The use of water in agriculture
For vegetative growth and development, plants require, within reach of their roots, water of adequate quality, in appropriate quantity and at the right time. Most of the water a plant absorbs performs the function of raising dissolved nutrients from the soil to the aerial organs, from where it is released to the atmosphere by transpiration: agricultural water use is intrinsically consumptive. Crops have specific water requirements, and these vary depending on local climatic conditions. Whereas an indicative figure for producing one kilogram of wheat is about 1000 litres of water that is returned to the atmosphere, paddy rice may require twice this amount. The production of meat requires between six and twenty times more water than for cereals, depending on the feed/meat conversion factor. Specific values for the water equivalent of a selection of food products are given in Table 3. Water required for human food intake can be derived from these specific values in a grossly approximate way, depending on the size and composition of the meals (see Box 1).
Table 3 Water requirement equivalent of main food products
This table gives examples of water required per unit of major food products, including livestock, which consume the most water per unit. Cereals, oil crops, and pulses, roots and tubers consume far less water.
BOX 1 ASSESSING FRESHWATER NEEDS FOR GLOBAL FOOD PRODUCTION
The amount of water involved in food production is significant, and most of it is provided directly by rainfall. A rough calculation of global water needs for food production can be based on the specific water requirements to produce food for one person. Depending on the composition of meals and allowing for post-harvest losses, the present average food ingest of 2 800 kcal/person/day may require roughly 1 000 m3 per year to be produced. Thus, with a world population of 6 billion, water needed to produce the necessary food is 6 000 km3 (excluding any conveyance losses associated with irrigation systems). Most water used by agriculture stems from rainfall stored in the soil profile and only about 15 percent of water for crops is provided through irrigation. Irrigation therefore needs 900 km3 of water per year for food crops (to which some water must be added for non-food crops). On average, about 40 percent of water withdrawn from rivers, lakes and aquifers for agriculture effectively contribute to crop production, the remainder being lost to evaporation, deep infiltration or the growth of weeds. Consequently, the current global water withdrawals for irrigation are estimated to be about 2 000 to 2 500 km3 per year.
Non-irrigated (rainfed) agriculture depends entirely on rainfall stored in the soil profile. This form of agriculture is possible only in regions where rainfall distribution ensures continuing availability of soil moisture during the critical growing periods for the crops. Non-irrigated agriculture accounts for some 60 percent of production in the developing countries. In rainfed agriculture, land management can have a significant impact on crop yields: proper land preparation leading surface runoff to infiltrate close to the roots improves the conservation of moisture in the soil. Various forms of rainwater harvesting can help to retain water in situ. Rainwater harvesting not only provides more water for the crop but can also add to groundwater recharge and help to reduce soil erosion. Other methods are based on collecting water from the local catchment and either relying on storage within the soil profile or else local storage behind bunds or ponds and other structures for use during dry periods. Recently, conservation agriculture practices such as conservation tillage have proven to be effective in improving soil moisture conservation.
The potential to improve non-irrigated yields is restricted where rainfall is subject to large seasonal and interannual variations. With a high risk of yield reductions or complete loss of crop from dry spells and droughts, farmers are reluctant to invest in inputs such as plant nutrients, high-yielding seeds and pest management. For resource-poor farmers in semi-arid regions, the overriding requirement is to harvest sufficient food stuff to ensure nutrition of the household through to the next harvest. This objective may be reached with robust, drought-resistant varieties associated with low yields. Genetic engineering has not yet delivered high-yield drought-resistant varieties, a difficult task to achieve because, for most crop plants, drought resistance is associated with low yields.
In irrigated agriculture, water taken up by crops is partly or totally provided through human intervention. Irrigation water is withdrawn from a water source (river, lake or aquifer) and led to the field through an appropriate conveyance infrastructure. To satisfy their water requirements, irrigated crops benefit from both more or less unreliable natural rainfall, and from irrigation water. Irrigation provides a powerful management tool against the vagaries of rainfall and makes it economically attractive to grow high-yield seed varieties and to apply adequate plant nutrition as well as pest control and other inputs, thus giving room for a boost in yields. Figure 3 illustrates the typical yield response of a cereal crop to water availability and the synergy between irrigation, crop variety and inputs. Irrigation is crucial to the world’s food supplies. In 1998, irrigated land made up about one-fifth of the total arable area in developing countries but produced two-fifths of all crops and close to three-fifths of cereal production.
Figure 3 Typical response to water for cereal crops
The developed countries account for a quarter of the world’s irrigated area (67 million ha). Their annual growth of irrigated area reached a peak of 3 percent in the 1970s and dropped to only 0.2 percent in the 1990s. The population of this group of countries is growing only slowly and therefore a very slow growth in their demand and production of agricultural commodities is foreseen. The focus of irrigation development is consequently expected to be concentrated on the group of developing countries where demographic growth is strong. Increasing competition from the higher valued industrial and domestic sector results in a decrease in the amount of overall water allocated to irrigation. Figure 4 illustrates the case for the Zhanghe irrigation system in China.
Map 1 shows irrigated land as percentage of arable land in developing countries. A high proportion of irrigated land is usually found in countries and regions with an arid or semi-arid climate. However, low proportions of irrigated land in sub-Saharan Africa point also to underdeveloped irrigation infrastructure. Data and projections of irrigated land compared to irrigation potential in developing countries are shown in Figure 5. The irrigation potential figure already takes into account the availability of water. The graph shows that a sizeable part of irrigation potential is already used in the Near East/North Africa region (where water is the limiting factor) and in Asia (where land is often the limiting factor), whereas a large potential is still unused in sub-Saharan Africa and in Latin America.
Figure 4 Competing uses of water in the Zhanghe irrigation district, China
According to FAO forecasts, the share of irrigation in world crop production is expected to increase in the next decades. In particular in developing countries, the area equipped for irrigation is expected to have expanded by 20 percent (40 million ha) by 2030. This means that 20 percent of total land with irrigation potential but not yet equipped will be brought under irrigation, and that 60 percent of all land with irrigation potential (402 million ha) will be in use by 2030. The net increase in irrigated land (40 million ha, 0.6 percent per year) projected to 2030 is less than half the increase over the preceding 36 years (99 million ha, 1.9 percent per year). The projected slowdown in irrigation development reflects the projected lower growth rate of food demand, combined with the increasing scarcity of suitable areas for irrigation and of water resources in some countries, as well as the rising cost of irrigation investment. The first selection of economically attractive irrigation projects has already been implemented, and prices for agricultural commodities have not risen to encourage investment in a second selection of more expensive irrigation projects.
Map 1 Area equipped for irrigation as percentage of cultivated land by country (1998)
Source: FAOSTAT, 2002.
Most of the expansion in 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. The expansion of irrigation is projected to be strongest in South Asia, East Asia and Near East/North Africa. These regions have limited or no potential for expansion of non-irrigated agriculture. Arable land expansion will nevertheless remain an important factor in crop production growth in many countries in sub-Saharan Africa, Latin America and some countries in East Asia, although to a much smaller extent than in the past. The growth in wheat and rice production in the developing countries will increasingly come from gains in yield, while expansion of harvested land will continue to be a major contributor to the growth in production of maize.
Figure 5 Irrigated area as proportion of irrigation potential in developing countries
In many developing countries, investments in irrigated infrastructures have represented a significant share of the overall agricultural budget during the second half of the twentieth century. The unit cost of irrigation development varies with countries and types of irrigated infrastructures, ranging typically from US$1 000 to US$10 000 per hectare, with extreme cases reaching US$25 000 per hectare (these costs do not include the cost of water storage as the cost of dam construction varies on a case-by-case basis). The lowest investment costs in irrigation are in Asia, which has the bulk of irrigation and where scale economies are possible. The most expensive irrigation schemes are found in sub-Saharan Africa, where irrigation systems are usually smaller and developing land and water resources is costly.
In the future, the estimates of expansion in land under irrigation will represent an annual investment of about US$5 billion, but most investment in irrigation, between US$10 and 12 billion per year, will certainly come from the needed rehabilitation and modernization of aging irrigated schemes built during the years 1960-1980. In the 1990s, annual investment in storage for irrigation was estimated at about US$12 billion (WCD, 2000). In the future, the contrasting effects of reduced demand for irrigation expansion and increased unit cost of water storage will result in an annual investment estimated between US$4 and 7 billion in the next thirty years.
Typically, investment figures in irrigation do not include that part of the investment provided by the farmer in land improvement and on-farm irrigation that can represent up to 50 percent of the overall investment. In total, it is estimated that annual investment in irrigated agriculture will therefore range between US$25 and 30 billion, about 15 percent of annual expected investments in the water sector.
Assessing the impact of irrigation on available water resources requires an estimate of total abstraction for the purpose of irrigation from rivers, lakes and aquifers. The volume extracted is considerably greater than the consumptive use for irrigation because of conveyance losses from the withdrawal site to the plant root zone. Water use efficiency is an indicator often used to express the level of performance of irrigation systems from the source to the crop: it is the ratio between estimated plant requirements and the actual water withdrawal.
On average, it is estimated that overall water use efficiency of irrigation in developing countries is about 38 percent. Map 2 shows the importance of agriculture in the countries’ water balance, and Figure 6 shows the expected growth in water abstraction for irrigation from 1998 to 2030. The predictions are based on assumptions about possible improvements in irrigation efficiency in each region. These assumptions take into account that, from the farmer’s perspective, wherever water is abundant and its cost low, the incentives to save water are limited. Conversely, if farmers can profitably irrigate more land using their allocation in an optimum way, irrigation efficiency may reach higher levels.
Map 2 Agricultural water withdrawals as percentage of renewable water resources (1998)
Improving irrigation efficiency is a slow and difficult process that depends in large part on the local water scarcity situation. It may be expensive and requires willingness, know-how and action at various levels. Table 4 shows current and expected water use efficiency for developing countries in 1998 and 2030, as estimated by FAO. The investment and management decisions leading to higher irrigation efficiency are taken and involve irrigation system management and the system-dependent farmers. National water policy may encourage water savings in water-scarce areas by providing incentives and effectively enforcing penalties. When upstream managers cannot ensure conveyance efficiency, there may be no incentives for water users to make efficiency gains. With groundwater, this caveat may not apply since the incentive is generally internalized by the users, and in many cases groundwater users show much greater efficiency than those depending on surface resources. Box 2 provides an overview of different aspects of potential improvements in agricultural water use efficiency.
Irrigation water withdrawal in developing countries is expected to grow by about 14 percent from the current 2 130 km3 per year to 2 420 km3 in 2030. This finding is consistent with the one given in Box 1 earlier but it is based specifically on individual assessments for each developing country. Harvested irrigated area (the cumulated area of all crops during a year) is expected to increase by 33 percent from 257 million ha in 1998 to 341 million ha in 2030. The disproportionate increase in harvested area is explained by expected improvements in irrigation efficiency, which will result in a reduction in gross irrigation water abstraction per ha of crop. A small part of the reduction is due to changes in cropping patterns in China, where consumer preference is causing a shift from rice to wheat production.
Figure 6 Irrigation and water resources: current (1999) and predicted (2030) withdrawals
Table 4 Water use efficiency in 1998 and 2030 (predicted) in 93 developing countries
Source: FAO, 2002.
BOX 2 POTENTIAL FOR IMPROVEMENTS IN AGRICULTURAL WATER USE EFFICIENCY
Global water strategies tend to focus on the need to increase agricultural water use efficiency, reduce wastage and free large amounts of water for other, more productive uses as well as sustaining the environmental services of rivers and lakes. While there is scope for improved use of water in agriculture, these improvements can only be made slowly and are limited by several considerations. First, there are large areas of irrigated agriculture located in humid tropics where water is not scarce and where improved efficiency would not result in any gain in water productivity. Second, water use efficiency is usually computed at the level of the farm or irrigation scheme, but most of the water that is not used by the crops returns to the hydrological system and can be used further downstream. In these conditions, any improvement in water use efficiency at field level translates into limited improvement in overall efficiency at the level of the river basin. Finally, different cropping systems have different potential for improvement in water use efficiency. Typically, tree crops and vegetables are well adapted to the use of localized, highly efficient irrigation technologies, while such equipments are not adapted to cereal or other crops.
While some countries have reached extreme levels of water use for agriculture, irrigation still represents a relatively small part of total water resources of the developing countries. The projected increase in water withdrawal will not significantly alter the overall picture. At the local level, however, there are already severe water shortages, in particular in the Near East / North Africa region and in large parts of Asia.
Of the ninety-three developing countries surveyed by FAO, ten are already using more than 40 percent of their renewable water resources for irrigation, a threshold used to flag the level at which countries are usually forced to make difficult choices between their agricultural and urban water supply sectors. Another eight countries were using more than 20 percent, a threshold that can be used to indicate impending water scarcity. By 2030 South Asia will have reached the 40 percent level, and the Near East and North Africa not less than 58 percent. However, the proportion of renewable water resources allocated to irrigation in sub-Saharan Africa, Latin America and East Asia in 2030 is likely to remain far below the critical threshold.
Water contained in shallow underground aquifers has played a significant role in developing and diversifying agricultural production. This is understandable from a resource management perspective: when groundwater is accessible it offers a primary buffer against the vagaries of climate and surface water delivery. But its advantages are also quite subtle. Access to groundwater can occasion a large degree of distributive equity, and for many farmers, groundwater has proved to be a perfect delivery system. Because groundwater is on demand and just-in-time, farmers have sometimes made private investments in groundwater technology as a substitute for unreliable or inequitable surface irrigation services. In many senses, groundwater has been used by farmers to break out of conventional command and control irrigation administration. Some of the management challenges posed by large surface irrigation schemes are avoided, but the aggregate impact of a large number of individual users can be damaging, and moderating the ‘race to the pump-house’ has proved difficult. However, as groundwater pumping involves a direct cost to the farmer, the incentives to use groundwater efficiently are high. These incentives do not apply so effectively where energy costs are subsidized; such distortion has arguably accelerated groundwater depletion in parts of India and Pakistan.
BOX 3 FOOD SECURITY AND ITS INDICATORS
Food security is defined by FAO as physical, social and economic access for all people to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active and healthy life. Its converse, food insecurity, applies when people live with hunger and fear of starvation. Food security requires that:
The individual state of health is also relevant for food security as disease-stricken people are hampered or unable to contribute to their own and their household’s food security. By the same token, undernourished people are much more prone to disease.
For regional and global assessments, per capita food intake per day in kilocalories is used as the indicator for food security. This indicator is derived from agricultural production and trade statistics. At the national level, a per capita food intake of less than 2,200 kcal/day is taken as indicative of a very poor level of food security, with a large proportion of the population affected by malnutrition. A level of more than 2,700 kcal/day indicates that only a small proportion of people will be affected by undernourishment. As people are enabled to access food, per capita food intake increases rapidly but levels off in the mid-3,000s. It must be stressed that per capita food intake in terms of kilocalories is only an indicator of food security: adequate nutrition requires, in addition to calories, a balanced diversity of food including all necessary nutrients.
The technical principles involved in sustainable groundwater and aquifer management are well known but practical implementation of groundwater management has encountered serious difficulties. This is largely due to groundwater’s traditional legal status as part of land property and the competing interests of farmers withdrawing water from common-property aquifers (Burke and Moench, 2000). Abstraction can result in water levels declining beyond the economic reach of pumping technology; this may penalize poorer farmers and result in areas being taken out of agricultural production. When near the sea, or in proximity to saline groundwater, over-pumped aquifers are prone to saline intrusion. Groundwater quality is also threatened by the application of fertilizers, herbicides and pesticides that percolate into aquifers. These ‘non-point’ sources of pollution from agricultural activity often take time to become apparent, but their effects can be long-lasting, particularly in the case of persistent organic pollutants.
Fossil groundwater, that is, groundwater contained in aquifers that are not actively recharged, represent a valuable but exhaustible resource. Thus, for example, the large sedimentary aquifers of North Africa and the Middle East, decoupled from contemporary recharge, have already been exploited for large-scale agricultural development in a process of planned depletion. The degree to which further abstractions occur will be limited in some cases by the economic limits to pumping, and promoted where strong economic demand from agriculture or urban water supply becomes effective (Schiffle, 1998). Two countries, Libyan Arab Jamahiriya and Saudi Arabia, are already using considerably more water for irrigation than their annual renewable resources, by drawing on fossil groundwater reserves. Several other countries rely to a limited extent on fossil groundwater for irrigation. Where such groundwater reserves have a high strategic value in terms of water security, the depletion of such reserves to irrigate is questionable.