Previous Page Table of Contents Next Page


Rainwater for improving yields

Much of the future food needed by the increasing numbers of people in developing countries will have to come from rainfed rather than irrigated lands, because the possibilities for increasing the area under irrigation are limited. Subhumid to semiarid areas are characterized by rainless periods, both within and between rainy seasons, which are generally unpredictable. Because of this the output of crops, pastures and streamflow is affected not only by the total amount of rainfall in a particular season, but also by the frequency, duration and severity of water stress in the plants at different stages of growth.

Greater attention to the value, capture and use of rainwater in increasing production from rainfed lands in the tropics and subtropics is justified on two main counts:

Soil productivity should be maintained and improved overtime. Two features are fundamental, for without them plant growth will be limited and the productivity of soils will not be sustainable:

Successful water management in a dryland farming system is based on: (1) retaining precipitation on the land; (2) reducing evaporation; (3) utilizing crops that have drought tolerance and fit rainfall patterns (Stewart, 1985). This raises three questions:

To effectively address the rising concerns about the land's capacity to produce crops and sustain streamflow, it is no longer sufficient to consider macroscopic factors alone. A framework for action must be based on understanding at microlevel as well. This will include understanding of how plants and soils function together and how they are likely to react to proposed improvements, e.g.:

More widespread understanding of such factors may lead to a greater respect for the soil as an environment for biological activity, for meso- and micro-organisms as much as for roots themselves.

Deteriorating water supply

A deterioration of water supplies refers to diminished quantities of groundwater and surface water as well as to deteriorating water quality. Poor quality water may be the result not only of inappropriate land use and soil management practices which result in materials being transported by surface runoff, but also of industrial and urban pollution due to inadequate processing controls and poor sanitation.

Increased runoff at the expense of rainfall infiltration is a major cause of declining groundwater, as less water is then available to percolate through the soil down to the groundwater, i.e. less recharge occurs. Increased runoff is often the result of changes in land use that reduce the protective ground cover and decrease surface soil porosity, as for example when forest vegetation is converted into inadequately-managed annual cropping. Such land use changes often arise when rising population pressures force people to cultivate or graze land that is poorly suited to the use to which it is being put.

Changes in land use that increase the quantity of water used in transpiration, such as reforestation programmes, will be expected to diminish the frequency and amount of groundwater recharge, assuming no changes in the amount of rainwater lost by runoff or other processes. Conversely, deforestation followed by the cultivation of annual crops would be expected to decrease transpiration and so increase groundwater supplies, as long as no extra runoff occurs as a result.

Drainage of swampy areas in middle and upper watershed positions can also reduce the amount of water reaching the groundwater through deep drainage due to the diversion of water into drainage canals. Falling groundwater levels may arise as a result of increased water consumption by irrigation schemes. The lack of proper drainage in irrigation schemes may lead to deteriorating groundwater quality due to the accumulation of salts.

Greater runoff can also result from urbanization because of the replacement of agricultural land by extensive areas of tarmac and concrete, such as roads, pavements and buildings. These prevent water infiltration and generate high proportions of runoff. In many developing countries, as populations grow and industrialization and urbanization increase, the demand for water grows and eventually exceeds the quantities available. San Salvador for example, has suffered serious water shortages due to various factors, including increased urbanization and industrialization (PRISMA, 1995; Barry and Rosa, 1995).

Throughout the world's continents water tables are falling, and it has been estimated that by 2025 more than half the world's population will be living in regions suffering from a shortage of water (Rockstrom, 1999). The combination of falling groundwater and greater runoff will reduce the base flows of rivers and streams, and will greatly increase peak flows and the incidence of floods. High runoff often affects the quality of surface water by its load of eroded soil sediments, may make the water unsuitable for drinking, and may increase the costs of water treatment. High sediment loads in the reservoirs of hydroelectric schemes will reduce the life span of dam sites and increase turbine maintenance costs.

The recommended amount of water needed for one person per day for cooking, drinking and washing is about 50 litres, but the amount of water needed every day for a crop to transpire and produce sufficient grain for one person is some 10 to 20 times larger. Therefore, water shortages will have most effect on food production, rather than on the availability of water for domestic use.

Indicators of deteriorating water supply

The following are simple visual indicators of deteriorating quantity and quality of water supplies (FAO, 2000a).

Indicators of reduced surface water:

Indicators of reduced quality of surface water:

Indicators of reduced quality of groundwater:

Soil productivity and soil erosion

Increased possibilities for safe and sustainable intensification of production can be identified if the nature of soil productivity and the process and the effects of soil erosion are examined.

Soil productivity

Fertility is the inherent capacity of a soil to supply nutrients in adequate amounts and suitable proportions, whereas soil productivity is a wider term referring to the ability of a soil to yield crops (Brady, 1974). The chief factors in soil productivity are soil organic matter (including microbial biomass), soil texture, structure, depth, nutrient content, water-storage capacity, reaction and absence of toxic elements. A brief description would indicate that soil productivity depends on physical, hydric, chemical and biologic characteristics and their interaction.

Much can be known about the above-ground growth and development of plants by observing and measuring them and their functions. We know much less of what goes on beneath the surface in the soil ecosystem, where plants' roots are important constituents. In order to function unimpeded through cycles of growth to maturity, land plants require water to pass through them from soil to atmosphere.

Plate 28 shows the length of roots relative to the above-ground parts. The smaller roots and their root hairs also constitute a large part of any root system, but are not easily visible.

PLATE 28. Root systems exposed at a roadside cutting (Burgay, Ecuador)

[T.F. Shaxson]

Root systems are of astonishing dimensions, as illustrated by comparative figures from the same-sized samples of soil (Table 7). In favourable conditions, roots of some plant species may grow by as much as 10 mm/day.

Dimensions of roots of three grasses in sample of 0.688 litre taken to a depth of 15 cm (after Russell, 1961)




Kentucky Bluegrass

Total length
Total surface area

29 m
406 cm2

46 m
316 cm2

64 m
503 cm2

384 m
2 128 cm2

Root hairs
Total number
Total lengths
Total surface area

6.1 million
0.6 km
0.03 m2

6.3 million
8.1 km
0.3 m2

12.5 million
16.8 km
0.7 m2

51.6 million
51.7 km
15.8 m2

The relative proportions of solids, liquid and gases in the rooting environment are as important as the manner in which they are arranged. In this respect, the biotic content of soil is also important because together with plant roots it contributes to restructuring soil and the improvement of porosity after damage by compaction, pulverization or structural collapse.

Without good conditions for roots very high-yielding crops will be unable to express their full potentials. The key role of soil porosity is shown in Figure 10.

FIGURE 10. Soil structure and its impact on soil processes and agricultural sustainability (Lal, 1994)

Many degraded soil situations have arisen because of mould-board or disk plough-based farming practices, which have resulted in:

Each of these four factors negatively affects soil as a habitat for plant roots. The view that such conventional tillage methods limit the development of an optimum habitat for rooting is borne out by the experience of alternative agriculture systems.

The amount of moisture in the soil depends on how much rainfalls and enters the soil. Under rainfed agriculture, the amount of water entering the soil depends on what percentage is diverted above the surface as runoff. It may not always be possible to prevent all runoff, but improvement of soil physical conditions will help to reduce it to the unavoidable minimum.

A thin surface crust or subsurface compaction can be enough to reduce the rainwater infiltration rate, to provoke runoff and to cause loss of water, soil and consequently potential soil moisture (Plates 29 and 30). However, when the soil is covered with litter, more water enters and the soil surface is protected from the force of raindrops (Figure 11).

FIGURE 11. Where the soil is not crusted and is protected by litter from the force of raindrops more water enters the soil than where the surface is bare (Lilongwe, Malawi) [T.F. Shaxson]

Infiltration and availability of water to plants depends on how the water is held between, firstly, individual particles of the soil (e.g. microscopic plate-like clay particles, irregular coarse-sand particles etc.) and secondly, on the distribution of different sizes of pore space in soil. If the majority of pore spaces is very small, whether because of the inherent properties of the soil such as in a very clayey Vertisol, or because the soil has been compacted, the water may be held so tightly that plants cannot extract much of it. On the other hand if the soil has a predominance of large spaces, such as in coarse-textured sandy soils, much or all of the rainfall may enter the soil and pass easily and rapidly down through the profile without much of it being retained. A wide range of pore sizes is therefore desirable for enabling both retention and transmission of rainwater.

PLATE 29. A thin surface crust caused by raindrop impact on a bare soil of poor structure (Tabatinga, Brazil)

[T.F. Shaxson]

PLATE 30. Crusting and subsurface compaction can result in serious losses of water and soil (Khobotle, Lesotho)

[T.F. Shaxson]

Reduction of pore space may be at least as important as losing soil particles with respect to yield. It affects water movement and the soil's tenacity of water retention, root expansion and gas exchange of O2 and CO2 with the atmosphere. Its loss is similar to losing spaces in a block of apartments when they are demolished: the same quantity of materials remains, but the value of the architecture is lost because there are no longer usable voids/rooms.

Field observations in Malawi, Zambia and Tanzania indicate that repeated tillage to the same depth with hand hoes can cause subsurface pans of compacted soil at the base of the tillage layer. After a few years, they may become so dense that neither roots nor water can penetrate them easily. This increases surface runoff, severely limits soil depth and causes stunting of roots (Plate 31). With root access to soil moisture restricted to the shallow soil layer above the compacted layer, plants are prone to suffer water stress after only a few days of dry weather.

PLATE 31. Roots of a cotton plant stunted and diverted sideways by a very compact subsurface layer (São Paulo, Brazil)

[T. F. Shaxson]

Much of the blame for this damage to soils can be attributed to inappropriate tillage practices. Such problems are found not only in tropical areas but also are now widely seen as well in temperate zones across North and South America, Europe, Asia, Australia and New Zealand. In temperate zones damage is due to compaction by machines. In tropical regions much damage is also done by high-intensity rainfall on unprotected soils.

Physical degradation of the soil is the precursor of excessive runoff, reduced soil moisture and root restriction, and is a primary limitation to crop growth. There are already widespread and serious problems in soils as rooting environments, characterized not so much by erosion as by unexpectedly poor performance of crops in large parts of the world, in both rainfed and irrigated areas.

Soil erosion

Runoff and erosion occur because soil porosity has been damaged, whether at or below the surface of the soil. They are the consequences, not primary causes of land degradation. In many instances tillage aimed at loosening the soil to let in more rainwater can also result in soil collapse, which then leads to increased erosion and loss of potential soil moisture through runoff.

It is sometimes assumed that yield reductions following soil erosion can be directly related to the quantities of soil materials lost. This may not always be the case, for example where erosion removes a similar quantity and quality of material from three soils with different subsoils, yields may be different from each other and equal, lower or higher than before the erosion (Figure 12). Where a subsurface layer is of better quality for rooting than that which overlays it, erosion could be followed by higher, not lower, yields though such situations are not common.

The differences in yields are related to the differences in the characteristics of the subsurface habitats in which the roots will grow before and after the erosion i.e. differences in depth, organic matter content, infiltration capacity, plant nutrient supply, biotic activity and architectural stability (Plates 32 and 33).

PLATE 32. Erosive loss of the pale surface soil would expose the dark soil layer whose characteristics might provide a completely different quality of potential seed bed (Thabana Morena, Lesotho)

[T.F. Shaxson]

PLATE 33. Runoff, due to damage to soil porosity, removed the topsoil and the subsoil is exposed; accumulation of fine soil fractions and organic matter at the lower margin (bottom left of the picture) is at the expense of soil depth and quality at the upper side (Iracemápolis, Brazil)

[T.F. Shaxson]

FIGURE 12. Yield after erosion is related to the quality of soil remaining, not to quantity and quality of soil removed (Shaxson, 1997a)

Differences in soil surface porosity affect water infiltration rates. Compaction of the soil by trampling or machinery has the same effect, changing the hydraulic conditions of the soil. Loss of porosity in the soil increases surface runoff, increasing infiltration increases soil moisture. A failure to understand this relationship has often led to inappropriate actions to stop erosion, such as construction of physical works or overuse of fertilizers (Box 2).


In the Philippines, investigations were made of loss of soil productivity associated with erosion. This problem is widely recognized by small farmers who have observed natural terraces forming between contoured buffer strips on agricultural land with 20-30 percent slopes. They noted that crop performance on the upslope side of these terraces was not as good as that on the lower side. To rectify this difference, they applied up to three times their usual rates of N, P and K fertilizers on the degraded upper parts of the strips between the buffers.

But even though the eroded soil had accumulated along the lower side, as in Plate 33, the yield increase owing to the fertilizers was insufficient to make up for the drop in yields averaged between those from the upper and lower halves of these terraces taken together. It was then found that placing all the residues from the previous maize crop on the upper part of the terrace significantly increased fertilizer efficiency on the degraded zones of the terrace; adding lime - to reduce the acidity of the soil - also helped to raise yields and increasing soil organic matter was recommended as a long-term measure to sustain crop yields.

This experience suggests that part of the problem was the poor physical, chemical, hydric and biological condition of the soil for rooting, which could not readily be improved until more water reached the roots and the soil quality had been improved (Stark, 2000). Comparable observations have been made in other countries, including in El Salvador

(Vieira et al., 1999).

Conventional physical means of Soil and Water Conservation (SWC) have often proved less than satisfactory and not widely acceptable to farmers, because they tried to halt runoff and erosion rather than concentrating first on improving the absorptive capacity and productivity of the soil in situ, thereby minimizing runoff and erosion as a consequence. As an example Plate 34 shows a field where compaction caused by excessive disking has created a subsurface pan, which has reduced the effective depth to about 4 cm, resulting in a waste of rainwater from excessive runoff. In undisturbed conditions under native vegetation, the effective rooting depth for this soil is more than 3 metres. Breaking the pan to restore favourable conditions for rooting and water infiltration would have been more appropriate than using sandbags.

PLATE 34. Hardpan, runoff and inappropriate physical measures (Tabatinga, Brazil)

[T.F. Shaxson]

Each time erosion occurs, the rooting environment for the subsequent crop is altered. This understanding shows the need to:

This approach is radically different from and should precede any physical measures that may still be necessary to catch and redirect runoff once it has begun. Soil characteristics, which favour water infiltration and gas exchange are the same as those that minimize runoff and erosion. In this way conservation concerns can be fully integrated with the production process.

Plant-damaging drought

The figure for the annual average rainfall is no indicator of the frequency of drought, either between or within years, as shown in Table 8. The mean rainfall between 1956 and 1977 was 1 025 mm, but the variation was from 507 to 1 917 mm.

Annual rainfall totals at Indore, India (Shaxson et al., 1980)














1 208


1 054


1 743








1 246




1 103


1 917






1 084


1 209




1 221




1 025

Drought periods within a particular year may show up as a delay in the onset of a rainy season; as dry spells of a week or more at critical periods of crop growth within the season; or as an earlier-than-expected end of the rainy season.

Unreliability of rainfall within the rainy season can be shown graphically (FAO, 1999a). Figure 13 shows an example of short-duration droughts within the monsoon season in Hyderabad (India) with serious consequences for annual crops.

Making droughts worse

Only rainwater that enters the soil can be effective with respect to plant growth and dry season streamflow. Avoidable surface runoff can reduce soil moisture and groundwater. Induced drought means that plants may become stressed earlier than need be, even though there is sufficient rainfall above-ground to provide for the crop.

A recent study in Karnataka, India shows that a precarious situation develops when the combined demands for soil moisture for plants and liquid water exceed average recharge of soil and groundwater (Box 3).


In three watersheds, there has been a dramatic increase in groundwater extraction for irrigation during the last 10 years. This has been driven by the relatively higher profitability of irrigated agriculture when compared with rainfed agriculture. Although there may be some small areas of unexploited aquifer in two of the watersheds, the evidence points to the conclusion that current levels of groundwater extraction are approximately equal to annual recharge. Over large areas, wells are pumped for irrigation each year until they fail.

As a direct consequence of increased groundwater extraction, groundwater levels have fallen and shallow wells have failed as tube wells have been constructed and as extraction from the deep aquifer has become the norm. Falling groundwater levels have led to changes in the surface hydrology of the project watersheds. Springs and seepage zones have dried and now only flow or become saturated after exceptionally wet periods. Flow in ephemeral streams is less prolonged after large rainfall events and as a consequence flows into reservoirs are reduced.

Although runoff for individual or sequences of rainfall events is often higher than the 2 and 6 percent average annual runoff recorded from plots and fields, this finding shows that there are no large volumes of additional surface water that can be harvested in the project watersheds.

Estimates of groundwater use on a village-by-village basis show that extraction is far from uniform. Levels of groundwater extraction in some villages are more than 2.5 times higher than recharge values.

In some villages there are already problems of water shortage in the dry season. In these cases, it is the poor, particularly women and children, who suffer most. Even more worrying is the prospect of a major groundwater drought in the region. Levels of groundwater extraction are such that, in many areas, there is no longer a groundwater buffer that can be used as a source of supply during periods of meteorological drought when no recharge will take place.

The results of the study show clearly that the focus of [the project] should be on water resource management as opposed to water resource development. Water resources in the watersheds are close to being fully developed and, in general, constructing check dams or new wells will only change the pattern of water abstraction and use, but will not make additional water resources available.

A fundamental need is to consider trade-offs associated with changing patterns of water use and select options that maximize the social and economic value of water in any given setting at the watershed scale. In most cases, this means giving drinking-water the highest priority and then allocating water to uses that have the next highest social and economic value.

(Adapted from Batchelor et al., 2000)

Areas may be increasingly desertified by land management practices, which result in soil damage (Box 4).


Near Arusha in Tanzania there have always been drier and then wetter years, and the long-term average did not show a decline. But despite this people complained about worse droughts occurring. This is an increasingly common phenomenon in many parts of the world, where usually due to land mismanagement the soil surface layers become less porous, allowing smaller proportions of rainfall to infiltrate with increasing runoff. Soil moisture therefore is not replenished to the extent indicated by rainfall figures alone (Christiansson, 1988). The mean annual rainfall (500 to 800 mm) near Kondoa used to support dense vegetation. The erosion has been terrible following clearing, excessive cultivation and overgrazing. The waste of potential soil moisture and groundwater over the years, more than the erosion process itself, has resulted in an impressive reduction and change of vegetation (Plate 35).

PLATE 35. Mismanagement of the forests of Ugogo (Tanzania) has led to a major reduction in vegetation density and a change towards more drought-tolerant species

[C. Christiansson]

Shortening the duration of drought

Climatic drought is unavoidable. Extending the period in which soil moisture remains available to plants shortens the duration of potentially damaging water stress in plants. At the same time this shortens the length of the non-producing period of the year during which the stored food will be eaten before the next harvest.

In seasonally dry regions the focus of attention should be on how much rainwater can be caught and stored in the soil much more than on emphasizing how much runoff has occurred across the land surface. Root systems are more extensive when water is not a limiting factor, as illustrated by differences in root growth by plants of the same clone of tea grown under rainfed and irrigated conditions (Figure 14). After 9 months, the root systems of one representative plant from each treatment were exposed by root washing and drawn to scale on paper. Dry periods within the 1968-1969 rainy season led to inhibition of root growth in the rainfed treatment, whereas the regular provision of sufficient irrigation water avoided water stress in the other treatment and resulted in a more profuse root system. Roots grow and extend within those volumes of soil where soil moisture is available.

This indicates that enabling more rainwater to enter the soil and minimizing losses from runoff and evaporation from the surface will be beneficial for root growth, provided that other factors such as nutrient levels and physical barriers to root growth are not limiting.

FIGURE 13. Within-season droughts with annual rainfall totals of 1 275 mm (1915), 776 mm (1965) and 340 mm (1972) in Hyderabad, India (Krantz and Kampen, 1978)

The development of a wide range of stable pore spaces cannot be achieved by mechanical tillage and can only result from soil biological activity. Soil organisms make a major contribution in developing and maintaining porosity and may allow plant survival even after water stress may have caused active growth to cease. Considered in this way, the severity of climatic drought can be diminished, inasmuch as plant persistence can be extended and the possibility of post-drought recovery is increased. The provision of permeable soil cover, preferably crop residues, moderates high temperatures in the upper root zone. Soil cover also prevents rain splash and encourages infiltration and markedly reduces the rate of evaporation of water from the upper layer of the soil. This conserves moisture, delaying the onset and shortening the duration of severe stress.

FIGURE 14. Root systems of two young tea plants of the same clone (MT12) without and with irrigation, after 9 months in the field (Fordham, 1969)

Even where little residue cover is available, soils in good condition under minimal tillage may provide better conditions for seedling growth and survival than those damaged by inappropriate heavy cultivation. For instance, in the dry wheat-growing lands of Western Australia, it was stated that in 2000: "...even no-tilled crops suffered severely with drought. However, their revival was markedly better than in situations where the soil structure had been damaged by tillage. Tilled soil did not receive the rain as well as soil that had softened through the years of no-tillage... The crusts from tilled soils are in strong contrast to the soft furrows [made by press-wheels at drilling-time] in the paddocks with a history of no-till....(Western Australia No-Tillage Farmers Association, 2001.)

Changing the perspective on saving soils

From this different viewpoint, important changes in emphasis include:

PLATE 36. Roots, other organisms and organic materials develop the soil ecosystem from the top downwards - here, on chalk. Purbeck, England

[T.F. Shaxson]

Plate 37. Topsoil as a rooting environment developing in an exposed marine clay, through the action of roots and of organic solutes (Poole, England)

[T.F. Shaxson]

Care about roots, soil organisms and water

The way the soil is managed as an environment for roots influences the onset, duration and severity of drought, since the roots are suppliers of water and nutrients to the other parts of the plant. Without sufficient water to satisfy plants yields may be limited after even a few days in hot weather. The more severe and prolonged the dry period, the greater the damage to final yields (FAO, 1999a).

A good understanding of the below-ground environment and of how this ecosystem functions is necessary so that it can be managed more appropriately. The key propositions for this understanding are:

The organisms in the soil ecosystem break down and transform organic materials, and contribute to:

Soil organisms are subsurface workers, who perform many soil improving activities without cost to the farmers. They deserve more attention than they generally receive on how best to provide for their requirements.

Rainwater for plants' needs should be retained by saving it in the soil where it may benefit all inhabitants of the root zone. Water in excess of these requirements should be able to pass further downwards to contribute to the groundwater, available for use downstream.

These three 'care' suggestions propose an approach aimed at determining the needs of the soil ecosystem of which plant roots are part. This is a necessary step in deciding ways to increase land use intensity without damage to the basic natural resources of water and soil.

Previous Page Top of Page Next Page