Chapter 3. Farming systems in drylands

Introduction

According to FAO (2001a), a farming system is defined as “a population of individual farm systems that have broadly similar resource bases, enterprise patterns, household livelihoods and constraints, and for which similar development strategies and interventions would be appropriate”. Depending on the scale of the analysis, a farming system can encompass a few dozen or many millions of households. The understanding of the major farming systems in drylands provides the necessary framework for the development of agricultural strategies and interventions. Based on the classification of farming systems of developing regions specified by FAO (2001a), most of the farming systems in drylands fall into the category of rainfed farming systems in dry lowpotential areas. These systems are characterized by mixed crop - livestock and pastoral systems merging into sparse and often dispersed systems with very low current productivity or potential because of extreme aridity or cold.

Understanding the world of smallholders in dryland environments is the key to designing appropriate and successful CS activities. It is important to understand that CS for poverty alleviation must be much broader in terms of the range of both practices and benefits (i.e. not only in monetary terms) than similar schemes in commercial agriculture and forestry.

There are various potential partners, or target groups, for CS programmes in drylands. From a purely scale-driven perspective, large-scale capital-intensive agriculture might be the most attractive. However, from a biophysical standpoint, as discussed in Chapter 4, systems that use significant quantities of fertilizers or that depend heavily on fossil fuel to supply irrigation water should not generally be considered because they are usually net carbon emitters. Only if a switch from high fertilizer and fossil-fuel dependence to more carbon-friendly inputs, technologies or land use is foreseeable in the short-run, should current large-scale agriculture be considered. There are some systems of capital intensive land use, such as the mechanized farming schemes in eastern Sudan (where large areas of land have been severely degraded) that offer great potential for soil CS if rehabilitated through low-intensity land use.

Apart from these technical reasons, large-scale, capital-intensive agriculture systems are probably not potential partners for soil CS because the small additional income that sequestration might bring would be unattractive in comparison with profits from other sources, many of which depend on carbon-emissive techniques.

Thus, the main target groups of soil CS in degraded agro-ecosystems are primarily small-scale, resource-poor farmers in uncertain and risk-prone environments for whom anticipated benefits could constitute an enhancement of their livelihood. Reference to these groups of farmers is made as smallholders. They depend on lowinput, subsistence-based agriculture, and they are usually characterized by diversity, variability and flexibility (Mortimore and Adams, 1999).

Characteristics of smallholder agriculture

The primary characteristics of smallholder agriculture in semi-arid developing countries are its diversity in space, its variability through time, and its multidimensionality in terms of the ways it operates and survives (Mortimore and Adams, 1999). This is largely because dryland smallholders must be highly responsive to a varied, changeable and hazardous environment. Thus, their operations are very different from those of large-scale farms driven by commercial goals, equipped with credits and efficiencyoriented technologies and covered by insurance systems against hazards and losses. This diversity, variability and multidimensionality means that each particular system must be approached with careful attention to its unique mix of characteristics.

Another important characteristic of smallholders, which also differentiates them from commercial farmers, is that few are motivated solely by the goal of agricultural profit. Instead, smallholders pursue basic subsistence and survival goals, balancing daily risks and opportunities directly through their livelihood options and management practices rather than through external institutions (Collinson, 2000). Many smallholders have deep attachment to their land, which they continue to farm, even when profitability is low, for reasons such as maintaining tenure, and maintaining family ties. At the same time, many have additional, and often higher, incomes from non-agricultural sources. These include: petty trading; the gathering of wild produce, including firewood; labouring; and remittances from family members. The result is “multi-enterprise production units” (Hunt, 1991).

Smallholders are further differentiated from high-input commercial farmers by their need to manage multiple risks. Almost all of their inputs and outputs are subject to large variation and uncertainty, such as labour, which is often the most critical variable. Another critical risk arises from the high variability in rainfall, which itself has two major consequences as far as sequestration is concerned. One is variation in the timing of bioproductivity, which means that planting and harvesting (and most other agricultural and non-agricultural activities) may have to be readjusted rapidly, sometimes within a season, and often between seasons. For example, fallows that appeared secure for years may have to be cleared after a particularly poor season. The other consequence is variability between fields, some of which may receive sufficient rainfall, and some of which may not. There are other risks that have similar consequences. These include: attacks by pests (against which pesticides are too expensive); illness, resulting in the unavailability of labour at some critical point in the season; and variability regarding prices of inputs such as seed, labour, food, and of outputs, mainly crops.

According to Mortimore and Adams (1999), smallholder responses to these various constraints follow three key avenues: (i) diversification of natural, economic, technical and social resource endowments with the underlying rational to spread risks as efficiently as possible; (ii) flexibility in day-to-day management of these resources in the form of active decisions to cope with and adapt to short-term variability; and (iii) adaptability over the longer term, perceived as cumulative and purposeful decisionmaking that will result in new or altered systems or livelihood pathways. When spreading risks, it is important for farmers to have a mix of products where both the type of products and the price of these products are independent of each other, a criteria that potentially applies very well to CS.

A further characteristic of smallholder agriculture is variable access to resources of all kinds. Within a village, some have ready access, and others have less access, to: secure landholdings; wild produce, such as fuelwood; credit; hired labour; livestock; and markets. Access also varies between villages and between countries. The implications of such uneven access to resources for sequestration schemes are discussed below.

Finally, these agricultural systems are and have long been undergoing continual change in response to environmental and social changes. Dry environments are now widely recognized as having a complex history of change, based on nonequilibrium dynamics rather than predictable, gradual and linear change (Leach and Mearns, 1999; Scoones, 1999; Scoones, 2001), sometimes referred to as event-driven systems (Reenberg, 2001; Sorbo, 2003). Thus, agricultural systems have had to adapt continuously to environmental conditions and to changing political and economic processes. In the lifetime of a soil CS scheme, one could expect many changes in the configuration of the agricultural landscape, apart from the changes that the project itself might bring. Planning in such an environment will be challenging. Instead of simplified and standardized approaches and predefined technical solutions, CS schemes in these systems will need to offer a range of technological and management options from which farmers can choose according to their needs.

Examples of smallholder farming systems

Within this broad description of the characteristics of dryland smallholder agriculture, there are various farming systems. These are systems such as annual croplands, plantations, forests, savannahs, natural pastures, fallow lands and vegetable gardens. Within each, there is a specific interaction between crops, livestock and trees, and between cultivated and non-cultivated land (FAO, 2000a).

Farming systems in drylands range from shifting cultivation embedded in extensive wooded grasslands to intensive smallholder farming where all land is under cultivation and the integration between cropping and animal husbandry maximized. However, these two extremes should not be understood as fixed points along an axis of agricultural development, but rather as examples of “pathways” of agricultural and environmental change (Scoones, 2001) that are possible both between and within sites. Such pathways of change reflect farmers’ livelihoods, constraints and opportunities within a historical context. Figure 7 provides a schematic illustration of dryland smallholder farming systems.

Agricultural intensification

Intensification, as defined by Tiffen and Mortimore (1993), implies “increased average inputs of labour or capital on a smallholding, either on cultivated land alone, or on cultivated and grazing land, for the purpose of increasing the value of output per hectare”. Intensification takes many forms, which can be classified in many ways. In the case of smallholder dryland farming systems, intensification tends to be related to increased local labour inputs per hectare and low-cost technologies rather than capitalintensive innovations. Mortimore and Adams (1999) describe such intensification as an “indigenous and adaptive process” whose path can be reconstructed through historical analyses.

There are many examples of such “indigenous” intensification. In rainfed farming systems, intensification often occurs as a consequence of growing population pressure. In many places, fallow periods have become shorter and shorter and eventually even abandoned. All fields may then be under cultivation and soil fertility is maintained by greater labour intensity. Techniques may include: intercropping with N-fixing legumes; time-intensive weeding and harvesting; the utilization of manure and mulch; and the protection of certain tree species. Crop rotation is practised where possible to ensure differential nutrient use and uptake between crops, such as millet and sorghum, and Nfixing crops, such as groundnuts and cowpeas. Trees, especially those known for their N-fixing and soil-restoring capacities, are protected. The application of manure, either from cattle or small ruminants, is a key element. In order to maintain supply in the face of increasing land scarcity, herds must be managed more intensively, e.g. feeding them with agricultural residues and weeds.

In dryland areas where sufficient surface water is available, irrigation has been a key method of intensifying land-use systems since ancient times. It requires supplies of water and of energy to take the water to the fields and gardens. The water may come from streams, rivers, springs and wells. Streams and rivers may be range from small ephemeral water courses, as in many parts of central Asia, to major rivers such the Nile, Niger, Amu-Darya, Hwang He and Indus.

Where there is a good “head” of water, as in mountainous terrain (e.g. in parts of highland Yemen and Oman), or in the large systems on floodplains, as of the Indus and Nile rivers, the water may be taken by gravity in small channels to the fields or gardens. Where the river flows in a gently sloping floodplain, the methods raising it to the fields of small schemes are much the same as from wells: animal or human-driven devices such as saqqias, Archimedes screws and shadufs. The qanat systems, which are particularly well developed in Iran and neighbouring areas, but are also found in other parts of Asia and in north Africa, are more elaborate, involving wells from which water is channelled underground, and fed to the fields by gravity. Another ancient system, which has seen major expansion and development in recent years, is water harvesting (or runoff agriculture). In this system, highly intermittent runoff is concentrated and then held in shallow troughs, where it is usually used for tree-crops.

FIGURE 7 Smallholder farming systems in the Sahel and management strategies in the context of carbon

Source: Tschakert, field work, 2001.

Extensive land use

In areas where both population densities and rainfall are low, patterns of extensive land use predominate either as a longer-term system state or a more recent major pathway of change (Mortimore and Adams, 1999). The latter is true for some areas in central Senegal where entire compounds have migrated recently to the city of Touba, leaving relatives and neighbours with more land available than in previous decades (Tschakert and Tappan, 2004). As land scarcity does not represent a constraint in this case, fallow lands constitute an important element of the farming system, allowing for short- and medium-term soil regeneration. In general, field sizes are significantly larger than in areas under intensification. Given the amount of land available to individual households, manure is generally only used for fields that are under continuous cultivation, primarily those adjacent to the settlements and others in close proximity. Remote fields and those left fallow are accessible to grazing animals all year round. Unlike animals in intensified systems, herds are not forced to leave for transhumance and thus contribute to a continuous flux of organic matter input. Weeding and harvest activities might occur with less intensity, while more agricultural residues are left on the fields.

Agroforestry may play an important part in these extensive systems. One example is the Sudanese system of gum arabic production, where a tree that regrows on fallow land is a major source of income for smallholders (Elmqvist and Olsson, 2003). In other long fallows, trees that yield other useful products, such as fruits, nuts, fibre and medicines, are planted. Trees also provide an important source of emergency food.

The above examples of intensive and extensive farming systems illustrate that a context-specific approach based on multiple pathways of change offers useful guidelines for potential CS schemes. Project design and implementation should start with a local understanding of environmental change and its underlying processes. The next step is to identify positive pathways of change at the local level and then finally to assess opportunities to encourage such pathways on a larger scale.

Soil fertility management

The concept of CS in degraded agro-ecosystems is typically based on two assumptions. The first is that any improvement in soil fertility management and land use will result automatically in higher amounts of C sequestered from the atmosphere and stored in soils. The second is that local smallholders and herders, who are anticipated to be the prime beneficiaries of planned interventions, need to be made aware of and trained in such improved management practices.

Given the complex, diverse and dynamic world of smallholder farming in dryland environments, these two assumptions seem oversimplified. In general, proposed management practices and land-use options merely reflect the most efficient technical options, focusing on achieving an optimal agronomic situation. However, as illustrated above, smallholders are more concerned about day-to-day risk management and longer-term adaptive strategies than the achievement of an assumed new equilibrium. Opportunistic farming is all about spreading risk, an adaptive process during which both losses and gains occur, often intentionally. Pure “efficiency would leave no room for flexible maneuver” (Mortimore and Adams, 1999).

What in fact constitutes “improved” soil fertility management or land-use options might be understandable only from a holistic farming-system research approach. Farmers who have developed highly dynamic and flexible soil-fertility management practices to cope with variability and uncertainty are often in the best position to bring this holism to a development project. Although farmers often have much experience in deliberating technologies within a much broader framework of “real life”, they are most often considered as passive recipients of outside assistance rather than key resources in the process itself.

Thus, a first step towards linking soils and C to people is to investigate practices that smallholders in dryland environments currently know and use, to understand their underlying rationale as well as driving factors for change, and to identify examples of positive pathways of change that could be replicated on a larger scale (Tschakert and Tappan, 2004).

Soil fertility management practices can be grouped according to the movement of nutrients into, within, and out of a system. Here, practices are sorted into four groups (Hilhorst and Muchena, 2000): (i) adding nutrients to the soil; (ii) reducing losses of nutrients from the soil; (iii) recycling nutrients; and (iv) maximizing the efficiency of nutrient uptake. The examples below are based mainly on the Senegal and Sudan case studies.

Adding nutrients to the soil

Fallowing

Fallowing is a well-known practice for replenishing nutrients in soils. Ideally, fallow periods are rotated with cropping periods, allowing the land to recover from years of cultivation. However, in many parts of the world’s drylands, both fallow area and duration have decreased over time. Most often, this decline is caused by increasing population pressure, the introduction of modern agricultural machinery, such as the plough, and periods of droughts, or a combination of all three. Some believe that this process is reaching crisis proportions. Today, in many drylands, fallow duration is reduced to only one year. In areas with severe land scarcity, it has disappeared altogether. As a consequence, farmers have shifted to other soil-fertility management practices, such as manuring and composting (see below) or they continue to farm with exceptionally low and decreasing yields. At the same time, less land in fallow also means reduced grazing possibilities or less fodder for animals, thus reducing the amount of manure that can be produced (Breman, Groot and van Keulen, 2000). Nevertheless, in areas with less population pressure, fallowing still constitutes an important option for soil-fertility management. This is particularly true for countries where structural adjustment packages have been implemented and subsidies for fertilizers removed.

Stubble grazing

Many farming systems include the grazing of animals on fields immediately after crop harvests. Animals graze stubble and stalks left on the field, while soils benefit from faeces deposition throughout the duration of the practice. Depending on the size and the forage situation of a field, as well as the overall number of animals, livestock is usually kept for 1 - 7 months on the same field, where it is rotated between different parts during shorter time intervals.

Overall, the gain in organic matter from stubble grazing can be substantial. In the Sahel, deposition of droppings ranges from 1 tonne/ha to 50 tonnes/ha depending on the time that animals are kept on the same field (Sagna-Cabral, 1989; Garin and Faye, 1990; Hoffmann and Gerling, 2001). However, direct exposure to the elements can reduce the nutrient value of dung and droppings considerably. Although stubble grazing has a long tradition in drylands, increasing land scarcity, limited purchasing power among many smallholders, and increased risks of animal theft in many areas have contributed to a general decline in herd sizes and, in some cases, led to the abandonment of stubble grazing altogether.

Inorganic fertilizers

The use of inorganic fertilizers has been one of the most widely promoted means for increasing production since the early twentieth century. In many of the drylands of the developing world, this kind of fertilizer was subsidized and made available to farmers with the aid of government and the support of non-governmental organizations (NGOs). Under structural adjustment programmes, subsidies were often removed and, hence, fertilizers became increasingly expensive for farmers. As is shown in the Senegal example (and illustrated in Figure 8), the use of fertilizers decreased in the 1990s. From a CS point of view, the use of synthetic fertilizers does not result in any net gain of carbon fixation (Schlesinger, 1999). The emission of CO₂ during the manufacture, transport and application of the fertilizers offsets any gain in biological production.

FIGURE 8 Changes in land use and soil fertility management, expressed in weighted points of importance/extent (1-10), as perceived by farmers in an intensified farming system in Senegal

Source: Tscharkert, field work, 2001.

Crop rotation and association

The practice of crop rotation and association, especially where it involves both cereals and legumes, is well known among farmers as a soil-fertility management practice. In many places, N-fixing crops include beans and groundnuts. However, in farming systems where land scarcity has become a limiting factor, priority is often given to cereals. Moreover, the availability of seeds for legumes might be dependent on state subsidies or credits, as was the case with groundnuts in the Sahel.

Reducing losses of nutrients from the soil

Woody vegetation

Trees can be an important component in many agro-ecosystems. With their deep and extensive root systems, they can capture nutrients not accessible by crops and make them available to crop production again through litter fall. From a CS point of view, not only do the trees store C in their aboveground biomass, they also contribute to belowground biomass through their root systems and their input of litter to the soil (branches and leaves). Of particular use are leguminous (N-fixing) trees, among which Faidherbia albida and Acacia senegal are two of the most appreciated. Trees may also play a role in reducing nutrient losses from wind erosion. Special consideration must be given to the use of biofuel instead of fossil fuel.

Living hedges and fences can capture silt and clay particles suspended in the air and could, hence, locally increase the clay content of the soil, a factor beneficial for CS (El Tahir and Madibo, in press). Litter produced by woody plants is beneficial because of its higher content of polyphenols (lignines and tannins), which decreases the decomposition rate (Abril and Bucher, 2001), when compared with grasses and annual herbs.

Erosion control

Erosion and subsequent transport and deposition have a complex relationship to soil carbon storage. Where water erosion dominates, a high proportion of soil C may be washed into alluvial deposits close to the erosion site, and stored there in forms which decay more slowly than in the parent soils. Therefore, this kind of erosion may have a positive effect on soil CS. Erosion does not always decrease productivity, but if it could be shown to do so, it would be perverse to favour decreased productivity for a mediumterm and perhaps one-off gain in sequestered C. The same arguments probably do not apply where wind erosion is the main erosional process, for organic matter is usually blown great distances and dispersed to places where it may decay rapidly and release its C. Management options that increase the amount of live and dead biomass left in agricultural areas decrease erosion in general while simultaneously increasing the C input to the soil (Tiessen and Cuevas, 1994).

Field clearing and weeding

Clearing the fields from weeds before planting, as well as weeding during the cropping season, is an important practice for reducing competition between the crop and weeds. However, from the point of view of soil fertility and CS, it is important to recycle as much of the weeds back into the soil as possible. Selective clearing and weeding implies that only weeds competing directly with the crop are removed while others remain in the field.

Recycling nutrients

Manure

Spreading manure from livestock that is kept within or close to the compounds is one of the most widespread soil fertility management practices. Farmers are well aware of the fertilizing effects of manure but they also appreciate it for the fact that it stabilizes sandy topsoil and reduces wind erosion. Farmyard manure (FYM) and manure produced in pens is usually of higher quality than dung and droppings left on fields by grazing animals. It may be combined with agricultural residues, household waste and ashes accumulating within a household. The most limiting factor of the usage of manure, in addition to lack of animals, is the lack of transporting material, often resulting in well-manured fields closer to the homestead rather than on more remote fields.

Crop residue management

Crop residues such as stalks and hay can be left on or returned to the field in the form of mulch or ploughed under at the end of the cropping season. However, in most dryland farming systems, crop residues are contested and are removed after harvest either as fodder, construction material, fuel or litter for composting. What remains on the field is often burned before the next cropping season. In some cases, crop residues are also sold at the local market, generating additional income.

Management of other organic matter

Household waste, fish scales, ashes, leaf litter, prunings, and surplus crop residues are also used to increase soil fertility. Often, these additional inputs are accumulated within the homestead, sometimes added on to manure piles, and then transported to fields where they are spread according to nutrient needs. In various places, composting has resulted in improved decomposition rates. Although the use of such alternative organic matter, primarily household waste, has been increasing, overall quantities are rarely sufficient to fertilize entire fields in a sustainable manner.

Maximizing the efficiency of nutrient uptake

Reduced land tillage

Although some farmers appreciate tillage for weed control and soil aeration, there seems to be a growing recognition that tillage also destroys the protective vegetative cover and, as a result, exposes soil nutrients to the elements. In areas where ploughs and draught animals are available to smallholders, tillage is still widely practised. In other areas, such as the Senegalese Peanut Basin, farmers have replaced deep tillage by superficial tillage, primarily because of the lack of machinery (Tschakert and Tappan, 2004). For the purpose of CS, reduced or no-tillage are preferable, simply because they enhance the storage of C in the soil.

Precision agriculture

Many farmers often match crops and management practices with the fertility status of specific fields and, on a smaller scale, with specific spots within one field. Given the relative scarcity of organic and inorganic inputs, the available quantities are spread following a patch-by-patch scheme.

Fire management

Fire is a very common tool for land managers in drylands. Fire is often used to clear the fields of weeds before planting. Another important reason for preplanting fire is to kill a range of agricultural pests. The role of fire in the soil carbon balance was investigated by modelling and found to have a significant effect on SOC. When the fire-return period was increased from 3 to 15 years, the SOC level increased by 30 percent (Poussart and Ardö, 2002).

These descriptions of individual methods of soil fertility management do not capture the full complexity of the ways in which they are combined. Some of this complexity is described in more detail below.

Soil fertility management practices in the Sahel

In drylands, farmers know and use a whole range of soil fertility management practices. However, these practices may vary from farming system to farming system, from farmer to farmer, and from field to field, and even within fields, depending on differential access to and utilization of resources. To illustrate the complexity of soil fertility practices, Table 9 presents a detailed example from Senegal.

In addition to spatial variability, soil fertility management practices tend to vary over time. As farmers adapt to risks, shocks and uncertainty over the long run and new or altered systems or livelihood pathways emerge, farmers’ portfolios of management practices also change. Figure 8 illustrates changes in management practices in a village that has followed a pathway of “indigenous and adaptive” intensification (Tschakert and Tappan, 2004). With increasing population pressure and land scarcity, a general shift from extensive management practices (fallowing and stubble grazing) to more intensive strategies (application of manure, household waste, compost, planting trees and fences) has occurred. This transition is overlaid with a change in governmental policies, reflected in the disengagement of the state after 1980, primarily following structural adjustment, implying reduced or no subsidies or credits for mineral fertilizer, groundnut seeds, and agricultural equipment.

TABLE 9
Example of soil fertility management practices used in the Old Peanut Basin, Senegal, 1999/2000

¹ Between December 2000 and December 2001, fourteen villages in the Départements de Thiès, Fatick, Bambey and Diourbel participated in a study on soil fertility management and carbon sequestration.

² Classification of soil fertility management practices after Hilhorst and Muchena (2000).

³ “dior”, “deck” and “ban” are Wolof names for the dominant soil types in the Old Peanut Basin. According to Badiane, Khouma and Senè (2000), “Dior” are common on former dune slopes and usually contain >95% sand and <0.2% organic C; “deck” are hydromorphic with 85 - 90% sand and carbon contents between 0.5 and 0.8%. “Ban” are similar to “deck”, usually found along waterways and basins (“bas-fonds”).

Source: Tschakert, fieldwork, 2000-01.

Building on local knowledge

In addition to the practices, it is important to understand farmers’ theories of soil fertility, soil formation, and the processes that cause losses and gains of soil fertility over time.

For example, smallholders in the Senegalese Peanut Basin perceive soil fertility as “saletés” (dirt), a generic term for organic matter inputs (manure, decomposing plant material, household waste, etc). To farmers, this “dirt” contains nutritious elements, referred to as “vitamins” or “tasty ingredients” that determine the strength and health of a soil. Although the majority of farmers lack detailed knowledge with respect to the origin of such “vitamins”, they are aware of the various processes resulting in soil degradation and fertility losses. The most frequently cited causes of fertility decline include: continuous cultivation without external inputs or crop rotation; reduction of protective vegetation cover; and exposure of SOM to the elements as a consequence of tree removal, deep tillage, and too short fallow periods; bush fires; and harmful insects. Accordingly, farmers’ preferred options to restore soil fertility focus on crop rotation, increased organic matter inputs, and accumulation of vegetative cover, primarily through longer-term fallowing and increased tree density.

In parts of Niger, soil fertility is seen much more holistically by farmers than by agronomists, who disaggregate the influences on crop productivity into factors such as water supply, water uptake, individual nutrients, and soil structure (Osbahr and Allen, 2002). The farmers know that the productivity of different soils is determined by a combination of factors. In a wet year, clay-rich soils in depressions may be waterlogged and unproductive, while sandy soils, where managed adequately, yield at acceptable levels. Clay-rich soils on better-drained sites may be very productive, and very responsive to inputs of manure or fertilizer. In a dry year, the sandy soils are barren, the well-drained clay-rich soils are too hard to till, and only the clay-rich soils of the depressions yield anything at all. Hard-won experience of each subtly different village environment yields a huge variety of different “knowledge” and practices, and a soil CS project could only succeed if all this experience were tapped.

Using farmers’ knowledge and practices as an entry point for CS activities offers several advantages: (i) it stimulates farmers’ participation in research and project design from the outset; (ii) it facilitates the introduction of C, N and other minerals unknown to the majority of smallholders in a way that is readily understandable and easy to integrate in their own construct of soil theories; and (iii) it opens new doors for extension services to engage, together with farmers, in a more participatory and holistic approach to problem solving instead of delivering predefined agronomic packages.

Realizing the biophysical potential for carbon sequestration in farming systems

In dryland environments, SOC in the first 100 cm soil amounts to about 4 tonnes/ha (Batjes, 1999). This is considerably lower than in other environments. Batjes’ estimates for current SOC are: 7 - 10 tonnes/ha in the tropics; 7 - 13 tonnes/ha in the subtropics; 11 - 13 tonnes/ha in temperate regions; and 21 - 24 tonnes/ha in boreal, polar and alpine areas. Few reliable numbers exist for the entire Sahel, with the exception of estimates for semi-arid savannahs and dry forests in Senegal (the West-Central agricultural region) as reported by Tiessen and Feller, 1998, Ringius, 2002 and Tschakert, Khouma and Senè, 2004, ranging from 4.5 tonnes C/ha for continuously cultivated areas without manure input to 18 tonnes C/ha for non-degraded savannahs (top 20 cm soil).

There is a suite of recommended practices and land-use types that are recommended to increase both the uptake of C from the atmosphere and the duration of storage in soils. As for croplands, FAO (2001b) differentiates practices that decrease carbon losses from the soil from those that increase organic matter inputs into the soil, and considers a combination of both. The first category includes reduced/conservation/zero tillage, crop residue management, green manuring, cover crops, and integrated weed control. The second category is based on increases in biomass resulting from manure, compost, mulch farming, mineral fertilization and irrigation as well as improved crop-residue management and green manuring with leguminous species. All these practices simultaneously increase CS, improve soil fertility, and decrease erosion through soil restoration in drylands, thus offering real potential for a win - win situation for local smallholders.

However, reliable dryland estimates on how much C could be sequestered under the various management practices and farming patterns are still sparse. The most comprehensive estimates (Table 10) range from 0.05 to 0.3 tonnes C/ha/year for croplands and from 0.05 to 0.1 tonnes C/ha/year for grasslands and pastures (Lal et al., 1998). The estimates by Lal et al. for tropical areas are about twice as high as those for drylands. For the Old Peanut Basin in Senegal, Tschakert, Khouma and Senè (2004) report a possible range of 0.02 - 0.43 tonnes C/ha/year for improved crop-fallow systems.

TABLE 10
Effects from land management practices or land use on carbon sequestration potential in drylands

Source: After Lal et al. (1998)

Dryland farming systems with access to adequate water resources may benefit from developing their irrigation potential. In small-scale irrigation systems, a high potential for CS arises from four characteristics: from four characteristics:

• The supply of water allows high primary productivity, more so in some (qanat irrigation) than in others (runoff agriculture).

• In most of these systems, the soils are fine-textured, permitting high quantities and slow decay of soil C, much of which is bound closely to clay particles.

• The extensive use of manure, both from animal and crop residues.

• The generally carbon-neutral energy sources, such as animal and human power.

In qanats, the large investment of human power in the early years may be drawn on for centuries, or even millennia.

Larger-scale irrigation schemes should not be dismissed outright as systems for soil CS. Some date back to times when they were no inputs of fossil energy. Therefore, their construction might be regarded as carbon-neutral. Even more recent systems, as in the vast systems in Pakistan and India, were constructed largely with animal or human power. With their high productivity, use of many fine-textured soils and sometimes large extent, these may be potentially valuable sinks of soil C. However, the total carbon budget of each, from the time of construction through to recent interventions, such as deep drains or tubewells, would have to be evaluated separately before it could be considered as suitable as a soil carbon sink. Some would almost certainly not qualify. These would be those in which large amounts of fossil energy had been used in construction, or in which large quantities of fertilizer were being used.