated) were made for the developed countries. However, given the share of developing countries in world irrigation and the much higher crop production growth projected for this group of countries, it is reasonable to assume that the world irrigation scene will remain dominated by events in the developing countries.

4.4.2 Irrigation water use and pressure on water resources

One of the major questions concerning the future of irrigation is whether there will be sufficient freshwater to satisfy the growing needs of agricultural and non-agricultural users. Agriculture already accounts for about 70 percent of the freshwater withdrawals in the world and is usually seen as the main factor behind the increasing global scarcity of freshwater.

The estimates of the expansion of land under irrigation presented in the preceding section in part provide an answer to this question. The assessment of irrigation potential already takes into account water limitations and the projections to 2030 assume that agricultural water demand will not exceed available water resources. Yet, as discussed above, the concept of irrigation potential has severe limitations and estimates of irrigation potential can vary over time, in relation to the country's economic situation or as a result of competition for water for domestic and industrial use. Estimates of irrigation potential are also based on renewable water resources, i.e. the resources replenished annually through the hydrological cycle. In those arid countries where mining of fossil groundwater represents an important part of water withdrawal, the area under irrigation is usually larger than the irrigation potential.

Renewable water resources available to irrigation and other uses are commonly defined as that part of precipitation which is not evaporated or transpired by plants, including grass and trees, which flows into rivers and lakes or infiltrates into aquifers. The annual water balance for a given area in natural conditions, i.e. without irrigation, can be defined as the sum of the annual precipitation and net incoming flows (transfers through rivers from one area to another) minus evapotranspiration.

Table 4.10 shows the renewable water resources for 93 developing countries. Average annual precipitation is around 1040 mm. In developing regions, renewable water resources vary from 18 percent of precipitation and incoming flows in the most arid areas (Near East/North Africa) where precipitation is a mere 180 mm per year, to about 50 percent in humid East Asia, which has a high precipitation of about 1250 mm per year. Renewable water resources are most abundant in Latin America. These figures give an impression of the extreme variability of climatic conditions facing the 93 developing countries, and the ensuing differences observed in terms of water scarcity: those countries suffering from low precipitation and therefore most in need of irrigation are also those where water resources are naturally scarce. In addition, the water balance presented is expressed in yearly averages and cannot adequately reflect seasonal and interannual variations. Unfortunately, such variations tend to be more pronounced in arid than in humid climates.

The first step in estimating the pressure of irrigation on water resources is to assess irrigation water requirements and withdrawals. Precipitation provides part of the water crops need to satisfy their transpiration requirements. The soil, acting as a buffer, stores part of the precipitation water and returns it to the crops in times of deficit. In humid climates, this mechanism is usually sufficient to ensure satisfactory growth in rainfed agriculture. In arid climates or during the dry season, irrigation is required to compensate for the deficit resulting from insufficient or erratic precipitation. Consumptive water use in irrigation therefore is defined as the volume of water needed to compensate for the deficit between potential evapotranspiration and effective precipitation over the growing period of the crop. It varies considerably with climatic conditions, seasons, crops and soil types. In this study, consumptive water use in irrigation has been computed for each country on the basis of the irrigated and harvested areas by crop as estimated for the base year (1997/99) and projected for 2030 (see Box 4.3 for a brief explanation of the methodology applied). As mentioned before, in this study the breakdown by crop over rainfed and irrigated land was performed only for the 93 developing countries.

However, it is water withdrawal for irrigation, i.e. the volume of water extracted from rivers, lakes and aquifers for irrigation purposes, which should be used to measure the impact of irrigation on water resources. Irrigation water withdrawal normally far exceeds the consumptive water use in irrigation because of water lost during transport and distribution from its source to the crops. In addition, in the case of rice irrigation, additional water is used for paddy field flooding to facilitate land preparation and for plant protection.

For the purpose of this study, irrigation efficiency has been defined as the ratio between the estimated consumptive water use in irrigation and irrigation water withdrawal. Data on country water withdrawal for irrigation has been collected in the framework of the AQUASTAT programme (see FAO, 1995b, 1997a, 1997b and 1999b). Comparison of these data with the consumptive use of irrigation was used to estimate irrigation efficiency at the regional level. On average, for the 93 developing countries, it is estimated that irrigation efficiency was around 38 percent in 1997/99, varying from 25 percent in areas of abundant water resources (Latin America) to 40 percent in the Near East/North Africa region and 44 percent in South Asia where water scarcity calls for higher efficiencies (Table 4.10).

To estimate irrigation water withdrawal in 2030, an assumption had to be made about possible developments in the irrigation efficiency of each country. Unfortunately, there is little empirical evidence on which to base such an assumption. Two factors, however, will have an impact on the development of irrigation efficiency: the estimated levels of irrigation efficiency in 1997/99 and water scarcity. A function was designed to capture the influence of these two parameters, bearing in mind that improving irrigation efficiency is a very slow and difficult process. The overall result is that efficiency will increase by 4 percentage points, from 38 to 42 percent (Table 4.10). Such an increase in efficiency would be more pronounced in water-scarce regions (e.g. a 13 percentage point increase in the Near East/North Africa region) than in regions with abundant water resources (between 0 and 4 percentage points in Latin America, East Asia and sub-Saharan Africa). Indeed, it is expected that, under pressure from limited water resources and competition from other uses, demand management will play an important role in improving irrigation efficiency in water-scarce regions. In contrast, in humid areas the issue of irrigation efficiency is much less relevant and is likely to receive little attention.

For the 93 countries, irrigation water withdrawal is expected to grow by about 14 percent, from the current 2 128 km³/yr to 2420 km³/yr in 2030 (Table 4.10). This increase is low compared to the 33 percent increase projected in the harvested irrigated area, from 257 million ha in 1997/99 to 341 million ha in 2030 (see Table 4.8). Most of this difference is explained by the expected improvement in irrigation efficiency, leading to a reduction in irrigation water withdrawal per irrigated hectare. A small part of this reduction is also a result of changes in cropping patterns for some countries such as China, where a substantial shift in the irrigated area from rice to maize production is expected: irrigation water requirements for rice production are usually twice those for maize.

Irrigation water withdrawal in 1997/99 was estimated to account for only 7 percent of total water resources for the 93 countries (Table 4.10). However, there are wide variations between regions, with the Near East/North Africa region using 53 percent of its water resources in irrigation while Latin America barely uses 1 percent of its resources. At the country level, variations are even higher. Of the 93 countries, ten already used more than 40 percent of their water resources for irrigation in the base year (1997/99), a situation which can be considered critical. An additional eight countries used more than 20 percent of their water resources, a threshold sometimes used to indicate impending water scarcity. Yet the situation should not change drastically over the period of the study, with only two more countries crossing the threshold of 20 percent. If one adds the expected additional water withdrawals needed for non-agricultural use, the picture will not be much different since agriculture represents the bulk of water withdrawal.

Nevertheless, for several countries, relatively low national figures may give an overly optimistic impression of the level of water stress: China, for instance, is facing severe water shortages in the north while the south still has abundant water resources. Already by 1997/99, two countries (the Libyan Arab Jamahariya and Saudi Arabia) used volumes of water for irrigation larger than their annual renewable water resources. Groundwater mining also occurs in parts of several other countries of the Near East, South and East Asia, Central America and in the Caribbean, even if at the national level the water balance may still be positive. In a survey of irrigation and water resources in the Near East region (FAO, 1997c), it was estimated that the amount of water required to produce the net amount of food imported in the region in 1994 would be comparable to the total annual flow of the Nile river at Aswan.

In concluding this discussion on irrigation, for the 93 developing countries as a whole, irrigation currently represents a relatively small part of their total water resources and there remains a significant potential for further irrigation development. With the relatively small increase in irrigation water withdrawal expected between 1997/99 and 2030, this situation will not change much at the aggregate level. Locally and in some countries, however, there are already very severe water shortages, in particular in the Near East/North Africa region.

4.5 Land-yield combinations for major crops

As discussed in Section 4.2, it is expected that growth in crop yields will continue to be the mainstay of crop production growth, accounting for nearly 70 percent of the latter in developing countries. Although the marked deceleration of crop production growth foreseen for the future (Table 4.1) points to a similar deceleration in growth of yields, such growth will continue to be needed. Questions often asked are: will yield increases continue to be possible? and what is the potential for a continuation of such growth? There is a realization that the chances of a new green revolution or of one-off quantum jumps in yields are now rather limited. There is even a belief that for some major crops, yield ceilings have been, or are rapidly being reached. At the same time, empirical evidence has shown that the cumulative gains in yields over time resulting from slower, evolutionary annual increments in yields have been far more important than quantum jumps in yields, for all major crops (see Byerlee, 1996).

In the following sections, the land-yield combinations underlying the production projections for major crops will first be discussed. Subsequently some educated guesses will be made about the potential for raising yields and for narrowing existing yield gaps.

4.5.1 Harvested land and yields for major crops

As explained in Section 4.3.2, for the developing countries the production projections for the 34 crops of this study⁸ are unfolded into and tested against what FAO experts think are«feasible» land-yield combinations by agro-ecological rainfed and irrigated environment, taking into account whatever knowledge is available. Major inputs into this evaluation are the estimates regarding the availability of land suitable for growing crops in each country and each agro-ecological environment, which come from the FAO agro-ecological zones work (see Section 4.3.1). In practice they are introduced as constraints to land expansion but they also act as a guide to what can be grown where. It is emphasized that the resulting land and yield projections, although they take into account past performance, are not mere extrapolations of historical trends since they take into account all present knowledge about changes expected in the future. Box 4.4 shows an example of the results, tracked against actual outcomes.

The findings of the present study indicate that in developing countries, as in the past but even more so in the future, the mainstay of production increases will be the intensification of agriculture in the form of higher yields and more multiple cropping and reduced fallow periods. This situation will apply particularly in the countries with appropriate agro-ecological environments and with little or no potential of bringing new land into cultivation. The overall result for yields of all the crops covered in this study (aggregated with standard price weights) is roughly a halving of the average annual rate of growth over the projection period as compared to the historical period: 1.0 percent p.a. during 1997/99 to 2030 against 2.1 percent p.a. during 1961-99. This slowdown in the yield growth is a gradual process which has been under way for some time and is expected to continue in the future. It reflects the deceleration in crop production growth explained earlier.

Discussing yield growth at this level of aggregation however is not very helpful, but the overall slowdown is a pattern common to most crops covered in this study with only a few exceptions such as pulses, citrus and sesame. These are crops for which a strong demand is foreseen in the future or which are grown in land-scarce environments. The growth in soybean area and production in developing countries has been remarkable, mainly as a result of explosive growth in Brazil and, more recently, in India (Table 4.11). Soybean is expected to continue to be one of the most dynamic crops, albeit with its production increasing at a more moderate rate than in the past, bringing by 2030 the developing countries' share in world soybean production to 58 percent, with Brazil, China and India accounting for three-quarters of their total.

For cereals, which occupy 58 percent of the world's harvested area and 55 percent in developing countries (Table 4.11), the slowdown in yield growth would be particularly pronounced: down from 2.1 to 0.9 percent p.a. at the world level and from 2.5 to 1.0 percent p.a. in developing countries (Table 4.12). Again this slowdown has been under way for quite some time. The differences of sources of growth and some regional aspects of the various cereal crops have been discussed in Section 4.2. Suffice it here to note that irrigated land is expected to play a much more important role in increasing maize production, almost entirely because of China which accounts for 45 percent of the developing countries' maize production and where irrigated land allocated to maize could more than double. Part of the continued, if slowing, growth in yields is a result of a rising share of irrigated production, with normally much higher cereal yields, in total production. This fact alone would lead to yield increases even if rainfed and irrigated cereal yields did not grow at all.⁹

It is often asserted (see, for example, Borlaug, 1999) that thanks to increases in yield, land has been saved with diminished pressure on the environment as a result, such as less deforestation than otherwise would have taken place. To take cereals as an example, the reasoning is as follows. If the average global cereal yield had not grown since 1961/63 when it was 1405 kg/ha, 1483 million ha would have been needed to grow the 2084 million tonnes of cereals produced in the world in 1997/99. This amount was actually obtained on an area of only 683 million ha at an average yield of 3050 kg/ha. Therefore, 800 million ha (1483 minus 683) have been saved because of yield increases for cereals alone. This conclusion should be qualified, however; had there been no yield growth, the most probable outcome would have been much lower production because of lower demand resulting from higher prices of cereals, and somewhat more land under cereals. Furthermore, in many countries the alternative of land expansion instead of yield increases does not exist in practice.

continued

⁵Adapted from Alexandratos and Bruinsma (1999).
⁶The role of agricultural land as a resource contributing to human welfare, as the latter is conventionally measured by GDP, has been on the decline. Johnson (1997) says that«agricultural land now accounts for no more than 1.5 percent of the resources of the industrial nations».
⁷ FAO (1997a) states concerning such estimates:«Irrigation potential: area of land suitable for irrigation development (it includes land already under irrigation). Methodologies used in assessing irrigation potential vary from one country to another. In most cases, it is computed on the basis of available land and water resources, but economic and environmental considerations are often taken into account to a certain degree. Except in a few cases, no consideration is given to the possible double counting of water resources shared by several countries, and this may lead to an overestimate of irrigation potential at the regional level. Wetlands and floodplains are usually, but not always, included in irrigation potential».
⁸For the analysis of production, the commodities sugar and vegetable oil are unfolded into their constituent crops (sugar cane, beet, soybeans, sunflower, groundnuts, rapeseed, oil palm, coconuts, sesame seed, etc.), so that land-yield combinations are generated for 34 crops.
⁹This is seen most clearly for rice in developing countries, excluding China (Table 4.12) where the growth in the overall average yield of rice exceeds that of rainfed and irrigated rice. This is because the rainfed rice area is projected to remain about the same but the irrigated area is projected to increase by about one-third.