Irrigation water withdrawal

Assessing the impact of irrigation on water resources requires an estimate of the water effectively withdrawn for irrigation, i.e. the volume of water extracted from rivers, lakes and aquifers for irrigation purposes. Irrigation water withdrawal normally far exceeds the net irrigation water requirement because of water lost in its distribution from its source to the crops.

For 118 out of the 165 countries and territories information on water withdrawal is available from national sources (i.e. not estimated). In order to fill the data gaps regarding the 47 countries for which this information is not available (or only estimated), a ratio of the estimated irrigation water requirement to the actual irrigation water withdrawal is calculated for countries for which such data is available:

WR Ratio = IWR / IWW

with:

WR Ratio = water requirement ratio (irrigation efficiency) (-)

IWR = irrigation water requirement per year (m3)

IWW = irrigation water withdrawal per year (m3)

This ratio is often referred to as "water use efficiency" (FAO, 2012c) in agriculture or "irrigation efficiency". However, the use of the expression is subject of debate (Perry and Kite, 2003). The word "efficiency" implies that that water is being wasted when the efficiency is low. This is not necessarily so. The recoverable fraction of the non-consumed water can be used further down-stream in the irrigation scheme, it can flow back to the river or it can contribute to the recharge of aquifers. It is for this reason that in this study the term "water requirement ratio" is employed when referring to the ratio between irrigation water requirement and the amount of water withdrawn for irrigation.

The average of the water requirement ratio calculated at sub-regional or regional level enables, in combination with the irrigation water requirement calculated in the previous step, the estimation of irrigation water withdrawal (below) for countries with missing data. In addition, it also permits to cross-check data and thus their correction.

---

Estimation of irrigation water withdrawal

Data on country agricultural water withdrawal (AQUASTAT variable code 4250) and water withdrawal for irrigation (AQUASTAT variable code 4475) have been collected through the AQUASTAT country surveys (FAO, 1995; 1997a; 1997b; 1999; 2000; 2005a; 2009b; 2012a). In addition, updating was carried out especially for this exercise and in some cases more recent national statistics were obtained. For improved accuracy, water withdrawal for irrigation was preferred when available over agricultural water withdrawal [code 4250], which also includes water withdrawal for livestock and aquaculture. However, because irrigation is the most significant agricultural user of water, in the absence of specific data for irrigation water withdrawal, the agricultural water withdrawal is assumed to refer only to irrigation. In addition, the water withdrawal data referring to the crop calendar’s year (or the closest year) was selected when available; thus the latest values are not always the ones used for calculation.

After selection of the most appropriate year for the water withdrawal data, water requirement ratios at country level have been calculated, by comparing them with the calculated figures on irrigation water requirement.

To assign a ratio to countries for which agricultural or irrigation water withdrawal information was unavailable or estimated, a regional rather than national approach was used in an effort to reduce the large uncertainties on the value at country level. Thus correction ratios have been calculated at sub-regional or regional level when at least two coherent country ratios were available. These exclude ratios either below 15 percent or above 85 percent considered inconsistent for being too low or too high. These were then also replaced by these correction ratios:

CRr = IWRr / IWWr

with:

r = region

CRr = correction ratio for the region

IWRr = total irrigation water requirement for the countries with coherent ratios only in the region

IWWr = total irrigation water withdrawal for the countries with coherent ratios only in the region

These sub-regional or, when not available, regional corrections ratios have been assigned to countries with previously estimated or unavailable water withdrawals. They have been used to make a new estimation of water withdrawal for irrigated agriculture per country by dividing the irrigation water requirement of the country by its assigned correction ratio. These new estimations refer to the year of their respective crop calendar. They are indicated as modelled data (with the symbol "L") in the AQUASTAT core database and in italic in Table 4. The total actual renewable water resources for each country are also presented in Table 4 and are used in the calculation of the pressure on water resources due to irrigation (irrigation water withdrawal as a percentage of total renewable water resources). Regional data and analysis are available in the regional summary below.

In addition to the above estimations of water withdrawal for irrigation in order to fill the corresponding variable [code 4475] in the AQUASTAT database, the water requirement ratios also identify incoherent pairs of irrigation water requirement and irrigation water withdrawal and therefore give the opportunity to apply some corrections.

---

Corrections of irrigation water requirement and withdrawal

The water requirement ratio identifies incoherent pairs of data on irrigation water requirement and irrigation water withdrawal, and has been fixed in this review at either below 15 percent or above 85 percent. In these cases, either the values of irrigation water requirement or the values of irrigation water withdrawal are corrected based on expert judgement.

Corrections of irrigation water requirement—based on the corresponding correction ratios and their respective water withdrawal (as explained above)—apply in the following three situations:

• For countries with high inter-annual variability of precipitation (Australia, Middle East countries, Tunisia), and high (> 85 percent) or low (< 15 percent) individual country ratios: The use of the long-term national average precipitation for the calculation of the irrigation water requirement results most of the time in distorted calculated ratios for a particular year: high ratio for a wet year or low ratio for a dry year. As a case in point, precipitation in Australia was predominantly larger than average in 2010 (ABS, 2011) resulting in an individual country average above 100 percent. In these cases, the individual country ratio was replaced by the correction ratio for its sub-region (or region when not available).
• For countries with moderate climatic conditions and individual country ratios over 85 percent (such as several European countries) or below 15 percent: Again, the long-term national average precipitation used in the model for the calculation of the irrigation water requirement explains the ratio’s bias. For countries with moderate climatic conditions AAIfull varies a lot from one year to the other depending on the early spring precipitation. When the irrigation water withdrawal and irrigation water requirement used for the calculation of the water requirement ratio do not refer to the same year, the distortion can be important. Another misrepresentation in such countries comes from the actual amount of water applied which varies according to the spring and summer rainfall. When summer rainfall is higher than the long-term average, irrigation is supplemental only so that the water requirement ratio obtained is extremely high. On the contrary, when rainfall is lower than the long-term average, a higher volume of irrigation water is required to compensate for the missing precipitation, so that the water requirement ratio obtained is extremely low. In European countries, for which both elements of the water requirement ratio refer to the same year, the actual volume of irrigation water applied explains any distortion. This puts a limitation of the model used for irrigation water requirement calculations in evidence. In these cases, the individual country ratio was replaced by the correction ratio for its sub-region (or region when not available).
• For countries with very limited (< 1 000 ha) AAIfull and AHIfull: Their calculated irrigation water requirement was nil, either because the unit used (km3, that is 10^9 m3) is too large for such limited irrigation or due to lack of geo-referenced data. This represents another limitation of the model used for the calculation of the irrigation water requirement. Their individual country ratio was replaced by the average ratio for its sub-region (or region when not available).

In addition, a correction of irrigation water withdrawal is also applied to some African countries:

• Where the portion of AEIfull represents less than 50 percent of the AEItot: Africa, as mentioned in the regional summary of irrigated crop calendars, is indeed the only continent where such small proportions exist ( Table 2a). Areas of equipped lowlands and spate irrigation outside of Africa exist only in a very limited number of countries (Cuba, Kazakhstan, Yemen, Georgia, Mongolia, Myanmar, Malaysia, Pakistan and Albania) and always representing only a small proportion of the total areas equipped for irrigation. In those African countries where AEIfull is less or close to 50 percent of AEItot, the irrigation water withdrawal figure was corrected relatively to AEIfull. This is based on the fact that often the data obtained for water withdrawal for irrigation include not only AEIfull but also equipped lowlands and sometimes even spate irrigation.

The resulting irrigation water requirement and irrigation water withdrawal for 167 countries are displayed in Table 4 together with their corresponding water requirement ratio. In addition, total actual renewable freshwater resources for each country are also presented and used in the calculation of the pressure on freshwater resources due to irrigation (water withdrawal for irrigation as a percentage of total actual renewable freshwater resources). The direct use of non-conventional water was not considered in this estimate of water pressure, but is done in the MDG water indicator. Regional data and analysis are available in the "Pressure on water resources due to irrigation" section (below).

---

Pressure on water resources due to irrigation: regional summary of water requirement ratio

Water requirement ratios obtained for individual countries (Table 4) are compiled and analyzed by continent and by income-based grouping (high, middle and low income, as defined as of October 2012). Also, like in the regional summary of irrigated crop calendars, special attention is given to LIFDCs and LDCs (Table 5).

On average, for the 167 countries, it is estimated that the water requirement ratio is around 56 percent, varying from 23 percent in areas of abundant water resources (Central America) to 72 percent in Northern Africa where water scarcity calls for higher water requirement ratios. In addition to geographical disparity based on water availability, water requirement ratios also may depend on financial resources availability as the income-based grouping evidences with increasing ratios for low, middle and high income countries: 48, 56 and 61 percent respectively. The LDCs display an even lower water requirement ratio with only 50 percent. The lack of financial resources may impede the appropriate operation and maintenance of irrigation systems and the development of smallholder/small-scale irrigation, as well as the development of pressurized irrigation systems—more expensive than surface irrigation systems but with higher irrigation efficiency at field scale. Financial resources also make possible the capacity building of irrigators and water officials and the monitoring of water resources among other. The higher water requirement ratios of the high income countries counterbalances both their lower cropping intensity and lower portion of areas equipped actually irrigated (explained in the regional summary of irrigated crop calendars). At country level, variations are even higher with water requirement ratios varying from 18 percent to 85 percent. In addition, relations to both water availability and income disappear, with a number of water-stressed or high income countries having relatively low water requirement ratios and some water abundant or low income countries having relatively high water requirement ratios (see also the discussion about the limitations of this review).

Irrigation water withdrawal was estimated to account for only 5 percent of total renewable water resources for the 167 countries studied (Table 5). However, there are wide variations between regions, with the Northern Africa region using 77 percent of its water resources in irrigation and the Middle East 40 percent (and the Arabian Peninsula, one of its sub-regions, uses 472 percent of its resources), while Latin America barely uses 2 percent and Europe 1 percent. Ten countries (mostly from the Arabian Peninsula, but also from Northern Africa or Central Asia) used volumes of water for irrigation which are several times larger than their annual renewable water resources in their respective reference year. Despite the distortion of these ratios by the use of significant volume of secondary freshwater (water previously withdrawn and returned to rivers and groundwater) and non-conventional sources of water—direct use of treated or untreated wastewater, agricultural drainage water and even desalinated water for agriculture—as well as fossil groundwater in some cases, the situation remains critical in these countries. An additional 22 countries used more than 20 percent of their water resources, a threshold that could be used to indicate impending water scarcity. For other countries, relatively low national figures may give an overly optimistic impression of the level of water stress: China, for instance, is facing severe water shortage in the north while the south still has abundant water resources. Overexploitation of renewable groundwater also occurs at the local level in several countries of the Near East, Northern Africa, South and East Asia, Central America and in the Caribbean, even if at the national level the water balance may still be positive.