Synthesis report of the FAO electronic conference "Drought-resistant soils: Optimization of soil moisture for sustainable plant production"

Topic 1 - Causes and consequences of soil moisture scarcity

Poor yields are often related to an insufficiency of soil moisture rather than an insufficiency of rainfall.

One of the major threats to agricultural production in dryland conditions is the distribution of rainfall in time and space and not so much the total rainfall. This was illustrated with an example from the Niger (Box 1).

However, where residue cover or cover crops are present even under low rainfall conditions (200-350 mm/year), there can be more soil moisture available to the crop compared with bare soils. An example from the north of the United Republic of Tanzania indicated an increase in rainwater productivity of 200-300 percent. Even in years with 400 mm or less of rainfall, maize yields of about 2 tonnes/ha were realized while maize crops failed completely in neighbouring conventional systems. Several examples noted that crop yields in dryland areas are often higher in some years despite less total rainfall. This is because of soil moisture storage from past years. The rainfall intensity, duration, date and hour of rainfall event, and soil properties and management are the factors that affect the soil moisture status and particularly plant-available water (PAW), which can vary significantly with the same total amount of rainfall falling under different circumstances. Therefore, in any season, soil moisture is a better criterion for predicting the yield potential of a soil than is the actual rainfall amount.

Furthermore, the discussion highlighted the fact that the amount of rainfall is not really a good measure for predictions of crop production. Especially in conventional farming, there is only a loose relationship between rainfall quantity and subsequent soil moisture, i.e. rainfall-water efficiency is reduced by unproductive losses through surface runoff, unproductive soil evaporation through bare soil surfaces and unproductive evaporation resulting from the aeration of soil during tillage operations. It was noted that the actual rainfall limits for growing certain crops have been established considering conventional practice. With proper soil moisture management, these limits can be shifted towards the lower values, i.e. in general terms, more crops can be grown with less water. Field evidence exists to support this, i.e. crops yielding under conservation agriculture (CA) in rainfall conditions that would normally not allow the growing of a crop. Figure 1 illustrates the frequency, date of onset and duration of wilting-point-induced water stress under conventional, minimum and zero tillage.

Source: Derpsch et al. (1991).

In an environment where the natural vegetation is sparse (resulting in minimal remaining surface residues) and two normal cropping seasons can pass with virtually no rain, insufficiency of rainfall remains the problem.

Conventional farming accelerates loss of soil moisture through reduced ability of the soil to capture, drain and store rainwater.

Various examples were shared in relation to PAW, especially situations in which conventional farming was compared with the practice of no till. Cases from Australia, Germany, Kazakhstan, and the United States of America were discussed. All showed PAW increased significantly with no till (Box 2).

One of the reasons for this effect in the long run is the buildup of organic matter under no-till conditions. An example that demonstrated the difficulties of increasing organic matter in the Sudano-Sahelian Arenosols was presented.

Rainwater was often lost as runoff owing to surface sealing because of clay fractions in the surface soil. An example from the Niger demonstrated that the application of calcium (Ca) and potassium (K) fertilizer might improve water infiltration capacity. The addition of organic matter gives stability to sandy soil aggregates, as the sand particles as such do not have any surface forces to bind them. Box 3 illustrates two more examples that were presented in which organic matter addition is responsible for the reclamation of shifting sands.

Despite the economic value of utilizing residues in the Sahel, two contributors demonstrated that they are by far the most promising means for improving soil moisture conditions in this zone. Discussion also highlighted that it is necessary to increase crop yield through fertilization in order to have enough residues to cover all the needs at farm level. Their alternative value as soil stabilizer and as a production input should be better publicized. It is known that biomass production under CA would increase to an extent that the alternative needs for crop residues and the soil fertility needs can be covered. The difficult part is to start this process. Farmers would need a "credit" on biomass.

An alternative to the buildup of organic matter through retaining crop stubble and residues is the application and/or the incorporation into the immediate topsoil of a mulch of composted organic wastes. Although the application of composted organic waste improves SOM, soil moisture content, bulk density, water infiltration, nutrient distribution and a whole series of soil parameters, it was argued that the effect of ploughing and incorporating fresh crop residue on the biological oxidation of the residual organic matter is similar to "opening the air supply and stirring kindling into the smouldering coals of an old coal stove". In both cases, the oxidation is accelerated by the improved oxygen supply and accessibility of new readily oxidizable material, resulting in further oxidation of the residual carbon in the system. On the other hand, the retention of crop residues on the soil surface in association with reduced tillage not only reduces erosion, but it also reduces the physical release of carbon dioxide (CO₂) and biological oxidation of soil carbon, which has been a less obvious but usually the greater cause of organic matter depletion in soils.

In addition to these SOM effects, the evaporation from bare soil surfaces and the direct loss through tillage have to be considered. A classic example is the common habit of farmers to wait for rain for seeding, where, in conventional agriculture, the seedbed is prepared shortly before seeding. The often intensive tillage operation associated with seedbed formation dries the soil to tillage depth. Under humid conditions with good soil moisture, the re-compaction of the seedbed at sowing might re-establish the capillary flow of the water to the seed. However, this takes time. In arid conditions, the seedbed remains dry to the planting depth, so that seeding is only possible after rains that sufficiently moisten the seedbed. This situation can also be problematic in terms of gaining good germination and early plant vigour. Seed is often placed deep to give the seed a chance of staying moist for as long as possible. However, this can lead to slow early growth as roots proliferate quickly only once the plant develops leaves and establishes an energy supply from photosynthesis. In a no-till, permanent soil cover situation, soil moisture for seedling germination is usually supplied from beneath the seed from stored soil water. Therefore, seeding occurs without the necessity of rain. Moreover, the seeding depth can be shallower (as a result of soil cover), so enabling faster development of the plants and improved resilience. The efficiency of rainfall utilization in such a system is also higher.

It was emphasized that soil cover (as a result of and in association with zero tillage) is the most important factor influencing water infiltration into the soil. Unlike tilled soils, no-till soils are most receptive to rain and irrigation water as a result of several interacting factors, particularly: the potential for slow infiltration of water through crop residues and enhanced organic-matter-enriched topsoils; the reduction in damaging raindrop impact with subsequent reduction in crust formation; the maintenance of the vertical pore system created by roots, earthworms and other soil animals; improved soil structure and aggregate stability; and improved organic matter mineralization. These conditions continue to improve with time under no till.

Soil moisture monitoring is usually a better decision-making tool for farmers than rainfall predictions and weather forecasts

This applies only in no-till farming systems. In tillage-based systems under arid conditions, soil moisture loss from tillage is considerable, causing the actual supply of water from above to be an important and unpredictable feature for decision-making.

Disasters, like floods and landslides, and the costs for road maintenance and water treatment would be minimized where rainwater infiltration and storage in agricultural soils were maximized.

Sufficient evidence for this exists. Moreover, commercial evidence was cited where electricity companies (generating hydroelectricity) have ensured that CA is practised on the lands around the hydrodams. In this way, not only are damaging flash floods prevented but eroded soil does not slowly fill the reservoirs.

This was illustrated by the experiences in Paraná, Brazil, where, even in the 1980s, the unexpectedly rapid improvements in catchment hydrology became apparent as a consequence of implementing residue-based no till on significant areas of land/catchments.

Farmers cannot cope with strong climate trends and their consequences for agricultural practices because their production systems are not flexible enough

As the occurrence of extreme climate events seems to be increasing even in moderate climates, the risk in conventional farming is increasing. There is evidence that crops grown under CA suffer less variation in yields in the event of extreme weather conditions. This provides a strong argument for promoting CA as a strategy to adapt to climate change.

Good knowledge of climate patterns and crop requirements enables the development of optimal strategies (sowing dates and variety selection). Farmers have to deal with several parameters before deciding on a sowing date and crop varieties. Different methods were suggested (from Bosnia and Herzegovina, Nigeria, and South Africa) for optimizing and selecting options. Agrometeorology techniques, which may be based on farmers experiences, and/or simple crop water models give good results, and may be combined with appropriate in-field management. What is expected is a minimum effect of drought on future yield.

The above dilemma and insecurity of farming is a major problem of conventional farming in arid and unpredictable conditions. The aggravation in a conventional system is that the soil is pulverized and transformed into a very fragile condition for preparing the seedbed while at the same time heavy winds or rainstorms happen regularly.

Topic 2 - Creating drought-resistant soil: Technologies and impacts of improved soil moisture management at field level

Restoration of soil porosity by mechanical means is less satisfactory than by biological means.

The reason for increased runoff in plough-based systems is that it is most difficult to keep the soil open to capture rain. If the soil is left untouched, it will settle and lose moisture by direct evaporation. Therefore, with conventional farming the soil surface is regularly loosened to break crusts and provide a coarsely aggregated horizon as an evaporation barrier. However, these aggregates tend to disintegrate with every rain. Crust formation was also cited as a major problem in plant-stand establishment in savannah soils in both central and northern Nigeria as well as in northeastern Brazil that were under continuous ploughing. However, the problem disappeared immediately when no till was introduced.

In addition to increasing water infiltration and controlling erosion, soil cover has a major impact in reducing soil temperature, reducing evaporation, increasing available water for plants, enhancing biological activity, contributing to reduce soil compaction and soil crusting, as well as having positive effects on soil chemical, physical and biological properties. All of these are advantageous for the farmer and lead to increased productivity. Furthermore, cropping systems that have permanent soil cover are essential to achieving long-term agricultural sustainability.

Mulch is critical to effective no-tillage systems. This is more difficult to achieve in countries where crop residues are harvested (Box 4) or where long growing seasons and unequal rainfall distribution or wind lessen crop growth. With no-till systems, the soil surface evolves a better-defined structure that results in pores connected continuously to the surface (continuous macropores). Mulch will prevent moisture loss through these pores. Equally important, the surface mulch protects the soil surface from the kinetic energy of rainfall and maintains the surface porosity so that infiltration capacity is maximized.

When speaking about the importance of crop residues or roots in the system, it is important to identify what type they are and how they contribute to soil structure, infiltration and nutrient cycling. Different crop types have a range of lignin content and ratios of carbon (C) to nitrogen (N) that can be used to advantage. Moreover, different crops will have different residue geometry and structure. Combined with chemical factors (e.g. lignin content), this will contribute to differential decomposition. Similarly, the differences between annual crop, perennial forage and tree roots are important. They explore different and complementary regions of the soil profile; and they have different decomposition rates that allow for different residence times. Diverse cropping systems have a net soil benefit as well as other economic benefits (drought strategy, pathogen management, labour distribution, farm income risk, etc.). Perennial crops have higher water-use efficiency (WUE) and, because of their deep rooting system, are more drought tolerant. Thus, a combination of trees and crops may optimize the system. WUE also increases with fertility levels. Nutrient cycling can be enhanced through aboveground and belowground residue management planning.

Runoff farming and water harvesting aim to concentrate rainfall water and favour local crop growth through more effective use of rainwater. However, they are restricted to specific conditions (environmental and socio-economic) for their effectiveness in improving water-use efficiency

Water harvesting is a very old practice. It is gaining more and more attention in western Sudan, and results are very encouraging. In dealing with agricultural production in western Sudan (semi-desert to semi-arid climates), new technologies need to take into account the socio-economic aspects of the area. One example presented reported that farmers used to sow and weed in one operation. This practice led to the loss of valuable amounts of moisture at the beginning of the growing season before seeds could be sown. In short, this meant that runoff was harvested when there was no crop, and the crop was grown when there was no runoff to harvest. An integration of technologies was required.

Runoff farming is site specific. However, while importance is given to the diversion of rain-runoff water into reservoirs, very little is done about the unabated soil erosion in catchment areas. This erosion results in a buildup of sediment concentration in the runoff water, which in turn results in the siltation of these reservoirs. The outcome is a very short life span (perhaps 2-3 years) for these reservoirs.

Other methods of rainwater harvesting at a small scale were discussed, e.g. demi lunes (Niger) and zai pits (Mali and Niger). The demi lunes are 3 m x 3 m at their widest part and are connected with each other by earth bunds producing a continuous water-harvesting structure. The assumption is that farmers will maintain these structures because they will benefit from the increased fruit yields that water harvesting will give. The demi lunes not only harvest water but also the fine soil particles loaded with organic material that "fertilize" the trees continuously. Spectacular growth, beyond expectations, was generated through improvements to the traditional planting pits in Mali, Niger, etc. The so-called zai/tassa improved planting pits catch more of the sparse rainfall. They also concentrate organic matter in the form of dung/compost into these pits, and the plant nutrients it contained enable the plants to make efficient use of the rainwater that is captured more effectively. "Improved water infiltration" could be subdivided into both the rate of true infiltration and the quantity of water entering the soil.

Runoff farming and water harvesting are noteworthy. One alternative to prevent sedimentation problems is the use of agroforestry in catchment areas where cattle grazing is forbidden or restricted. It was recognized that this last suggestion is the most difficult to implement in the field.

Soil moisture management practices and organic matter decomposition processes have a major impact on vegetal, animal and soil biodiversity

Soil is a combination of inert material, roots and soil biota and it is in the soil that the transformation begins of not-living environmental substances to living ones, which is expressed by soil fertility. The use of cover crops changes the competition and dominant species distribution in the soil and plant cover and life cycles of plants and insects involved in plant pollination. Effective microbes (EM) can increase porosity by increasing the favourable organisms in the soil, and some experimental results seem to show that avocado trees need less water when these organisms are present. When crop residues are left undisturbed on top of the soil, a kind of "soil skin" is generated. This "skin" normally provides food and shelter to both microlevel and mesolevel flora and fauna, thereby increasing the soil biotic load and biological activity. In addition, macrofauna contributes to organic matter transport and long-term storage in deep soil layers.

Although the principles that lead to good soil moisture management are not new, farmers do not apply them widely, mainly because adapted technologies are not developed by researchers

It was suggested that in many places there has been little effort by researchers to visit field sites to monitor and measure the effects on soil conditions (biological, chemical, physical and hydric) of techniques that farmers themselves have devised and/or developed, put into practice, and found to be worth continuing. This is particularly evident in the case of soil moisture conditions. Most effects of treatments are judged on the basis of yield alone even though the expression of yield is the result of many different resource-interactions that occur over time within both the soil and the plant. There is considerable information available on responses to nutrient ions, but much less is known about how such responses are affected by the sufficiency or insufficiency of soil moisture, and by interactions with organic materials and processes in soil.

The question was raised: "Is this scarcity of in-field investigations on soil moisture causes/effects because this sort of muddy-boots work does not represent 'whizz-kid' science, or because of other reasons for being unwilling to leave the research stations?" At present, research seems to be following and explaining why some farmers' practices work well. Research should also promote integrated and appropriate techniques. Any such information could add to that which researchers already have, or could also be the spark for new lines of research work. It also provides an opportunity for farmers and researchers alike to see the principles in action in real-life situations.

Four requirements for the adoption of good land management practices were cited:

• For any new technology to be adopted successfully by farmers, it must bring the farmer a visible and immediate benefit - economic or otherwise.

• This benefit must be substantial enough to convince the farmer to change existing practices (i.e. benefit/cost ratio should be substantial).

• For the technology to be disseminated on its own, the costs incurred in its adoption must be affordable for the farmer.

• The introduction of the new technology should be followed up by an extension service for a long period.

Farmers themselves will be the major source of solutions once they are engaged in the development process. The problems cannot be resolved without working with farmers in real field situations. As one of the contributors stated: "the reason why technologies are not adopted by farmers is very often simple, at least for the people who are working directly with the farmers: the issue and these techniques have not been thought and conceived with the farmers but by technicians or scientists studying the problem only from their discipline and their technical point of view." A systemic approach that can integrate all the parameters (organic matter availability; organic matter share; work-time availability; etc.) could be the solution to this problem. Cultural and socio-economic knowledge and an excellent capacity for understanding and exchanging with farmers are fundamental to the sharing of concepts and practices.

In addition, in-field experimentation is required in order to observe the interaction of the different components and variables of the system.

The major barriers are often not agronomic but a complex mix of governance, policy, economic and institutional deficiencies. The successful adoption of no till in Argentina and neighbouring countries has in great measure to be attributed to the common sense of their farmers and their ability to detect the economic, physical and other advantages of the system, and how it fits well with general conditions and development bottlenecks. In addition to being more cost-effective and environmentally friendly, no-tillage systems are no more complex or challenging than other methods of crop production. The complexity tends to be in the minds of researchers and advisers who are unable or unwilling to consider the multitude of factors that contribute to the emergence of a problem and, hence, its possible solutions.

A large component of the problem of limited technological change among resource-poor farmers arises from inefficient research and extension systems that utilize a linear model of information and knowledge flow. Impact has been small because the resource-poor farmer is not integrated into efficient knowledge or information systems. Another problem associated with resource-poor farmers is their aversion to risk. Their livelihoods and those of their families depend on the decisions they take, and one bad decision can lead to hunger, farm loss or worse. CA is less risky than conventional, tillage-based agriculture with respect to most of the common environmental hazards that affect agriculture. However, innovative farmers do need some support during the early stages. They do not need paternalism and free inputs, but just some assurance that someone will look after them if things go "horribly wrong". With CA, things seldom go "horribly wrong" because it is far more knowledge intensive than it is technically intensive. Changing from ploughing to direct seeding is not a major undertaking, but changing the whole system is difficult to carry out, involves a lot of stakeholders at the same time, and needs to be understood by the farmers. Furthermore, the socio-economic environment for farmers is often not beneficial. Critical to achieving acceptance and widespread uptake of CA are the establishment of long-term integrated partnerships between farmers, researchers, agribusinesses, engineers, financiers and local politicians, and the combining of new technologies and farmers' initiatives with zero financial risk to the farmers. One major issue with no-till farming is that it has to be continuous before it can be effective both in terms of soil quality improvement and yield increase. It is not inevitable that yield decreases in the first years after changing to no till. Different reasons and management solutions were discussed to prevent yield losses. These included level of technology, full plant stand, and placed fertilizers, etc. that lead to increased rainwater productivity, and stable and even better yields.

The discussion highlighted the fact that the knowledge and technology for no-till systems are not always available in developing countries. Thus, the development of systems that require less management and knowledge and which allow the crop residues to remain on the soil surface until the time for the next cropping season might be more useful in such areas. In this way, moisture is conserved and the soil surface is protected from rain and wind impact. Moreover, owing to the absence of crust as the result of the residue cover, more rainwater will infiltrate into the soil and less will run off. At the time of planting, a one-off cultivation prepares the bed for planting, eliminates most weeds, and incorporates the residues into the soil, so providing the conditions for some organic matter buildup and additional nutrient in the soil. However, the key issue is probably to understand the necessity and the reason to change and evolve.

"Good practices", like those limiting runoff and favouring water infiltration, may have negative environmental effects at farm and watershed level

In a given watershed area (village level), improved agricultural practices cause water to infiltrate slowly into the subsoil and aquifers of the catchment. This water regenerates in surface wells and in ponds/tanks downstream. Results are positive and many: runoff is negligible; soil erosion is negligible; agroforestry becomes well established (grazing forbidden); and water is available in wells and tanks for domestic and farming purposes. This changes the whole socio-economics of the village.

Topic 3 - Environmental consequences of drought-resistant soil and improved soil moisture management

Local soil management practices will only modify water redistribution towards rivers, aquifers, etc. at local scale

This statement generated contrasting views. The discussion highlighted the significant downstream effects of local soil moisture management practices. Thinking in larger dimensions of space and time, it could be suggested that such practices could have a marginal effect on the rising of the sea level. Other contributors indicated that the effects (cleaner surface water and reduced sedimentation of lakes and dams) may be expected more at the regional than at the global level (Box 5).

Carbon sequestration, especially under tropical conditions, was considered by some contributors to be of minor importance. Optimization of green water (on-site) management improves rainwater-use efficiency and thus yield potential. At the same time, it improves blue water (off-site) resources, in particular through reduced runoff - hence, fewer flash floods and reduced erosion and water turbidity. Thus, it increases groundwater recharge and creates a more stable river baseflow. However, a recent study concluded that downstream impacts of soil management practices on hydrological regime would be felt more readily at small scales (i.e. in watersheds of less than 100 km²). Off-site effects are less clear and the impacts are site-specific and dependent on many factors. As one participant stated: "We should be careful when generalizing that off-site, or "blue water", impacts of soil conservation are always significant, and positive. Rather, we need to evaluate carefully where positive land-water linkages exist from the downstream users' perspectives that merit an investment on their part."

One such evaluation was shared during the conference. It concerned dairy farms in KwaZulu-Natal, South Africa. The most pertinent aspect was a report from Umgeni Water, in whose catchment area these farms lie. Each year, water quality is tested in addition to ongoing water flow measurements. The latest annual report states that the water sampled was of the highest quality measured anywhere in KwaZulu-Natal and the best recorded in 40 years of testing. In addition, water production had increased with minimal silt loads. This is after only about four years of no-till farming. This report is important because it contains probably the best justification for large-scale adoption of CA practices, emphasizing additional environmental benefits of this farming method: "This, in addition to carbon sequestration benefits, will surely help greatly in convincing several governmental departments of the need to seriously consider adopting CA as the national standard for agricultural production."

One participant made the point that where the impact of soil water conservation on downstream hydrology is small, then at the watershed-scale "badlands" or degraded lands that are not suitable for agriculture may contribute much of the sediment. However, as another participant asked: to what extent does a systematic approach to soil moisture management on a large scale actually influence water regimes downstream? In parts of India where the pressure on land and water is great, there is concern that watershed management approaches focusing on a better use of water at field or microwatershed level reduce the amount of water available downstream.

It was suggested that conventional systems and no-till systems should be compared with the natural system that existed in the watershed before humans began to place large demands on the soil and water resources. Moreover, perhaps we should imagine what the landscape looked like 200 years ago under native, perennial vegetation and wonder whether a mature no-till system looks more like that older system. In every case when this comparison is made, optimized no-till systems are clearly superior to any system that has tillage involved. Natural systems cycle water through macropores in a soil that is covered with litter. No-till systems do the same. Tilled systems have water running off the soil into the streams. Carbon does not decline in healthy natural systems and proper no-till systems, but it does decline in tilled systems.

As one contributor stated: "In many cases, human pressure on the resources (soil and water) is too great. In the case of competition between soil and animals for crop residues, if we put such pressure on natural ecosystems, they would also be degraded rapidly. This is exactly where the problem is. There are large areas in semi-arid regions where population pressure is probably too great for a natural ecosystem to survive. The question is then: What can we propose to the rural populations in those regions and what are the implications for agriculture and soil moisture management?"

The production of biomass for soil cover, as is recommended for local soil moisture management, will also have costs in terms of water use, and have a negative influence on the water balance

The costs of biomass as "production input" are easily offset by the cost reduction in land preparation and downstream damage. If soil conservation measures are adopted on a large part of the watershed area, increased evapotranspiration resulting from the maintenance of permanent soil cover could reduce water availability at the outlet of the watershed. This effect may be offset by the improved infiltration and aquifer recharge. However, this depends very much on the rainfall and the conservation measures adopted. Put simply, annual rainfall of less than 400 mm causes little or no recharge, while rainfall of more than 800 mm (particularly with large intense events) causes significant recharge. Between these two limits, recharge will only take place in areas where localized runoff ponds.

Soil moisture management practices do not only influence the hydrologic regime but modify nutrient cycling and organic matter decomposition processes as well

Where rain falls on a dry soil, the air in the soil pores is trapped beneath a downward-moving wetting front. Where the air cannot escape, the weight of water causes pressure buildup in the soil air until it is able to support the weight of the water, and thus slow down and halt further downward percolation. Such a problem of obstructed infiltration is worse in a de-structured soil than in one that has wormholes, old root-channels and large porous soil aggregates.

Only life-processes in soil can create such structures as continuous macropores, which create matrix areas of different levels of moisture and aeration within the soil. These facilitate soil life processes that affect nutrient dynamics. Nutrient cycling can be enhanced through aboveground and belowground residue-management planning. Diverse cropping systems have a net soil benefit as well as other economic benefits (drought strategy, pathogen management, labour distribution, farm income risk, etc.). Different crop types have a range of lignin and other organic products that can be used to advantage. Different crops will have different residue geometry and structure. Combined with chemical factors (e.g. lignin content), this will contribute to differential decomposition.

Relatively uniform moisture conditions in the soil are important for several reasons. The population and diversity of biota will be enhanced, resulting in improved nutrient availability for plants. In addition, crops are more tolerant to salinity, even where it is high, compared with the high levels of fluctuation in salt levels under flood irrigation as the soil moves from saturation at field capacity to varying degrees of desiccation. Therefore, improved soil moisture management (including irrigation techniques), especially combined with the use of organic fertilizers, can be a very effective means of overcoming micronutrient deficiencies. Reduced tillage and diverse cropping systems utilizing different rooting morphology, perennial crops, etc. can also stop or mitigate salinization. Where adoption of no till reduces the incidence or duration of runoff and flooding to depressional landforms, then less methane (CH₄) is generated - a good thing for greenhouse-gas mitigation.

The reason for mass movement of saturated soil from slopes was put forward as a result of the substitution of deep-rooted trees with shallow-rooting grasses (Box 6). One case presented illustrated that good practices can be harmful in regions where landslides are a prevalent phenomenon. An example was presented where, in order to avoid landslides, the infiltration rate was reduced to a minimum, and crude oil byproducts and other materials were used to reduce the infiltration rate and soil moisture storage.

One intervention discussed the disturbance imposed by driving over the soil with wheels carrying loads of more than 1 tonne (or perhaps less). A study found that this had a devastating effect on earthworms and all soil macrofauna. Microfauna effects were more varied.

Topic 4 - Adequate tools and technologies to support efficient soil moisture management

The modelling of the effects of reduced runoff (by improved soil and water conservation techniques) on the water-balance components (crop transpiration, soil evaporation and deep percolation) may assist in the understanding of the process and contribute to practical questions for the whole watershed, which includes downstream effects. Validation of model-simulation results by groundtruthing is necessary with present available indicators such as groundwater level and river discharges.

Available equipment and techniques are not adequate for on-farm measurement of plant-available soil water

There is a lack of available equipment and techniques for simple, routine, low-cost measurement and monitoring of soil water. Tools in the form of equipment to support soil moisture monitoring are usually inaccessible to most poor farmers. It is important to develop low-technology field-useful approaches that give farmers and land users simple field-applied techniques and models for determining the onset and severity of drought. Preferably, methods based on the experience of the farmers themselves in estimating watering needs should be developed. These methods should be based on the observation of plants and on empirical evaluation of the humidity by touching soil samples.

One such method is the measurement of evaporation. Compared with the calculation of potential evapotranspiration, it is very simple, acceptable by farmers and includes meteorological conditions (solar radiation, relative humidity, air temperature and wind). Based on the measured evaporation, farmers know exactly the quantity and time to irrigate their crop. Another method is the use of "indicator plants" in order to avoid relying on instrumentation altogether. The purpose is to provide a visible warning of impending drought stress within a commercial crop. In addition, it is important to consider plant roots as their distribution better integrates the soil volume than do most installed field sensors.

One system of providing a simple, cheap, field-usable measure of soil hydraulic conductivity (soil water infiltration) that does not require trained people to measure or simulate and process the information is now available. It is called the Visual Soil Field Assessment Tool (VS-FAST). The method is simple and robust, being based on fundamental, globally tested and accepted soil physical principles. This ensures that the operator (seen principally as being farmers and local extension staff in developing countries) can use the technique after basic training.

The soil probe, mentioned in the discussion paper "Adequate tools and technologies to support soil moisture management", is useful for determining stored soil moisture in order to make decisions on continuous cropping and what crop type to pick. Dr Paul Brown of Montana State University developed the probe in the 1960s. It gained popularity in the 1970s and 1980s as farming practices in the Northern Great Plains in the United States of America decreased the use of a summer fallow year in order to conserve soil moisture. A major feature of the probe is the inclusion of a piece of a wood auger below the ball on the tip. At any point of insertion of the probe the handle can be given a quick twist and the probe removed. A small sample of the soil at the depth of insertion can be brought to surface and the degree of soil moisture can be determined by feel and appearance. Similarly, where there is more resistance to the probe insertion, a small sample of soil can be brought up in order to verify that it is a dry layer.

For some years now, many Australian irrigation farmers have successfully used capacitance probes to monitor soil water content. The probes are installed, calibrated and maintained; and the data analysed and interpreted for the farmer by local commercial companies.

Plate 1
Improved soil probe with an auger tip to bring small amounts of soil to the surface to be examined by feel and appearance method

T. GODDARD

The type of model for on-farm estimation of plant-available soil water does not secure reliable results

"Breaking down reality" - the reductionist approach - does not necessarily reveal or anticipate how actual combinations of factors will affect crop development.

A new family of models called artificial neural networks (ANNs), which are neither reductionist nor mechanistic, have the potential to be useful in the management of soil moisture. The ANN methods embody to some degree the non-linearities that characterize the combined effect of variously timed water applications. The notion of relating water inputs directly to yield may seem unreasonably simplistic to some. However, for irrigation planning purposes at least, there appears to be some value in applying a non-linear algorithm in a straightforward manner.

Another approach is the "least limiting water range" (LLWR) - this being the water content range within which plant growth is least limited. The strength of this approach is the integration of three physical factors affecting plant growth directly, but its use still is limited by difficulties of some soil related property measurement. As it is conceptually well based, it should be used as a simple procedure by farmers and field technicians to evaluate soil quality and propose mitigation management techniques in order to overcome the soil physical constraint factors.

Plant physiology should also be taken into account when water availability to plants is assessed. This is because the efficiency with which the plant uses water can vary considerably.

Soil moisture monitoring needs trained people to measure or simulate and process the information

The available methods for soil moisture measurement (e.g. gravimetric, tensiometry, gypsum-porous blocks, electrical resistance method, neutron probes, capacitance probes, etc.) all need trained people to apply them and they are not acceptable to or usable by small farmers. Some of them are very expensive and need laboratory measurements and field calibration. In addition, the results of measurements by these methods are not precisely exact. Moreover, they need accompanying methodologies for translating the results into a soil management strategy. Many provide only an approximation or a trend until properly and fully calibrated or cross-checked against crop variables.

Measured (laboratory) or simulated concepts, variables or indicators are needed to generate information to support soil moisture monitoring

Process-based models are difficult to design. There are issues of selection of variables and functional forms, and they need testing with observed data. It is extremely difficult to include interactions between factors in models. Any model is a simplification of the reality. However, from the scientific point of view, choosing a model is also selecting a priori phenomena that will be more important than others for describing, exploring or explaining a situation. Based on the example from the working document and the reality faced by farmers in Brazil, it was suggested that "different kinds of farmers may decide to go or not to go for residue application for entirely different reasons and, so it seems, on perfectly rational grounds." Therefore, a coherent analytical framework, clearly stating for what circumstances the observations and experiences are valid, is needed so that a greater focus can be achieved.

A holistic approach to soil moisture management is needed. While studying a system functioning, and in order to find the proper explanation for a given outcome, it is normally necessary to examine the area of interactions and to understand and quantify them.

The short-term benefits and long-term externalities induced by good soil moisture management must be evaluated, discussed and published

There is a need to concentrate on identifying natural soil properties that lead to some soils providing more moisture to crops particularly at mid-season droughts or at the end of a generally short growing season. In addition, the identification of management practices that optimize soil moisture supply to crops is vital, which must depend on local conditions of climate, soil and crop. It is the scientific understanding of how such practices actually operate that will enable:

• generalization into general principles;
• upscaling to larger areas;

• identification of alternative options that can vary with local circumstances and availability of resources.

Topic 5 - Conditions for adoption of drought-proofing practices by farmers

Technology adoption and adaptation has only been successful where adequate attention was paid to farmers' initiatives and their interactions on "products" released by research, extension services and the private sector

Farmers have to be involved in the development of new technologies from the outset. The more changes required of farming systems, the more essential farmer involvement becomes. Most technologies on soil drought-proofing require a profound change in farming practices, not just the change of a component as with improved varieties or fertilizers. This means that farmers have to learn how to integrate new practices into their systems. They will have to pay for this learning with setbacks in the first few years. Successful adaptation usually takes some time. Pioneers in a community are viewed critically and risk becoming outsiders. This is especially valid for Africa, where strong farmers associations are lacking. Added to this is insufficient access to information, inputs, credit, etc. Support programmes are often of too short a duration, e.g. 2 years, followed by programmes with new foci. Farmers want advice on how to improve their way of farming and how to become more profitable. They do not want technological packages as the interactions of the components are difficult to understand, nor do they want to feel that they are being used as guinea pigs by researchers and extensionists. Researchers and extensionists are often too output oriented, forgetting the production costs and the risk of not covering even production costs when adequate markets and prices are lacking.

The implementation of improved soil moisture management systems will be hindered by government legislation, incentives and subsidies

The word subsidy is loaded with different interpretations in different countries and different circumstances. In most industrialized countries, agriculture subsidies consist of input and price subsidies. These subsidies have had devastating impacts on the components of agricultural systems through soil erosion, water pollution, loss of biodiversity and genetic erosion, loss of habitats and buffer zones, etc. and on farmers (e.g. smallholder poverty, abandonment of small family farms, low food prices, and dependency of farmers on price fluctuations). Recently, in Europe and elsewhere, subsidies have also been applied as a payment for environmental services of environment-friendly agricultural practices. Although subsidies and legislation can be an obstacle to the adoption of sound soil management practices, they can also become a very effective tool for their promotion when they are used to support the development of sustainable on-farm soil management investment. "Moderate" subsidies can become a useful incentive for farmers to change existing habits. For example, in the Netherlands, the Government assisted the large-scale introduction of greenhouses using a well-designed subsidy strategy. If this kind of subsidy was "right" for a developed European country, why is it "wrong" for developing countries? The issue is to what point subsidies will prevent producers from improving their practices because they make as much money with the subsidies as they would by improving their production. Farmers do not feel they need to change anything. Moreover, those who are willing to change find it difficult to adopt crop rotations that do not include the subsidized crops. Apart from industrialized countries, few governments now pay subsidies or provide other incentives to farmers. Governmental support of the agriculture sector in the United States of America was mentioned as an example during the discussion: "If the majority of farmers in the United States of America were playing by the real-world market rules, as the majority of the South American farmers are, they would go out of business unless they adopted technological changes to reduce costs, like using CA extensively or, directly, no till."

The main reason for farmers and communities not to implement appropriate soil moisture management is lack of information, education and training

A variety of factors can contribute to the adoption or rejection (and, hence, failure or success) of technically perfect technologies. While economics is not the least of these, other factors are also important, such as labour requirements, different priorities, and local sensitivities (e.g. reluctance or suspicion of government-driven initiatives).

It was argued that among all factors influencing the adoption rate of soil moisture management techniques, the economic factor ranks far ahead of considerations such as education, training and information. The simple fact that it is difficult to maintain or replicate pilot experiments shows that, even when a technology is known, other factors must play an important role. The adoption of CA by farmers in the states of Paraná and Santa Catarina in Brazil undoubtedly contributed to their being able to survive the combined effects of low output prices and increasing input prices in the 1990s. It was found that the most important reasons for not adopting soil fertility options in the Sahel were:

• farmers lack capital, and credit systems are poorly developed and relatively inaccessible;

• support is in many cases required during the transition period as soil depletion is financially far more attractive for farmers than soil improvement;

• creation of an enabling environment on a regional or national level requires important investments.

Microeconomists would say that farmers are perfectly rational and maximize some utility function. It should be well understood that this maximum utility consists of different attributes, whose mix varies with each farmer. Farmers know the advantages and disadvantages of their production systems very well. Consequently, they will often be reluctant to adopt technologies that deviate drastically from a system that they know well and with which they have considerable experience with respect to its performance under variable weather conditions. For example, the farmers in the Shiwaliks Belt (one of the most fragile regions from a soil erosion point of view in India) of the Himalayan ecosystem are illiterate and bound by tradition. It has been observed that, being illiterate, they are not aware of various programmes of research and extension being carried out by different agencies. As such, they are not in a position to know about emerging technologies. Second, being bound by tradition, the farmers are not prepared to forgo the existing practice and adopt new ones (Box 7). Farmers should be involved while carrying out research on a particular problem. In this way, they will learn about the new technique and will consider it their own invention. The major obstacles identified in southwest Nigeria were:

• Solutions not relevant (research agenda). Farmers perceived them as unnecessary because of tradition and old methods.

• Traditional and cautious behaviour of farmers. Existing knowledge attitude and practices were not initially considered carefully by researchers.

• Insufficiently qualified and motivated extension services.

• Communication plan. Research results were not communicated properly and adequately to end users.

In this case, the role of a scientist should be to develop of a drought-resistant variety of black gram rather than insist on the adoption of green gram.

Extension staff can appear not motivated when, in fact, owing to a lack of training, they are fearful or ignorant. They need to have the training and practice. Then they will gain in confidence and be successful. Farmers are reluctant because they are comfortable with what they have always done and they "fear" failure with adopting a new system. Moreover, they usually have the equipment to continue doing what they have always done and they are either unable or unwilling to invest in the equipment necessary for a change in the production system. This confirms the view that farmers who are comfortable with their current situation would not look for a change, especially if it is an unknown and, hence, risky one. A number of conditions necessary for adoption can be summarized:

• the technology should result in an immediate and significant improvement in farmers' profits;

• the benefit/cost ratio should be high;

• the investment (monetary and labour) should be within the fiscal and physical capacity and within the existing work habits and traditions of the farmers;

• there should be frequent follow-up;

• an integrated (holistic) participatory approach should be taken;

• credit should be available;

• a well-designed marketing strategy is necessary.

As there are alternatives to feeding livestock, farmers do not really have to choose between using crop residues for soil cover or as feed for cattle in arid regions

This is not valid for most dry regions and it is a significant problem that needs to be addressed. Livestock feed is generally in short supply, and the use of crop residues is the easiest way to feed livestock. In the short term, there are some farmers, particularly in dryland areas of developing countries, who do not have a choice and need to use crop residues for animal feed. In most parts of Africa, the right of land use is limited to the growing season. Fields are open for communal grazing after harvest. It is difficult for individual farmers to prevent livestock entering their fields. This is also valid for cover crops, which are commonly grazed by freely roaming cattle or goats.

This is also the case in the dairy industry in Uruguay, which has a completely different climate (temperate subhumid) from the semi-arid areas of Africa. In dairy farms, almost all the aerial biomass is grazed, harvested for hay or silage, or trampled by the animals. Nevertheless, the crops of these systems (oats in winter, and corn or sorghum in summer) are in rotation with perennial pastures that are also grazed. However, they leave a very important amount of roots in the surface ploughed horizon and develop progressively a mulch of residues on the surface in the 3-4 years that they occupy in the rotation.

In the Shiwaliks region, farmers are reluctant to use crop residues for mulching as they have inadequate sources of fodder for their cattle. Farmers prefer to use locally available mulch materials such as the leaves of some local trees for moisture conservation or they adopt other techniques for this e.g. breaking capillaries in order to check evaporation losses. Here again, the aim should be to develop the existing technology of using locally available mulch materials.

A primary role for the international breeding institutions is the selection of varieties that produce more straw (especially important in integrated livestock systems).

In general, conservation agriculture is perceived as more profitable by farmers, while researchers and extensionists perceive it as more risky.

Only about 3 percent of the land in the world is being farmed using CA principles.

As one participant stated: "If farmers truly perceived CA as more profitable, I cannot believe that more of them would not be adopters. There are many reasons for not adopting, but tradition and fear are among the main ones. My perception is that they do not see enough upside potential for them to make a change. They would if they really perceived more profit in changing."

Another participant stated that the reason for the lack of uptake is that farmers usually do not know of CA and its potential. The critical mass of this knowledge has not yet been reached. In addition, there are forces working against it, e.g. the oil industry and the agricultural machinery industry, that have paramount importance and power in developed countries. It is not difficult to imagine that they would not be happy with a technological change that reduces oil use and replaces the use of huge machinery with a few machines and medium-sized tractors.

REFERENCE

Derpsch, R., Roth, C.H., Sidiras, N. & Köpke, U. 1991. Controle de erosão no Paraná, Brasil: Sistema de cobertura do solo, plantio direto e preparo conservacionista do solo. Sonderpublikation der GTZ, No. 245. Rossdorf, Germany, TZ-Verlagsgesellschaft GmbH. 272 pp.