Agricultural Water Management 80 (2006) 57–73

www.elsevier.com/locate/agwat

Water harvesting and supplemental irrigation for

improved water productivity of dry farming
systems in West Asia and North Africa
Theib Oweis *, Ahmed Hachum
Natural Resource Management Program, International Center for Agricultural Research in the
Dry Areas (ICARDA), P.O. Box 5466, Aleppo, Syria Arab Republic

Available online 10 August 2005

Abstract

In the dry areas, water, not land, is the most limiting resource for improved agricultural
production. Maximizing water productivity, and not yield per unit of land, is therefore a better
strategy for dry farming systems. Under such conditions, more efficient water management
techniques must be adopted. Supplemental irrigation (SI) is a highly efficient practice with great
potential for increasing agricultural production and improving livelihoods in the dry rainfed areas. In
the drier environments, most of the rainwater is lost by evaporation; therefore the rainwater
productivity is extremely low. Water harvesting can improve agriculture by directing and concen-
trating rainwater through runoff to the plants and other beneficial uses. It was found that over 50% of
lost water can be recovered at a very little cost. However, socioeconomic and environmental benefits
of this practice are far more important than increasing agricultural water productivity. This paper
highlights the major research findings regarding improving water productivity in the dry rainfed
region of West Asia and North Africa. It shows that substantial and sustainable improvements in
water productivity can only be achieved through integrated farm resources management. On-farm
water-productive techniques if coupled with improved irrigation management options, better crop
selection and appropriate cultural practices, improved genetic make-up, and timely socioeconomic
interventions will help to achieve this objective. Conventional water management guidelines should
be revised to ensure maximum water productivity instead of land productivity.
# 2005 Elsevier B.V. All rights reserved.

Keywords: Water scarcity; Supplemental irrigation; Water harvesting; Rainfed agriculture; Water productivity

* Corresponding author. Tel.: +963 21 213433; fax: +963 21 213490.

E-mail address: t.oweis@cgiar.org (T. Oweis).

0378-3774/$ – see front matter # 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.agwat.2005.07.004
58 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73

1. Introduction

Water scarcity in West Asia and North Africa (WANA) is a well-known and alarming
problem because water in this region is the scarcest in the world and water-related issues
have become extremely acute and even critical. Today, this problem is of increasing
concern to national governments and research institutions. In WANA countries, agriculture
accounts for over 75% of the total consumption of water. With rapid industrialization,
urbanization and high population rate increase (up to 3.6%), economic realities seem
certain to reallocate water increasingly away from agriculture to other sectors though the
demand for more food and fiber is steadly increasing.
The average annual per capita renewable supplies of water in WANA countries is now
below 1500 m3, well below the world average of about 7000 m3. This level has fallen from
3500 m3 in 1960 and is expected to fall to less than 700 m3 by the year 2025. Most of
WANA countries had per capita water availability of less than 1000 m3, the threshold for
water poverty. Some countries like Jordan, the annual per capita share has dropped to less
than 200 m3 (Margat and Vallae, 1999).
Despite its scarcity, water continues to be misused. Mining groundwater is now a
common practice in the region risking both water reserves and quality. Land degradation is
another challenge in the dry areas, closely related to water. Climatic variation and change,
mainly as a result of human activities, is leading to depletion of the vegetative cover, loss of
biophysical and economic productivity through exposure of the soil surface to wind and
water erosion and shifting sands, salinization of land, and water logging. Although these
are global problems, they are especially severe in the dry areas of WANA.
Two distinct environments may be identified along a transect within the dry areas of
WANA. The first is the relatively wet zone where winter rainfall is enough to support purely
dry farming systems. Because rainfall amounts and distribution in this zone are usually
suboptimal, moisture stress periods often occur during one or more stages of crop growth
causing very low crop yields. Variation in rainfall amounts and distribution from one year
to the other causes substantial fluctuations in production that can range in the case of wheat
from 0.3 to over 2.0 t/ha. This situation creates instability and negative socio-economic
impacts.
The second zone in the WANA dry areas is the drier environment (steppe or badia),
adjacent to the first zone. It is characterized by an annual rainfall of less than 250, too low to
support economical dryland farming. Much of WANA lands lie in this zone. Small and
scattered rainstorms in these regions fall on lands that are generally degraded with poor
vegetative cover. These areas have been exposed to mismanagement, overgrazing, and
removal of bushes for fuel wood, and are subject to desertification. Rainfall, although low
in annual average, when multiplied by the vast areas results in a large volume of ephemeral
water. Although this water forms a major resource, it is lost almost completely through
direct evaporation or through uncontrolled runoff. Therefore, agricultural productivity of
rainwater in these areas is extremely low.
Most of the agricultural area of WANA is rainfed and a large proportion of the region’s
agricultural livelihood is based on dryland farming systems. While irrigated areas may
produce far higher yields and marketable surpluses, the overall value of dryland production
is much greater than its market value due to social and other indirect benefits associated
 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73 59

with these systems. In the dry areas, water, not land, is the most limiting resource for
improved agricultural production. Therefore, maximizing yield per unit of water used
(water productivity), and not yield per unit of land is a better strategy. Agricultural
production and livelihoods in dry areas can be sustained, only if priority is given to
improving water productivity and enhancing the efficiency of water procurement. In other
words, more food, feed and fiber must be produced using less water.
Many technologies to improve water productivity and the management of scarce water
resources are available. Among the most promising and efficient-proven technologies are:
(i) supplemental irrigation (SI) for optimizing the use of the limited water available from
renewable resources in rainfed areas, and (ii) water harvesting (WH) for improved farmer
income in drier environment (Oweis and Hachum, 2003). Improving crop water
productivity, however, requires exploiting not only water management but rather all other
inputs such as improved germplasm, fertility and cultural practices.
This paper addresses the potential role and major research findings of supplemental
irrigation and water harvesting for improved water productivity in the dry areas of WANA.

2. Supplemental irrigation

Rainfall in WANA rainfed areas, especially in the dominant Mediterranean-type

climate, is characterized by low annual amount, unfavorable distribution over the growing
season and large year-to-year fluctuations. In Tel Hadya, a major dryland farming area in
northern Syria, the annual rainfall ranges from 230 to 504 mm with an overall average of
343 mm. Rainfall occurs mainly during the winter months (December–February) and early
spring (March), so that crops must often rely on stored soil moisture when they are growing
most rapidly during April and May. In the wet months, stored water is ample, plants sown at
the beginning of the season (November) are in their early growth stages, and water
extraction rate from the root zone is limited. Usually little or no moisture stress occurs
during this period (Fig. 1). However, during spring, plants grow faster with high
evapotranspiration rate and rapid soil moisture depletion due to higher evaporative demand
conditions. Thus, a stage of increasing moisture stress starts in the spring and continues
until the end of the season.
Shortage of soil moisture in the dry rainfed areas occurs during the most sensitive
growth stages (flowering and grain filling) of cereal and legume crops. As a result, rainfed
crop growth is poor and yield is consequently low. The mean grain yield of rainfed wheat in
WANA is about 1 t/ha, but ranges from 0.5 to 2.0 t/ha depending on the precipitation
amount and distribution, and on agronomic factors such as soil fertility and crop variety.
These yield levels are far below the yield potential of wheat (over 4–5 t/ha). Supplemental
irrigation can, using a limited amount of water, if applied during critical crop growth
stages, result in substantial improvement in yield and water productivity.
Research results from the International Center of Agricultural Research in the Dry
Areas (ICARDA) and others, as well as harvest from farmer’s fields, showed substantial
increases in crop yield in response to the application of relatively small amounts of
irrigation water. This increase covers areas with low as well as high annual rainfall. Table 1
shows increases in wheat grain yields under low, average, and high rainfall in northern
60 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73

Fig. 1. Typical soil moisture pattern over a Mediterranean-type wheat growing season.

Syria, with application of limited amounts of SI. Applying 212, 150, and 75 mm of
additional water to rain-fed crop increased yields by 350, 140, and 30% over that of rain-
fed receiving annual rainfall of 234, 316, and 504 mm, respectively. By definition, rainfall
is the major water supply source for crop growth and production, thus the amount of water
added by SI cannot by itself support economical crop production. In addition to yield
increases, SI also stabilized wheat production from year to the other. The coefficient of
variation was reduced from 100 to 20% in rainfed fields that adopted SI.
The impact of SI goes beyond yield increases to substantially improve water
productivity. Both the productivity of irrigation water and that of rainwater are improved
when both are used conjunctively (Oweis et al., 1998, 2000). Average rainwater
productivity of wheat grains in WANA is about 0.35 kg/m3. However, it may increase to as
high as 1.0 kg/m3 with improved management and favorable rainfall distribution. It was
found that one cubic meter of water applied as SI at the proper time might produce more
than 2.0 kg of wheat grain over that of rainfed.
Furthermore, using irrigation water conjunctively with rain was found to produce more
wheat per unit of water than if used alone in fully irrigated areas, where rainfall is
negligible. In fully irrigated areas water productivity in producing wheat ranges from 0.5 to
about 0.75 kg/ m3, one third of that achieved with SI. This difference suggests that
allocation of limited water resources should be shifted to the more efficient practice.

Table 1
Yield and water productivity for wheat grains under rain-fed and SI (SI) in dry, average and wet seasons at Tel
Hadya, North Syria (Oweis, 1997)
Annual rainfall Rain-fed Rainfall WP Irrigation Irrigated Yield increase GIWPa
(mm) yield (t/ha) (kg/m3) amount yield due to SI (t/ha) (kg/m3)
(mm) (t/ha)
Dry (234 mm) 0.74 0.32 212 3.38 2.64 1.25
Average (316 mm) 2.30 0.73 150 5.60 3.30 2.20
Wet (504 mm) 5.00 0.99 75 6.44 1.44 1.92
a
GIWP (irrigation water productivity) is taken as the ratio of increase in grain yield to the gross depth of SI applied.
 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73 61

Other field crops also respond positively to SI. Several barley genotypes were
supplementally irrigated to replenish 33, 66, and 100% of the deficit soil moisture in an
area with total rainfall of 186 mm. The mean grain yield (in t/ha) for the barley genotypes
tested was 0.26 t/ha (rainfed); 1.89 t/ha (33% SI); 4.25 t/ha (66% SI); and 5.17 t/ha (100%
SI). The highest yield of one of the genotype (Rihane-3) was 0.22, 2.7, 4.75, and 6.72 t/ha
for the four SI treatments, respectively. These dramatic results under SI were obtained
partly because of the drought during this season (ICARDA, 1989).
Northern Iraq is a typical WANA’s rainfed area where most of the grains of the country
are produced. In a rainfall zone (from 300 to 500 mm with non-uniform temporal and
spatial distribution), huge investments in SI systems were put to overcome rainfall
shortages. The results of studies conducted by ICARDA and Iraq showed that substantial
improvement can be made in yield and WP by using SI in conjunction with proper
production inputs and system management (Adary et al., 2002). In the growing season of
1997–1998 (annual rainfall 236 mm), rainfed wheat yield increased from 2.16 to 4.61 t/
ha by applying only 68 mm of irrigation water at the critical time. Applying 100–150 mm
of SI in April and May achieved maximum results. Early sowing (November) is the
optimal sowing date for wheat in northern Iraq. Every week of delay in sowing may result
in a grain yield loss of up to 0.5 t/ha of wheat. Yield, especially biological, significantly
increased with the increase in nitrogen fertilizer, and farmers were strongly advised to
continuously monitor the nitrogen level in the soil for economical and environmental
reasons.
Food legumes are important crops in WANA, particularly for providing low cost protein
for people with low incomes. Rainfed yields are low for the same reasons outlined earlier
for cereals. Analysis of 4-years’ data (1996–2000) of SI on winter-sown food legumes at
ICARDA’s experimental fields, northern Syria, under different water management options
has shown that in nearly all the cases GIWP is higher than RWP and TWP. Also, TWP is
higher than RWP in most of the cases (Table 2). It can be seen that lentil and faba bean are
more responsive to irrigation than chickpea.
In the highlands of WANA region, frost conditions occur between December and March
and put field crops in a dormant mode during this period. In most of the years, the first
rainfall, sufficient to germinate seeds, comes later than October resulting in the crop stand
being small when frost occurs in December and stops their growth. Rainfed yields as a
result are much lower than when the crop stand is good when the crop takes off in early
spring. Ensuring a good crop stand in December can be achieved by early sowing and
applying small amount of SI in October. SI given at early sowing dramatically increases
wheat yield because plants which emerge earlier in the autumn grow more vigorously and
yield much more in the following spring than plants which germinate late. A four-year trial,
conducted at the central Anatolia plateau of Turkey, showed that applying 50 mm of SI to
wheat sown early increased grain yields by more than 60%, adding more than 2 t/ha to
the average rainfed yield of 3.2 t/ha (ICARDA, 2003). Water productivity reached
5.25 kg grain/m3 of consumed water, with an average of 4.4 kg/m3. These are
extraordinary values for WP with regard to the irrigation of wheat. The study also
revealed that SI given later in the spring and early summer further increased yield, but
resulted in lower water productivity. Similar results were obtained in the highlands of Iran
at Maragheh (Tavakkoli and Oweis, 2004).
62 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73

Table 2
Mean values of water productivity of four seasons (1996–2000) of lentil, chickpea, and faba bean under rainfed
and SI at Tel Hadya, Aleppo, Syria
Water management Water productivity (kg production/m3 of water)
option Grain Biomass
a b c
RWP TWP GIWP RWP TWP GIWP
Lentil
Rainfed 0.31 1.26
1/3 SI 0.36 0.63 1.34 1.82
2/3 SI 0.37 0.55 1.31 1.45
Full SI 035 0.43 1.14 0.92
Chickpea
Rainfed 0.36 0.92
1/3 SI 0.36 0.4 0.94 1.00
2/3 SI 0.37 0.4 0.96 1.00
Full SI 0.35 0.32 0.90 0.85
Faba bean
Rainfed 0.34 0.99
1/3 SI 0.38 0.66 1.01 1.33
2/3 SI 0.39 0.59 1.00 1.11
Full SI 0.37 0.46 0.96 0.97
Data from Oweis et al. (2004a,b, 2005).
a
RWP is taken as the ratio of rainfed yield (grain or biomass) to rainwater.
b
TWP is taken as the ratio of irrigated yield (grain or biomass) to total water supply (rain + SI).
c
GIWP is taken as the ratio of increase in yield (grain or biomass) to the gross depth of SI water applied.

2.1. Optimization of supplemental irrigation

Optimal SI in rainfed areas is based on the following three basic aspects (Oweis, 1997):

1. Water is applied to a rainfed crop that would normally produce some yield without
irrigation.
2. Since rainfall is the principal source of water for rainfed crops, SI is only applied
when rainfall fails to provide essential moisture for improved and stable production.
3. The amount and timing of SI are scheduled not to provide moisture-stress-free
conditions throughout the growing season, but to ensure a minimum amount of water
available during the critical stages of crop growth that would permit optimal instead of
maximum yield.

2.2. Deficit supplemental irrigation

Deficit irrigation is a strategy for optimizing production. Crops are deliberately allowed
to sustain some degree of water deficit and yield reduction (English et al., 1990). The
adoption of deficit irrigation implies appropriate knowledge of crop water use and
responses to water deficits, including the identification of critical crop growth periods, and
 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73 63

of the economic impacts of yield reduction strategies. For wheat in northern Syria, the
results show significant improvement in SI water productivity at medium SI-application
rates than at full irrigation. Highest TWP was achieved at rates between 1/3 and 2/3 of
full SI.
There are different ways to manage deficit-irrigation. The irrigator can reduce the
irrigation depth, refilling only part of the soil water capacity of the root zone, or reduce the
irrigation frequency by increasing the interval between successive irrigations. In surface
irrigation, wetting furrows alternately or placing them further apart is one way to implement
deficit irrigation.

2.3. Water versus land productivity

Yield per unit of water consumed (ETWP) is a good indicator for assessing the
performance of SI. ETWP is the evapotranspiration water productivity defined as ratio of
yield to the seasonal evapotranspiration. Higher ETWP is linked with higher yields. This
parallel increase in yields and ETWP, however, does not continue all the way. At some high
level of yield, incremental yield increase requires higher amounts of water to achieve. This
means that ETWP starts to decline as yield per unit land increases above certain levels
(Fig. 2).
It is clear in Fig. 2, that the amount of water required to produce the same amount of
wheat at yield levels beyond 5 t/ha is much higher than the water requirement at lower
levels. It would be more efficient to produce only 5 t/ha while the saved water may better be
used to irrigate new land than to produce maximum yield with excessive amounts of water
at low ETWP. This, of course, applies only when water not land is the limiting resource and
without sufficient water to irrigate all the available land.

Fig. 2. Relationship between crop water productivity and land productivity for durum wheat under supplemental
irrigation in northern Syria (Zhang and Oweis, 1999).
64 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73

The association of high ETWP values with high yields has important implications for
the crop management for achieving efficient use of water resources in water scarce
areas (Oweis et al., 1998). Attaining higher yields with increased ETWP is only
economical when the increased gains in crop yield are not offset by increased costs
of other inputs. The curvilinear ETWP-yield relationship makes clear the importance
of attaining relatively high yields for efficient use of water. Policies for maximizing
yield should be considered carefully before they are applied under water-scarce
conditions. Guidelines for recommending irrigation schedules under normal water
availability (Allen et al., 1998) may need to be revised when applied in water-scarce
areas.

2.4. Water productivity in farmers’ fields

Research at ICARDA (Zhang and Oweis, 1999) has shown that applying only 50% of
full SI requirements causes a yield reduction of only 10–15%. A farmer-managed field
plots were established to demonstrate this finding, in collaboration with the Syrian
NARS, over 6 years in the rainfed areas with annual rainfall ranging from 300 to
500 mm. It was observed that farmers tend to over irrigate their wheat fields. When there
is not enough water to provide full irrigation to the whole farm, the farmer has two
options: to irrigate part of the farm with full irrigation leaving the other part rainfed or to
apply deficit SI to the whole farm. Assuming that under limited water resource only 50%
of the full irrigation required by the farm would be available (i.e., 4440 m3 for a 4 ha
field), the option of deficit irrigation was compared with other options. The results are
summarized in Table 3. They show that a farmer having a 4-ha farm would on average
produce 33% more grain from his farm if he adopted deficit irrigation for the whole area,
than if full irrigation was applied to part of the area. The advantage of applying deficit
irrigation increased the benefit by over 50% compared with that of the farmer’s usual
practice of over-irrigation.

Table 3
Wheat grain production scenarios for a 4-ha farm with various strategies of SI in Syria
Irrigation management strategy Rainfed Farmer’s Applying full Applying 50%
(342 mm) practice SI water of full SI
SI water depth applied (m3/ha) 2980 2220 1110
Grain yield (t/ha) 1.8 4.18 4.46 4.15
Rain or irrigation water productivity (kg/m3) 0.53 0.80 1.20 2.11
Farm production (t), water is not limiteda 7.2 16.7 17.8 16.6
Farm production (t) if only 50% of full 7.2 10.8b 12.5c 16.6d
irrigation requirements available
Per hectare average production (t) e 1.8 2.7 3.12 4.15
a
Entries in this row are obtained by multiplying the grain yield values by 4.
b
Calculated as: (4440/2980)  4.18 + [4 (4440/2980)  1.8.
c
Calculated as: 2  4.46 + 2  1.8.
d
Calculated as: 4  4.15.
e
Calculated by dividing the entries of the preceding row by 4.
 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73 65

2.5. Maximizing net profits

Increase in crop production per unit land or per unit water does not necessarily increase
farm profit because of the non-linearity of crop yield with production inputs, particularly
with water and its interactions with other input factors. Following English et al. (1990) and
English and Raja (1996) analysis, 10 years of SI data (1985–1996) on bread and durum
wheat were analyzed to evaluate water-yield relations and to develop optimal irrigation
schedules for various rainfall conditions (Zhang and Oweis, 1999). Quadratic crop
production functions with the total applied water (rain + SI) were developed and used to
estimate the levels of SI water for maximizing yield, net profit and levels to which the
crops could be under-irrigated without reducing income below that which would be earned
for full SI under limited water resources. The study concluded that SI scenarios
maximizing the profit under limited water resources conditions or for a targeted grain
yield of 4–5 t/ha should be recommended for sustainable utilization of water resources
and higher WP.
Unlike in full (or conventional) irrigation, the time of SI application cannot be
determined in advance. When possible, and for rational allocation of limited water
supplies, SI should be scheduled at the moisture sensitive stages of plant growth. For
rainfed cereals in WANA, sensitive growth stages are: seedling, flowering and grain filling.
Scheduling of SI should coincide with these sensitive periods to make certain that root zone
moisture does not limit growth.
ICARDA has developed methodologies to help farmers determining the right SI
management. Determining rainfed and SI production functions is the basis for optimal
economical WP. SI production functions for wheat (Fig. 3) are developed for each rainfall
zone by subtracting the rainwater production function from the total water (SI + rain)
production function. Since rainfall amount cannot be controlled, the objective is to
determine the optimtal amount of SI that results in maximum net benefit to the farmers.
Knowing the cost of irrigation water and the expected price for a unit of the product,
maximum profit occurs when the marginal product for water equals the price ratio of water

Fig. 3. Supplemental irrigation production functions of wheat in Syria at different levels of rainfall (Oweis, 1997).
66 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73

Fig. 4. Supplemental irrigation optimization chart for wheat in Syria (Oweis, 1997).

to the product. Fig. 4 shows the optimal amount of SI to be applied under different rainfall
zones and various price ratios.
One of the practical problems of SI is that all the fields may need irrigation at the
same time in the spring. The result is a very high water supply and large irrigation system is
required. A multi-sowing date strategy reduced the peak farm water demand rate by more
than 20%, thus potentially allowing a reduction in the irrigation system size and cost
(Oweis and Hachum, 2001). Also, the water demand of a larger area can be met with the
same water supply. However, optimal sowing dates that minimize farm water demand do
not always maximize total farm production and/or water productivity. The outcome
depends on the crop water requirements and yield for each sowing date. Furthermore,
selection of the optimal strategy is greatly influenced by the level of water scarcity.

2.6. Cropping pattern and cultural practices

Among the management factors for more productive farming systems are the use of
suitable crop varieties, improved crop rotation, sowing dates, crop density, soil fertility
management, weed control, pests and diseases control, water conservation measures (Pala
and Studer, 1999). SI requires varieties that are adapted to or suitable for varying amount of
water application. Appropriate varieties need first to manifest a strong response to limited
water applications and they should maintain some degree of drought resistance. In
addition, the varieties should respond to the higher fertilizer rates that are generally
required under SI (Oweis, 1997), and resist lodging, which can occur in traditional varieties
under irrigation and fertilization.
Delaying the sowing date prevents crop germination and seedling establishment
because of the rapid drop in air temperature starting generally in November. In the
lowlands of the Mediterranean region, where continuous cropping of pure cereal or cereal-
legume rotations prevails, mid-November was found to be the optimum sowing time for
cereals. Every week’s delay after this time may result in an average yield decrease of about
250 kg/ha. If the onset seasonal rain is delayed, early sowing can be realized by SI.
 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73 67

Given the inherent low fertility of many dry-area soils, judicious use of fertilizer is
particularly important. In northern Syria, 50 kg N/ha is sufficient under rainfed conditions.
However, with water applied by SI, the crop responds to nitrogen up to 100 kg N /ha, after
which no benefit is obtained. This rate of N greatly improves WP. It is also important that
there is adequate available phosphorus in the soil so that response to N and applied
irrigation is not constrained (Ryan, 2000).
To obtain the optimum output of crop production per unit input of water, the mono-crop
WP should be extended to a multi-crop WP. Water productivity of a multi-crop system is
usually expressed in economic terms such as farm profit or revenue per unit of water used.
While economic considerations are important, they are not adaquate as indicators of
sustainability, environmental degradation and natural resource conservation.

3. Water harvesting

The drier environments, ‘‘the steppe’’, or, as called in West Asia, ‘‘Al Badia’’, occupy
the vast majority of the dry areas of the world. The disadvantaged people generally live
there. The natural resources of these areas are subject to degradation and the income of the
people who depend mainly on grazing is continuously declining. Due to harsh conditions,
people are increasingly migrating from these areas to the cities, with associated high social
and environmental costs.
Precipitation in the drier environments is generally low compared to crop basic needs.
It is unfavorably distributed over the crop-growing season and often comes with high
intensity. As a result, rainfall in this environment cannot support economical dryland
farming. In the Mediterranean areas, rain usually comes in sporadic, unpredictable storms
and is mostly lost in evaporation and runoff, leaving frequent dry periods during the crop
growing season. Part of the rain returns to the atmosphere, directly from the soil surface by
evaporation after it falls, and part flows as surface runoff, usually joins streams and flows
to swamps or to ‘‘salt sinks’’, where it loses quality and evaporates; a small portion joins
groundwater. The overall result is that most of the rainwater in the drier environments is
lost with no benefits and/or productivity (Oweis et al., 1999).
Water harvesting in agriculture is based on the principle of depriving part of the land
of its share of rainwater (which is usually small and non-productive) and adding it to the
share of another part. This brings the amount of water available to the target area closer
to the crop water requirements so that economical agricultural production can be
achieved and thus improving the water productivity. Thus, water harvesting may be
defined as ‘‘the process of concentrating precipitation through runoff and storing it for
beneficial use’’.
Capturing rainwater and using it efficiently is crucial for any integrated development.
Water harvesting is an ancient concept with a wealth of indigenous knowledge available.
Indigenous systems such as jessour and meskat in Tunisia, tabia in Libya, cisterns in north
Egypt, hafaer in Jordan, Syria and Sudan and many other techniques are still in use (Oweis
et al., 1999, 2001). Water harvesting may be developed to provide water for human and
animal consumption, domestic and environmental purposes, as well as for plant
production.
68 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73

3.1. Water harvesting techniques

Experience with rainwater harvesting in WANA includes Micro-catchments and Macro-

catchments systems.

3.1.1. Micro-catchments techniques

Among the widely used microcatchment WH techniques are contour ridges, semi-
circular and trapezoidal bunds, and small runoff basins. Two success stories for
microcatchment WH will be reported.

3.1.1.1. Contour bunds and ridges in the Syrian badia. Mehasseh is a very dry area in
southern Syria with a Mediterranean climate that receives only 150 mm of annual rainfall.
Land is degraded with little vegetative cover. People are poor and depend on sheep
production with grazing as the most important source of animal feed. Rainwater falls in a
few sporadic and intensive storms on crusty soils causing runoff and soil erosion. Most of
the rainwater flows to salt sinks where it evaporates. As a result rainwater productivity is
extremely low. A water-harvesting project, based on contour bunds and ridges, was
initiated in 1995 to improve the natural resources management and peoples’ livelihood in
this area.
Half a meter high contour bunds or ridges were constructed along the contour lines,
spaced between 5 and 20 m apart. The first meter along the upper side of the ridge is
allocated for cultivation, whereas the rest is the catchment. Ridges were usually formed
manually, with an animal-driven implement, but then were mechanized to cut down labour
costs. Contour ridges are one of the most important techniques for supporting the
regeneration and new plantations of forages, grasses and hardy trees on gentle to steep
slopes.
To overcome the contouring difficulties, semi-circular and trapezoidal bunds are usually
used. Earthen bunds in the shape of a semi-circle, a crescent, or a trapezoid facing directly
upslope are created at a spacing that allows sufficient catchments to provide the required
runoff water, which accumulates in front of the bund, where plants are grown. Usually they
are placed in staggered rows. The diameter or the distance between the two ends of each
bund varies between 1 and 8 m and the bunds are 30–50 cm high. Cutting the soil to form
the bund immediately upstream creates a slight depression. Runoff is intercepted here and
stored in the plant root zone. These bunds are used mainly for the rehabilitation of
rangeland or for fodder production, but also used for growing trees, shrubs and field crops.
The bunds were planted with Atriplex shrubs. An adjacent field was planted also with
shrubs without constructing the water harvesting bunds. Rainfall in 1997 on the project
area was 174 mm annual. Planted shrubs with no bunds had less than 10% survival rate,
while those grew under micro-catchments had over 90% survival rate. The three following
years were very dry with annual rainfall of less than 50 mm. Most of the surviving shrubs
without bunds dried out during the 1st drought year. The shrubs supported with water
harvesting bunds survived three consecutive drought years and are still growing
vigorously. Table 4 shows the rate of survival and volume of dry matter produced under the
two water harvesting systems compared with the control where no water harvesting was
applied.
 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73 69

Table 4
Atriplex (salt bush) survival rate and increase in biomass production under two water harvesting systems and a
control in Mehasseh, Syria
Year Annual Survival rate (%) Increase in biomass
rainfall relative to no water
(mm) harvesting biomass (%)
No water Contour Semicircular Contour Semicircular
harvesting ridges bunds ridges bunds
1997/1998 174 23 95 98 90 700
1998/1999 36 12 91 95 390 990
1999/2000 42 5 90 93 240 630
Somme et al. (2005).

3.1.1.2. Small runoff basins for fruit trees in Jordan. This technique is sometimes called
negarim (Critchley and Siegert, 1991), and made of runoff basins of small diamond- or
rectangular-shaped grid plots each surrounded by low earth bunds. They are oriented to
have the maximum land slope parallel to the long diagonal of the diamond, so that runoff
flows to the lowest corner, where the plant is placed. The usual grid size is 50–200 m2. They
can be constructed on any gradient. They are most suitable for growing trees but may also
be used for other crops. When they are used for trees, the soil should be deep enough to hold
sufficient water for the whole post-rainy season.
The arid lands of Jordan receives about 160 mm of rainfall annually and has a
Mediterranean climate. No economic crop can be grown with this amount of rainfall.
Farmers in the area depend on livestock and other forms of agriculture using limited
groundwater. A project was launched in 1987 to diversify farmer’s production by
introducing trees in suitable areas. Lack of water resources limited this option. However,
the introduction of the negarim to support fruit trees was a great success (Oweis and
Taimeh, 1996). Plots of 25–75 m2 were constructed on deep soils and almonds and olives
trees were planted in the winter season. Polymers were added to the tree pit in order to
increase the water storage capacity of the soil so enough runoff is kept for the long dry
summer. Table 5 shows the overall micro catchment system efficiency for three soil surface
treatments of a catchment of 25 m2 supporting the almond trees in Mouaqar Jordan. It
indicates that 50–60% of the rain in this environment can be captured and utilized by the
plants.
Often one can achieve high runoff efficiency (defined as the ratio of runoff volume
collected at the target area to the volume of rainwater falling on the catchment area), but
water may be lost in deep percolation in the traget area, due to improper selection of the
catchment size relative to target area under the given soil and climatic conditions. In the
work reported earlier, the storage efficiency (as defined in the footnote of Table 5) ranged
from 38 to 100%. The overall efficienty of the system, which is the product of the runoff
efficiency by the storage efficiency, ranged for the plastic sheet treatment from as high as
83% to as low as 25%. This is despite the fact the runoff efficiency was very high. If the
catchment area is increased to 75 m2, the efficiency value goes down to as low as 8%.
Therefore, designing the catchment area to fit the crop and soil conditions is vital to achieve
high water harvesting system efficiency (Oweis and Taimeh, 1996).
70 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73

Table 5
Overall microcatchment system efficiency of 25 m2 area, supporting almond trees with three surface tratments in
the Mouaqar area of eastern Jordan (Oweis and Taimeh, 1996)
Rain storm(s) (mm) Overall water harvesting system efficiencya %
Natural surface Compacted surface Covered with plastic sheet
b
16 60 55 83
65c 61 58 81
8.5b 44 47 82
5.5b 32 34 73
5 61 49 78
15 43 51 82
23c 39 52 25
10.5 70 60 32

Mean 51 50 67
a
System efficiency is the product of runoff efficiency and storage efficiency. Runoff efficiency is the ratio of
water collected from the catchment to the rainwater fell on the catcment. Storage efficiency is the ratio of water
stored in the tree root zone to the water collected fom the catchment.
b
Sum of two rain storms occured within 24 h.
c
Sum of four rain storms occured within 24 h.

Planted trees survived the harsh climatic conditions and grew well only beasuse of
providing the water harvesting system. The production was so good that farmers in the area
started adopting the technology. Generally, it is important that the location, the design and
the crop are properly selected for a successful development.

3.1.2. Macro-catchments techniques

Macro-catchment systems are characterized by having runoff water collected from
relatively large catchments. Often the catchments are natural rangeland or a mountainous
area. Catchments for these systems are mostly located outside farm boundaries, where
individual farmers have little or no control over them. Water often flows in temporary
(ephemeral) streams called ‘‘wadi’’ and is stored in surface or subsurface reservoirs, but
also can be stored in the soil profile for direct use by crops. Sometimes water is stored in
aquifers as a recharge system. Harvested runoff may be stored, in a storage facility ranging
from an on-farm pond or tank to a small dam constructed a cross the wadi, and used later for
supplemental irrigation (Oweis et al., 1999). Several issues both technical and
socioeconomic need to be considered for optimal implementation of this water-harvesting
system (Oweis and Taimeh, 2001).
There are many scenarios for the management of the water harvesting reservoir for
supplemental irrigation. One scenario is to empty the reservoir as soon as possible after it is
filled and water is stored in the soil profile to ensure reservoir space for next runoff. More
water can be stored and utilized, but the risk of not having additional runoff after emptying
the reservoirs is real. It is true that the promotion of effective release conflicting restriction
of release in preparation for drought but the availability of rainfall and early irrigation
opportunity makes the risk of crop failure minimum. Information on the runoff events,
number and distribution, over the winter season is needed to assist the proper selection of
the management strategy.
 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73 71

Table 6
Probability (%) of having certain amount of rain or more at or after the 1st of the month in Muwaqar area at eastern
Jordan (Oweis and Taimeh, 2001)a
Rain (mm/day) October November December January February March April
7 100 100 100 100 93 83 35
10 100 100 100 100 98 76 26
13 100 100 98 89 76 67 22
15 93 93 89 76 61 50 17
20 83 83 74 63 48 37 7
30 46 46 46 35 26 20 2
40 22 22 22 15 11 7 2
50 11 11 11 7 7 2 2
a
Calculation was based on rainfall record during the period of 1939–1992.

As an example, it was found that the threshold daily rainfall for the occurrence of runoff
in the Muwaqar catchment is 7 mm/day and the threshold amount to fill the constructed
reservoir is 13 mm/day. These numbers are conservative and correspond to the lower
rainfall intensities in the area. Lower values of daily rainfall were found with higher
intensities. Based on these values, the chances for runoff occurrence are very good until
beginning of April after which it drops sharply (Table 6). The 13-mm/day amount needed
to fill the reservoirs is 100% assured in December and the farmer can empty his reservoirs
with no risk at all if occurred earlier. However farmer has a very good chance (about 90%)
having this runoff in January, a good chance (76%) in February, and an in two out of three
years he may be able to fill his reservoirs in March. It all depends on how much risk is
acceptable to continue emptying the reservoir as late as possible. Under the conditions of
the areas, taking risk for one out of 3 years is common and it would not seem unwise to wait
until March to start the conservative management of the reservoir water.

3.2. Socio-economic limitations

Constraints to the implementation and adoption of water-harvesting include:

 Difficulties due to farmers’ unfamiliarity with the technology;

 Conflicts and disputes on water rights, land ownership and use; and
 Lack of adequate characterization of rainfall, evapotranspiration and soil properties.

Micro-catchment systems are usually within individual farms. This is a simple and low-
cost approach, although farmers may experience some difficulty with elements requiring
precision, such as following the contour lines or determining maximum slope. The
community may be involved in micro- and macro-catchment WH systems, typically
through a careful locally- planned program.
In planning macro-catchment WH, it is important to consider any existing water rights
to avoid upstream or downstream conflict of interests. Ideally these should be planned at
the watershed level with farmers’ participation in their planning. Community-based
management, farmer participation in planning and cost sharing, or the establishment of a
cooperative can be among alternatives recommended to manage these reservoirs and to
72 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73

overcome the problem of small holdings. Selection of the appropriate reservoir site to
maximize storage and minimize sediments and nearness to the command areas is
important.
These experiences and many others show that the productivity of rainwater in the drier
environments can be substantially increased when a proper water harvesting technique is
implemented.

4. Conclusions

In water-scarce areas, water, not land, is the primary limiting factor to improved
agricultural production. Conventional practices, which have been developed for managing
water under normal water supply conditions, are not suitable under conditions of water
scarcity. Accordingly, maximizing yield or net return per unit of water, and not yield per
unit of land, is a more viable objective for on-farm water management in the dry farming
systems. Improving water productivity in water-scarce areas requires a change in the way
agriculture is practiced. More efficient irrigation practices must be adopted. Supplemental
irrigation is an option with high potential for increasing water productivity in rainfed areas.
Scarce water, now used for full irrigation, could be reallocated to supplement dry farming
for improved water productivity. However, to maximize the benefits of SI other inputs and
cultural practices must also be optimized.
Water harvesting can substantially increase rainwater productivity in the drier marginal
environments. It is the only option for economic agricultural and environmental protection
in these areas. Though technologies for various conditions are available, many of them are
not widely adopted. This can be attributed to technical, socioeconomic and policy factors,
but most importantly the lack of community participation in the development and
implementation of these technologies.

References

Adary, A., Hachum, A., Oweis, T., Pala, M., 2002. Wheat productivity under supplemental irrigation in Northern
Iraq. On-farm water husbandry research report series, no. 2. International Center for Agricultural Research in
the Dry Areas (ICARDA), Aleppo, Syria.
Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop evapotranspiration. Guidelines for Computing Crop
Water Requirements. FAO Irrigation and Drainage Paper 56. FAO, Rome, Italy.
Critchley, W., Siegert, K., 1991. Water Harvesting. FAO, Rome, Italy.
English, M., Musick, J.T., Murty, V.V.N., 1990. Deficit irrigation. In: Hoffman, G.J., Howell, T.A., Solomon, K.H.
(Eds.), Management of Farm Irrigation Systems. ASAE, St. Joseph, MI, pp. 631–663.
English, M., Raja, S.N., 1996. Perspectives on deficit irrigation. Agricultural Water Management 32, 1–14.
ICARDA, 1989. Farm Resource Management Program Annual Report for 1989. International Center for
Agricultural Research in the Dry Areas (ICARDA), Aleppo, Syria.
ICARDA, 2003. ICARDA Annual Report 2002. International Center for Agricultural Research in the Dry Areas
(ICARDA), Aleppo, Syria.
Margat, J., Vallae, D., 1999. Water Resources and Uses in the Mediterranean Countries: Figures and Facts. Blue
Plan, UNEP-Regional Activity Center.
Oweis, T., 1997. Supplemental Irrigation: A highly Efficient Water-Use Practice. ICARDA, Aleppo, Syria.
 T. Oweis, A. Hachum / Agricultural Water Management 80 (2006) 57–73 73

Oweis, T., Taimeh, A., 1996. Evaluation of a small water harvesting system in the arid region of Jordan. Water
Resour. Manage. 10, 21–34.
Oweis, T., Hachum, A., 2001. Reducing peak supplemental irrigation demand by extending sowing dates. Agric.
Water Manage. 50, 109–123.
Oweis, T., Taimeh, A., 2001. Farm water-harvesting reservoirs: Issues of planning and management in dry areas.
In: Adeel, Z. (Ed.), Integrated Land Management in Dry Areas. Proceedings of a joint UNU-CAS international
workshop, Beijing, China, 8–13 September 2001. UNU, Tokyo, Japan, pp. 165–182.
Oweis, T., Hachum, A., 2003. Improving water productivity in the dry areas of West Asia and North Africa. In:
Kijne, W.J., Barker, R., Molden, D. (Eds.), Water Productivity in Agriculture: Limits and Opportunities for
Improvement. CABI Publishing, Wallingford, UK, pp. 179–197.
Oweis, T., Pala, M., Ryan, J., 1998. Stabilizing rain-fed wheat yields with supplemental irrigation and nitrogen in a
Mediterranean-type climate. Agronomy J. 90, 672–681.
Oweis, T., Hachum, A., Kijne, J., 1999. Water Harvesting and Supplemental Irrigation for Improved Water Use
Efficiency in the Dry Areas. SWIM Paper 7. International Water Management Institute, Colombo, Sri Lanka.
Oweis, T., Zhang, H., Pala, M., 2000. Water use efficiency of rain-fed and irrigated bread wheat in a Mediterranean
environment. Agronomy J. 92, 231–238.
Oweis, T., Prinz, D., Hachum, A., 2001. Water Harvesting: Indigenous Knowledge for the Future of the Drier
Environments. ICARDA, Aleppo, Syria.
Oweis, T., Hachum, A., Pala, M., 2004a. Water use efficiency of winter-sown chickpea under supplemental
irrigation in a Mediterranean environment. Agric. Water Manage. 66, 163–179.
Oweis, T., Hachum, A., Pala, M., 2004b. Lentil production under supplemental irrigation in a Mediterranean
environment. Agric. Water Manage. 68, 251–256.
Oweis, T., Hachum, A., Pala, M., 2005. Faba bean productivity under rainfed and supplemental irrigation in
northern Syria. Agric. Water Manage. 73, 57–72.
Pala, M., Studer, C., 1999. Cropping systems management for improved water use efficiency in dryland
agriculture. Paper presented at the International Conference on Water Resources Conservation and Manage-
ment in Dry Areas. 3–6 December 1999. Amman, Jordan.
Ryan, J. (Ed.), 2000. Plant Nutrient Management under Pressurized Irrigation Systems in the Mediterranean
Region. Proceedings of the IMPHOS International Workshop organized by the World Phosphate Institute, 25–
27, Amman, Jordan. ICARDA, Aleppo, Syria.
Somme, G., Oweis, T., Abdulal, A., Bruggeman, A., Ali, A., 2005. Micro-catchment Water Harvesting for
Improved Vegetative Cover in the Syrian Badia. ICARDA, Aleppo, Syria.
Tavakkoli, A.R., Oweis, T., 2004. The role of supplemental irrigation and nitrogen in producing bread wheat in the
highlands of Iran. Agric. Water Manage. 65, 225–236.
Zhang, H., Oweis, T., 1999. Water-yield relations and optimal irrigation scheduling of wheat in the Mediterranean
region. Agric. Water Manage. 38, 195–211.