Session 6
"The management of these agro-ecosystems”

The management of the agro-ecosystems associated with sandy soils

Bell, R.W. ¹ and V. Seng ²

Keywords: agro-ecosystems; erosion, leaching, nutrient cycles, water balance, sandy texture

Introduction

Sandy soils as defined in the World Reference Base (FAO-ISRAC-ISSS, 1998) contain <18% clay and >65% sand in the first metre of the solum. Generally, the main Reference Group for sandy soils is the Arenosol (FAO-ISRAC-ISSS, 1998). The nearest Soil Taxonomy equivalent is the Psamment sub-order (Van Wambeke, 1992). However, amongst tropical soils, sandy textures in the surface layers are more prevalent than sandy textures throughout the profile, and the shallow sandy soils may share similar attributes and constraints as deeper sandy soils for shallow rooted crops such as padi rice. Hence, in areas where lowland rice is prevalent it may be important to broaden the scope of sandy soils beyond the Arenosols to include those Reference Groups with members that are sandy in the surface layers: Regosols, Leptosols and Fluvisols.

Tropical sandy soils are prevalent in landscapes where felsic volcanic, or siliceous sedimentary rocks and their erosional products are found. They are also prevalent in desert regions, and as beach deposits and dunal features in coastal zones. The lower Mekong River Basin in Southeast Asia has an extensive area of sandy soils (Mekong River Commission, 2002). Other major provinces for sandy soils in tropical environments include: the Sahel zone of West Africa, the Kalahari basin (covering two-thirds of Botswana and Angola), and Northern Australia. However, smaller but still significant areas of sandy soils occur in most tropical regions including parts of central Vietnam, Pakistan, Saudi Arabia, Iran, and Brazil (FAO, 2005).

In terms of the agro-ecosystems developed on sandy soils, the prime limiting factors and the main concerns for sustainability, vary according to topography, viz, steepland (>12% slope), uplands and lowlands; sub-divided further by agro-ecological zoning. Sandy soils occur in arid, semi-arid and humid rainfall zones of the tropics and from coastal lowlands to high altitudes. For the purposes of this paper, we will exclude sandy soils in desert regions due to their low potential for land use except where irrigation water is available.

Sandy soils in steeplands tend to be used for conservation reserves, forestry (including plantations) and for shifting cultivation, and may also serve as important water catchments. Sandy uplands comprise all those soils that are neither seasonally waterlogged nor steep. They are used for a range of field cropping systems depending on the agro-ecological zoning and elevation. Plantation crops and forestry are also prevalent. Sandy soils in the continuously or seasonally waterlogged lowlands are largely developed for irrigated and rainfed rice cultivation especially in Asia. While sandy soils are not highly suitable for rice cultivation, because rice is the major subsistence crop in Southeast Asia, sandy soils commonly are used for this purpose. Rice is also common in West Africa. Because of the shallow rooting zone of padi rice which is restricted to 0-20 cm, the surface texture has a more dominating influence on lowland rice than in upland cropping. Hence it is common to refer to soils in lowlands as sandy if they have sandy surface textures regardless of the subsoil texture. Indeed, White et al. (1997) in their Cambodian Agronomic Soil Classi­fication system (CASC), developed for rice soils, restricted their consideration of soil properties to the 0-50 cm depth. The continuously or seasonally waterlogged lowlands are also important fisheries habitats, for aquaculture and for flood control.

The environments in which tropical sandy soils occur are prone to degradation risks from nutrient decline, erosion, leaching, salinity, and acidification. Management of agro-ecosystems for sustainable production on these landscapes will need to find technologies to overcome these constraints for economic viability of the enterprises. Development of sustainable agro-ecosystems in these sandy terrains should be based on optimisation of key ecosystems processes: closing nutrient cycles, restoring hydrological balance; enhancing biodiversity and strengthening resilience of these processes to perturbations. A range of opportunities that exist to achieve sustainability of agro-ecosystems in sandy landscapes will be outlined.

Development of sustainable agro-ecosystems in sandy terrains

Sustainable agro-ecosystems need to simul­taneously satisfy three sets of criteria: economic viability; ecological processes; and social acceptability (Lefroy et al., 1992). This represents an advance on earlier hinking that focussed on economic viability of agro-ecosystems. It also recognises that focussing only on ecological processes will not be sufficient for sustainable agro-ecosystems since economic viability continues to drive many key decisions by farmers, and society increasingly is expressing a voice about the practice of agriculture in terms of the quality of food it delivers to markets and the off-site impacts of agriculture.

Economic viability of agriculture

Tropical sandy soils have a wide range of limiting factors for agricultural use: these include nutrient deficiencies, acidity, low water storage and poor physical attributes. Limiting factors may have a major bearing on the economic viability of agriculture on sandy soils. For example, a large province of sandy soils in Western Australia was deficient in the micronutrients, Zn, Cu and Mo (Bell et al., 2004). Until the discovery of these deficiencies and practical means of correcting them, it was not economically viable to use these sandy soils for agriculture.

Low nutrient levels are common on sandy soils, and crops grown on these soils commonly express multiple nutrient disorders which limit productivity of crops (e.g. Northeast Thailand, Bell et al., 1990). While fertiliser can correct these disorders, it is often difficult to achieve the optimal mix of nutrients and other soil amendments to make it economic (e.g. Ragland and Boonpukdee, 1987). Failure to diagnose all the limiting nutrients in a soil will lead to ineffective use of fertilizer and poor responses to those fertilizers that were applied. The widespread use of N alone often provides poor returns from fertilizer investment since on sandy soils deficiencies of P, S, K and/or micronutrients also commonly limit crop production.

In part the difficulty of fertilizing crops when there are multiple deficiencies is lack of availability of appropriate fertilizer products. In Thailand, for example, there are a large range of fertilizer products available, but limited understanding by farmers of the types most suited for particular soils and crops (Bell et al., 1990). A range of NPK formulations are commonly available in the Thai market but they vary in S content (Chunyanuwat et al., 1993). Since S deficiency is quite common in Northeast Thailand, NPK formulations with different S content may give quite different responses in crops (Bell et al., 1990). Similarly, NPK formulations in Thailand vary in B content which would affect responses since large areas of sandy soils in Northeast Thailand are low in B (Bell et al., 1990). Market supply of micronutrient fertilizers varies nationally and locally and in many places such fertilizers are not readily available, e.g. Bangladesh (C. Johansen, pers. comm.). Finally, according to Ragland and Boonpukdee (1987) responses to fertilizers alone on sandy soils in Northeast Thailand are poor without addition of organic matter.

The high percolation rates of the deep sands are a major limiting factor for rainfed agriculture. Drought is the most important consequence for crops. However, leaching of N and other nutrients may also limit productivity of these soils even when water is not limiting. For example, the deep sandy Prey Khmer soil in Cambodia has lower potential productivity for rice even with fertilizer application than the other major sandy lowland soil (Prateah Lang), which has a higher clay content in the subsoil (White et al., 1997).

Acidity is common on sandy soils. Where lowland rice is the main crop, flooding alleviates acidity (Kirk, 2004). However, for rainfed crops, acidity may give rise to a range of fertility constraints including Al and Mn toxicities and deficiencies of nutrients (Dierolf et al., 2001). Poor N fixation in legumes is often a consequence of soil acidity due to the low tolerance of Rhizobia to acidity, and to low lvels of plant available Mo.

Salinity is a common constraint on sandy soils wherever irrigation is used in semi-arid environments. In sandy coastal zones, salinity is associated with seawater intrusion (White et al., 1997). Less well recognised is dryland salinity, which arises from perturbation of the hydrological balance in rainfed environments. It is a major problem in Northeast Thailand where a large percent of soils are sandy: currently dryland salinity affects about 12% of the lowland soils but is predicted to spread to cover up to 30% (Yuvaniyama, 2001).

Optimisation of key ecosystems processes

In undisturbed ecosystems, key processes such as the cycling of nutrients, hydrological balance, energy capture and flow, biodiversity, and resilience maintain ecosystem function. Disturbance of eco-systems for agriculture alters each of these processes. It has been argued that sustainable agro-ecosystems could be developed by mimicking the operation of the key ecosystem processes (Lefroy and Stirzaker, 1999). Hence a way forward for sustainable agriculture in sandy terrains is to understand key ecosystem processes that operated in pre-existing ecosystems and to model agriculture on those processes. However, there are practical limits to the application of this approach. Firstly, there are as yet inadequate studies of ecosystem function in tropical sandy terrains. Secondly, where harvested products are exported from the location from which they were produced, the nutrient cycle is interrupted and nutrient supply needs to be maintained through inorganic and/or organic inputs.

Closing nutrient cycles

In undisturbed ecosystems, the cycling of nutrients through the biomass and soil compartments ensures that leakage of the store of nutrients is negligible (Grierson and Adams, 1999). Small amounts leave the ecosystem but may be offset by accretion in rainfall, by nitrogen fixation etc. However, most agricultural systems allow significant nutrient losses through harvested product removal, leaching, gaseous losses, and erosion. While these losses may be offset by fertilizer use or the return of crop residues to the field, the losses of nutrients still represent inefficiency in nutrient use and may have off-site consequences. Sandy soils in Northeast Thailand, for example, have lost considerable nutrient capital since the clearing of dipterocarp forests (Noble et al., 2000) and changed nutrient cycles and hydrology have made these soils prone to acidification. Nutrient budgets calculated for farm land in Northeast Thailand show large net losses at the field scale and at regional scale (Lefroy and Konboon, 1998).

Restoring hydrological balance

Water balance considers the partitioning of rainfall to evaporation, transpiration, runoff, deep drainage and to the change in soil water storage. In sandy terrain, the runoff component may be low especially in the pre-existing ecosystems. Change of land use to agriculture would normally alter hydrological balance by changing runoff, transpiration and deep drainage components (Lefroy and Stirzaker, 1999). Northeast Thailand is a relevant case study demonstrating the consequences of hydrological change in a sandy terrain. The development of dryland salinity appears to be related to a change in the landscape water balance following clearing of the forest for agriculture (Williamson et al., 1989). Rapid clearing in Northeast Thailand occurred in the 1960’s (Ruaysoongnern and Suphanchaimart, 2001). Prior to that, the salt stored in the halite strata of the near-surface sedimentary formation was not mobilised because the vegetation used most of the rainfall allowing little deep drainage to groundwater. However, under rice-based and upland cropping, significant deep drainage to the groundwater occurs annually and this has caused watertables to rise egionally over time. When groundwater reaches the soil surface or within 2 m of the surface, discharge of salt occurs annually at the soil surface. The gentle relief of the Northeast Thailand and the widespread shallow halite-bearing sediments place large areas of Northeast Thailand at risk of dryland salinity.

As the development of salinity in sandy soils of Northeast Thailand is essentially a water balance problem (Williamson et al., 1989), its long term solution will come from changes in land use that decrease deep drainage to regional groundwater. Given the current prevalence of lowland rice cultivation, this will prove a challenge in the short term. Tree planting or revegetation with perennial vegetation across a significant portion of the landscape may be needed to restore water balance, but the minimum amount needed to be effective is not known. Currently upland areas are mostly targeted for tree planting. More extensive agroforestry planting in lowlands may also be needed to help restore the water balance.

However, wherever large areas are seasonally flooded, deep drainage will continue. Lowland rice is uniquely dependent on surface hydrology and the duration of standing water in relation to crop growth stages (Fukai et al., 2000). Marked alteration of the hydrology of sandy terrain inevitably occurs with its cultivation. High deep drainage rates are a common problem in the sandy lowland rice soils of Cambodia (White et al., 1997), Laos and Northeast Thailand (Fukai et al., 1995). Deep drainage rates varied from 1 to 6 mm d^-1 on sandy soils (Fukai et al., 1995). Model simulations for Ubon in Northeast Thailand show about a 50% increase in rice yield if the deep drainage rate of a sandy soil (6.3 mm d^-1) under puddled conditions could be reduced to 1.4-1.8 mm d^-1 (Fukai et al., 2000).

Fields in the high or upper terraces of the sandy lowlands lose large amounts of water, particularly after heavy rainfall, through surface runoff and subsurface lateral water movement, while those in the lower terraces may intercept the flows from the upper paddies (Fukai et al., 2000). Moreover, location of on-farm drains, and road embankments and drains under roads can markedly affect where the runoff is directed. Lateral redistribution of water results in water availability and rice growth duration varying by 30 days or more, within quite small areas. Fukai and colleagues have used simulation models to estimate the sensitivity of rice yield to the effect of variation in one parameter while all others are held constant. In sandy terrain at Ubon in Northeast Thailand, the influence of run-on to the lower terrace diminished as the deep percolation rate was reduced from 6 to 1 mm d^-1. With 1 mm d^-1, there was almost no water stress throughout the growth period and hence the effect of water movement was small, whereas, with 4-6 mm d^-1, rice experienced periods with standing water interspersed with periods of water stress. In this case, simulated grain yield was strongly influenced by variation in lateral water movement.

Enhancing biodiversity

Agricultural ecosystems generally involve a decrease in biodiversity relative to the prior native ecosystems that existed. This is particularly the case where monocultures of cereals or plantation crops dominate the landscape. Biodiversity in these monoculture-dominated agro-ecosystems is usually greatest around villages and home gardens, and in the remnants of the prior ecosystems especially in less favoured agricultural environments such as riparian zones along rivers, streams and wetlands. Enhancing biodiversity within the agricultural system can be achieved through agroforestry, intercropping and crop diversification. In situ conservation of wild relatives of crop species has an important role in conserving gees that may be useful in future breeding programmes (Rerkasem, 2004). Flora in greatest need of conservation is in the sandy uplands due to their widespread use for agriculture. Conservation in the steeplands is often compromised by short rotation shifting cultivation, but there are cases in Southeast Asia of the compatible use of steeplands for agriculture and the conservation of biodiversity (Rerkasem, 2004).

Opportunities to improve sustainability of sandy landscapes

Sandy soils represent an ongoing challenge for water and nutrient management at landscape and field scales. Productivity on these soils tends to be low, even when recommended agronomic practices are followed. However, there are a number of promising avenues for sustainable development of agro-ecosystems dominated by sandy soils.

Maintenance of soil organic matter

Sandy soils generally have lower organic matter levels than heavier textured soils given similar rainfall, temperature, land use and tillage. Clay levels on deep sandy soils may be too low to protect organic matter from oxidation (Baldock and Nelson, 2000). Reducing tillage may help maintain organic matter levels. Perennial crops will also tend to maintain higher organic matter levels. However, there are practical limits to the levels of organic matter that are achievable on sandy soils (Shirato et al., 2005). Research with the aim of boosting organic matter levels in sandy soils often fails to recognise this limit and as a result there has been much wasted investment on organic matter management. Organic matter levels can be enhanced in sandy soils through minimum tillage, zero burning and retaining crop residues. Slowly decomposing litter appears to build organic matter levels over a period of several years on sandy soil in Ubon, Northeast Thailand (Naklang et al., 1999).

Clay and other mineral soil amendments

Possibly the most effective long term investment in improving productivity and sustainable use of sandy soils would be to increase their clay content. Application of clay to sandy soils has been suggested as a semi-permanent treatment to enhance water and nutrient retention in Northeast Thailand (Noble et al., 2004). Enhanced clay content would allow soils to accumulate increased organic matter levels, and hence retain more water and nutrients and buffer soils against significant change in chemical properties. However, this strategy is limited by the availability of clays, the cost of transporting the required amounts of clay and by a still rudimentary knowledge about potential benefits. Initial research on the sandy soils of Northeast Thailand suggests very strong responses in growth can be achieved by clay amelioration. Further work is ongoing to demonstrate the benefits of this technology. Northeast Thailand has numerous deposits of high activity clay in lacustrine sediments, S. Ruaysoongnern (pers. Comm.). The relevance of this technology for other parts of the region, particularly for the Prey Khmer (Arenosols) and Prateah Lang (Acrisols) soils of Cambodia (White et al., 1997; Bell and Seng, 2004) warrants further investigation.

Balanced fertilization

Sandy soils commonly suffer from multiple nutrient deficiencies. These problems can be compounded by Al toxicity. For this reason, the concentration of research on N or NP fertilizer use that has yielded significant benefits for farmers on loam and clay soils often gives disappointing results on sandy soils. Our research on farmers’ fertilizer use on sandy soils in Southern Cambodia suggest that they over-used N and under-used P, K and hardly used S (Ieng et al., 2002). Even a focus on NPK fertilizer use may lead to ineffective fertilizer programmes since S and micronutrient deficiencies such as B, Zn, Cu and Mo are common. Bell et al. (1990) reported extenive areas of multiple deficiencies on sandy soils of Northeast Thailand yet as discussed above S contents of fertilizers vary and relatively little use of B fertilizer has occurred despite there being solid evidence of crop yield responses to this element. The challenge for achieving balanced crop nutrition in sandy terrain is to provide simple advice and tailored fertiliser advice and products so that it can be efficiently delivered to large numbers of small farmers. Nutrient budgets may be a useful strategy for guiding farmers to consider the important issues for balanced nutrition and sustainable crop production. However, it needs to be supported by a parallel business development approach that seeks to supply fertilizers of the right type in the markets where they are needed, at an affordable price.

Plantation crops, forestry and agroforestry

Where the pre-existing ecosystems were forests or woodlands, a sustainable agro-ecosystem probably needs to incorporate a similar vegetation structure, as plantation crops, plantation forestry, or agroforestry. The soil cover provided by perennial vegetation minimises erosion and nutrient leaching, while the decrease in cultivation intensity and frequency helps maintain soil organic matter levels and soil structure. Ecosystems benefit from restoration of water balance, closed nutrient cycling, and enhanced biodiversity. Agroforestry may provide similar benefits to plantations (Young, 1997).

Breeding species for adaption to the constraints

Where a soil constraint is widespread and difficult to overcome by conventional agronomy, there is a strong case for breeding for tolerance to the stress. In sandy soils, this would commonly mean breeding for drought tolerance first, followed by tolerance of mineral disorders, including soil acidity.

Management of agro-ecosystems associated with sandy soils

Ecosystems provide a range of services that tend to be undervalued, and those existing on sandy terrains are no exception. Changed water balance which is the consequence of land use for agriculture in most cases, gives rise to land degradation that has on-site and off-site effects. Sandy soils in river basins deliver water that is used for a variety of purposes downstream. Changed water flow patterns and changed water quality may harm downstream wetlands and riparian zones on which the livelihoods of communities depend. Changed water flow patterns may give rise to downstream flooding events that impact on major settlements and cities. Excess nutrients or sediment from sandy terrains may cause harm to fisheries and fisheries breeding places. Hence there is a need to develop institutional arrangements that pursue integrated river basin management outcomes that recognise the connection between upstream and downstream users and achieve equity between their respective costs and benefits (Kam et al., 2001). Such institutional arrangements need to incorporate the range of stakeholders that have an interest in the structure and function of river basin management, including government natural resources management agencies representing both production and conservation, private sector business, educational institutions and civil society groups.

References

Baldock, J.A.; Nelson, P.N. 2000. Soil organic matter. In: Handbook of Soil Science. Ed. M.E. Sumner. pp. B25-84. CRC Press, Boca Raton, Florida.

Bell, R.W.; Seng, V. 2004. Rainfed lowland rice-growing soils of Cambodia, Laos, and Northeast Thailand. In: Water in Agriculture. Eds V. Seng, E. Craswell, S. Fukai and K Fischer. ACIAR Proceedings 116, pp. 161-173.

Bell, R.W.; Dell, B.; Huang, L. 2004. Importance of micronutrients in crop production. In: International Fertilizer Association Symposium on Micronutrients, Dehli, 23-25 Feb 2004, published at http://ww. fertilizer.org/ifa/news/2004_3.asp

Bell, R.W., Rerkasem, B., Keerati-Kasikorn, P., Phetchawee, S., Hiranburana, N., Ratanarat, S., Pongsakul, P.; Loneragan, J.F. 1990. Mineral Nutrition of Food Legumes in Thailand with particular reference to micronutrients. ACIAR Technical Report. 19, pp. 52.

Chunyanuwat, P.; Supavita, P.; Bell, R.W.; Loneragan, J.F.; Lefroy, R.; Blair, G. 1993. Secondary and micronutrients in chemical fertilizer. In: Fertilizer Analysis Methods. pp. 41-45. Eds J. Jermsiri and P. Chunyanuwat. Department of Agriculture, Bangkok.

Dierolf T.; Fairhurst T.; Mutert E. 2001. Soil Fertility Kit. GTZ-GmbH, FAO, PT Jasa Katom, and PPI and PPIC. Oxford Graphic Printer.

FAO (2005). fao.org/landandwater/agll/prosoil/accessed.

FAO-ISRIC-ISSS 1998. World Reference Base for Soil Resources. Acco Press, Leuven, Belgium.

Fukai, S.; Basnayake, J.; Cooper, M. 2000. Modelling water availability, crop growth, and yield of rainfed lowland rice genotypes in Northeast Thailand. In: Characterising and Understanding Rainfed Environments. Eds T.P. Tuong, S.P. Kam, L. Wade, S. Pandey, B.A.M. Bouman and B. Hardy. pp. 111-130. IRRI, Los Baños, Philippines.

Fukai, S.; Rajatsasereekul, S.; Boonjung, H.; Skulkhu, E. 1995. Simulation modelling to quantify the effect of drought for rainfed lowland rice in Northeast Thailand. In: Fragile Lives in Fragile Ecosystems, Proceedings of the International Rice Research Conference, 13-17 Feb 1995. pp. 657-674. IRRI, Los Baños, Philippines,

Grierson, P.F.; Adams, M.A. 1999. Nutrient cycling and growth in forest ecosystems of Southwestern Australia. Relevance to agricultural landscapes. Agroforestry Systems 45: 215-244.

Ieng, S.; Bell, R.W.; Ros, C.; Cox, P.G.; Pracilio, G.; Cook, S.; Mak, S. 2002. Farmers’ perceptions of rice response to environment and fertilizers in rainfed rice ecosystems in Takeo Province, Cambodia. Cambodian Journal of Agriculture 5: 36-47.

Kam, S.P.; Hoanh, C.T.; Trebuil, G.; Hardy, B. (Eds.) 2001. Natural Resource Management Issues in the Korat Basin of Northeast Thailand: An Overview. Limited Proceedings No. 7. IIRI, Manila, The Philippines.

Kirk, G.D.J. 2004. Biogeochemistry of Submerged Soils. John Wiley and Sons, Chichester, UK. pp. 291.

Lefroy, R.D.B.; Konboon, Y. 1998. Studying nutrient flows to assess sustainability and identify areas of nutrient depletion and imbalance: an example for rainfed rice systems in Northeast Thailand. In: Rainfed Lowland Rice: Advances in Nutrient Management Research. Eds J.K. Ladha, L.J. Wade, A. Dobermann, W. Reichardt, G.J.D. Kirk and C. Piggin. pp. 77-93. IRRI, Los Baños, Philippines,

Lefroy, E.C.; Stirzaker, R.J. 1999. Agroforestry for water management in the cropping zone of Southern Australia. Agroforestry Systems 45: 277-302.

Lefroy, E.; Salerian, J.; Hobbs, R.J. 1992. Integrating economic ad ecological considerations: A theoretical framework. In: Re-Integrating Fragmented Landscapes: Towards Sustainable Agriculture and Nature Conservation. Eds D.A. Saunders and R.J. Hobbs. pp. 209-244. Springer Verlag, New York.

Mekong River Commission 2002. Land Resource Inventory for Agricultural Development (Basin wide) Project. Part III Soil Database Final Report June 2002. Mekong River Commission, Phnom Penh.

Naklang, K.; Whitbread, A.; Lefroy, R.; Blair, G.; Wonprasaid, S.; Konboon, Y. ; Suriya-arunroj, D. 1999. The management of rice straw, fertilisers and leaf litters in rice cropping systems in Northeast Thailand. Plant and Soil 209: 21-28.

Noble, A.D.; Gillman, G.P.; Ruaysoongnern, S. 2000. A cation exchange index for assessing degradation of acid soil by further acidification under permanent agriculture in the tropics. European Journal of oil Science 51: 233-243.

Noble, A.D.; Ruaysoongnern, S.; Penning de Vries, F.W.T.; Webb, M. 2004. Enhancing the agronomic productivity of degraded soils in Northeast Thailand through clay-based interventions. In Water in Agriculture. Eds V. Seng, E. Craswell and K. Fischer. ACIAR Proceedings 116, 147-160.

Ragland, J.; Boonpuckdee, L. 1987. Fertiliser responses in Northeast Thailand. 1. Literature review and rationale. Thai Journal of Soils and Fertilisers 9: 65-79.

Rerkasem, B. 2004. Land and water resources. In Water in Agriculture. Eds V. Seng, E. Craswell and K. Fischer. ACIAR Proceedings 116, 105-118.

Ruaysoongnern, S.; Suphanchaimart, N. 2001. Land-use patterns and agricultural production systems with emphasis on changes driven by economic forces and market integration. In: Natural Resource Management Issues in the Korat Basin of Northeast

Thailand: An Overview. Eds T.P. Tuong, S.P. Kam, L. Wade, S. Pandey, B.A.M. Bouman and B. Hardy. pp. 67-78. IRRI, Los Baños, Philippines.

Shirato, Y. ; Paisancharoen, K.; Sangtong, P.; Nakviro, C.; Yokozawa, M; Matsumoto, N. 2005. Testing the Rothamsted carbon model against data from long term experiments on upland soils in Thailand. European Journal of Soil Science 56: 179-188.

Van Wambeke, A. 1992. Chapter 13 Entisols. Soils of the Tropics. Properties and Appraisal. pp. 253-266. McGraw-Hill, New York.

White, P.F.; Oberthür, T.; Pheav, S. 1997. The Soils Used for Rice Production in Cambodia, A Manual for their Recognition and Management. International Rice Research Institute, Manila, Philippines. 71 p.

Williamson, D.R.; Peck, A.J.; Turner, J.V. 1989. Ground-water hydrology and salinity in a valley in Northeast Thailand. In: Goundwater Contamination. IAHS Publ. No. 185, 147-154.

Young, A. 1997. Agroforestry for Soil Management. 2^nd Ed. CAB International, New York. 320 p.

Yuvaniyama, A. 2001. Managing problem soils in Northeast Thailand. In: Natural Resource Management Issues in the Korat Basin of Northeast Thailand: An Overview. Eds T.P. Tuong, S.P. Kam, L. Wade, S. Pandey, B.A.M. Bouman and B. Hardy. pp. 147-156. IRRI, Los Baños, Philippines.

¹ School of Environmental Science, Murdoch University, Murdoch, Western Australia 6150

² Office of Soil and Water Sciences, Cambodian Agricultural Research and Development Institute, P.O. Box 01, Phnom Penh, Cambodia

Long-term topsoil changes under pearl millet production in the Sahel

Anneke de Rouw¹ 

Keywords: Sandy soils, pearl millet, fallows, organic matter, manures

Introduction

In the African Sahel, between the 400 and 700 isohyets, people subsist on pearl millet, the only profitable crop. Agriculture and husbandry are often linked in this semi-arid region. However, both crops and livestock productions suffer from a combination of low soil fertility and scarce and unpredictable rains. Most soils in the Sahel are derived from acidic or aeolian parent materials which are poor in clay and nutrients, especially N and P (Bationo & Mokwunye, 1991). Almost all nutrients are found in the soil organic matter, a fraction that is also very low. The decline in nutrient status during cultivation is an inevitable consequence of clearing and is reinforced by the effects of cultivation (Ahn, 1970 p. 244; Feller & Beare, 1997). The ongoing physical land degradation in the Sahelian zone of West Africa, sealing, crusting, hardsetting of soils and the long-term reduction in amount of diversity of the natural vegetation, is enhanced by population growth and a marked drier climate since the 1970s (Sivakumar, 1992; Valentin, 1995).

Cultivation clears the soil and leaves the soil surface almost bare. Due to the heavy rains, the soil undergoes serious structural deterioration. The impacts of the drops separate the fine soil particles and the organic matter from the sand, and the soil pores get clogged. This makes the sandy soils of the Sahel very liable to surface crusting (Ambouta et al., 1996). Soil surface crusting reduces the infiltration rate, and thus triggers runoff and erosion. It constitutes a serious constraint for cultivation in dry areas (Casenave & Valentin, 1992). However, wind erosion is probably a greater constraint. Land clearing reduces the soil protection provided by the standing vegetation and the litter (Bielders et al., 2000). Moreover, any soil tillage breaks the surface crusts, and the detached material can be easily carried away by the wind during the storms (Valentin, 1995). The protection of the soil surface is improved by the presence of stable aggregates at the soil surface, which obviously resist better to the impact of the raindrops (Feller et al., 1989). It is also improved by any object, crop residue, dung or vegetation fragment lying on the soil, as they absorb the energy of the rains (Collinet & Valentin, 1985) and trap wind-blown particles during the storms (De Rouw & Rajot, 2004b).

Dust trapping is particularly important in the Sahel. In contrast with the local soils, which tend to be acid and ighly weathered, the dust from the Sahara exhibits appreciable quantities of water-soluble and exchangeable cations (Hermann, 1996). About 25% of the dust-load consists of clay, silt and organic matter, the rest being very fine sand (Möberg et al., 1991). Though dust constitutes an important long-term factor of nutrient renewal for these soils, and though dust is supposed to help maintain soil fertility, the actual quantities involved are difficult to assess. Estimations of annual dust input in Southwest Niger varies widely, depending on method and scale: 0.9 t ha^-1 (Buerkert & Hiernaux, 1998), 6.8 t ha^-1 (Chappell et al., 1998), and 1.9 t ha^-1 (Herrmann et al., 1994). When open buckets are used to measure dust deposition, only the airborne material that falls down vertically like rain is trapped. However, dust accumulates on vertical obstacles, thus tree- and shrub-covered fallows that constitute a considerable dust-trap. An appropriate method to determine the complementary dust input due to tree and shrub trapping would be to wash the dust from the vegetation and to estimate the leaf area. Though the sedimentation of dust is evident and the chief processes of soil restoration and soil losses is well documented, there is a need to investigate how pearl millet cultivation respond to this.

Most farmers in the Sahel are too poor to use external inputs (Powell et al., 1996). Subsequently, the long-term success of pearl millet cultivation depends on the recycling of nutrients in the topsoil. In practice this means either by manuring or fallowing. Model-based farm studies integrating livestock, pasture and cropland components suggest that the continuous cultivation of soils in the Sahel can be achieved by manure application, even if low quantities are applied (Harris, 1999; Abdoulaye & Lowenberg-de Boer, 2000; Buerkert & Hiernaux, 1998; Bielders et al., 2002). Minimum dung inputs observed in farmers’ fields are 1.3 t ha^-1 every two or three years (Powell & Williams, 1993) and 1.1 t ha^-1 year^-1 (De Rouw & Rajot, 2004a). No such farm studies are available for fallowing. This study focuses on the changes in top soil due to long-term cultivation of soils and aims to relate well-known processes of soil restoration and degradation to farming practices.

Materials and Methods

Study area

The study area is located 60km east of Niamey, Niger, near the village of Banizoumbou (13°31'N, 2°39'E). The village territory, approximately 80 km², was exploited by 84 farms with an average number of 10 persons per farm. Part of the best land had been cultivated for about 150 years in alternation with periods of bush fallow. By 1990 this type of cultivation has spread over 70% of the village territory, the remaining 30% being marginal land or unfit for cultivation yet suitable for pasture (Loireau, 1998). In 1978, cultivation with regular manure application started and this practice was used at the time of the study over about 10-15% of the annually cropped area. Only 15% of the farms had more than 10 domestic animals (zebus, goats and sheep), the others had less or no livestock. However, exchange contracts were often made between farmers and nomadic herders about the manuring of fields. In the dry season, animals are left free at night in the field to let the dung and urine fall on the soil surface and decompose in situ. In the day-time these herds move to permanent pasture or fallow land. This labour-extensive practice is widespread in the Sahel (Landais & Lhoste, 1993; De Rouw & Rajot, 2004a).

The climate is hot and dry most of the year. The mean annual rainfall is 550 mm, the rains falling between June and September. The four years of experiments (1993-1996) were near average as far as total rainfall was concerned (total rainfall: 461, 642, 509 and 523 mm) but the istribution of the rain events varied largely between years and sites.

Crop production was entirely rainfed. Pearl millet was cultivated on every field, the fields being usually very large (minimum and maximum area of individual fields 4 and 30 ha, respectively). The crop was sown in hills and generally two weeding rounds were required. Weeding was performed using the “hilaire”, a shallow cultivating hoe that not only cuts the roots of the weeds but, more importantly, breaks the superficial crusts to allow the rains to infiltrate. All tillage operations (cultivation, clearing, sowing, thinning, weeding) were entirely made by hand and no chemical fertilizers were applied. Thick aeolian sand deposits, up to 9 m thick, were the preferred areas for cultivation, provided the slopes were less than 4%. These deposits were uniformly sandy, from 91% sand, 6% silt and 3% clay in the topsoil upper slope to 90% sand, 6% silt and 4% clay downslope. The organic matter content in the topsoil was very low, but slightly higher downslope (0.27%) than upslope (0.23%) (D’Herbès & Valentin, 1997). These cultivated soils are classified as Psammentic or Cambic Arenosols (FAO).

Sites and cropping systems

A survey among 60 farms gave four cropping systems depending on the farmers’ access to arable land and manure.

Fallow system — long cycles. Farmers with land but no access to manure cultivate the same field for about 10 years then let it lie fallow for more than 15 years (four fields).

Manure system — new fields. Farmers with land and access to manure open up new fields and apply small quantities of manure each year thus cultivating the same site for over 15 years (four fields).

Fallow system — short cycles. Farmers with little land and no access to manure cultivate the same plot for short periods, 4-6 years, after which a fallow period is necessary of 3-5 years (four fields).

Manure system — old fields. Farmers with little land but access to manure recuperate impoverished land that they make productive by annual manure application (three fields).

It should be kept in mind that all fields are under the pressure of population increase, resulting in two long-term trends: (1) The gradual shortening of the fallow period, i.e. fields of the Fallow system - long cycle will progressively pass into the group Fallow system - short cycle. (2) Manure replacing the fallow, i.e. some fields of the Fallow system will evolve into field of the Manure system.

Plot size and other observations

Pearl millet stands can be extremely variable over short distances. Part of this variability is organized along the slope (Rockström & de Rouw, 1997). In order to capture this variability, the plots were arranged in transects of 100 m*5 m running down the slope. However, some variability appeared at random in the fields. This was mainly due to contrasting soil surface features ranging from highly permeable micro dunes to almost impermeable crusts. The size of the plots was adjusted to the scale of these heterogeneities (5*5 m).

Some transects were studied for four years (3 fields), others for two years (9 fields), and some for only one year (3 fields). The general description of each field (n = 15) included the area, the slope and the history of the site, with an estimate of the family’s access to manure, workforce and land. Annual records for each field (n = 33) included the daily rainfall and the cropping practices. Annual records of each plot (n = 1,320) included: (1) in the first week after sowing: cover by dung and crusts (typology after Valentin & Bresson, 1992), maximum height and origin of micro relief (between 5 and 50 cm), and the meso relief (over 50 cm, mainly gullies); (2 at harvest: number of woody plants, grain yield and total aboveground biomass of the crop. Fresh biomass was weight in the field and a sample was taken and ovendried (24H 70ºC) for dry weight determination.

Soil sampling and analysis

Topsoil (0-20 cm) samples were taken at harvest. As the analysis of every plot each year was too costly, two compromises were made: (1) as transects ran down the slope and the soil texture was known to slightly increase downslope, samples of two adjacent plots were mixed, up to a maximum of four plots in case they looked homogeneous; (2) When the soil conditions were different in adjacent plots (e.g. a gully or sand fan) the soil samples were analysed separately. As a result, the number of soil samples analysed per field varied between a minimum of 10 in transects with a relatively uniform soil surface, to a maximum of 14 in transects with heterogeneous soil surface. In order to get a representative set of data, both the mean and the standard deviation of each variable were weighed by the number of plots mixed in the soil sample.

Soil analyses included pH-water, total carbon, particle size distribution (<2 µm, 2-20 µm >20 µm), N-tot, P-Bray and P-tot. The cations Ca²⁺, Mg²⁺, Na⁺ and K⁺ were determined using the ammonium acetate method, H⁺ and Al³⁺ were determined using 1 M KCl. The effective cation exchange capacity (ECEC) was calculated as the sum of exchangeable bases and exchangeable acidity. Zebu dung was collected from the soil surface in May 1994, where it had dried in situ. Preliminary analysis of the data.

Data gathered over three or four years from the same transect were analysed to determine whether they showed a trend with time. None of the soil variables demonstrated such a trend. For example, the organic matter content did not decline as the number of years of cultivation increased from the third to the sixth year. Instead, all the data from a given field tended to remain clustered. The variables “height of micro relief”, “proportion of permeable surface”, “number of woody plants ha^-1”, “grain yield” and “crop residue” were also unrelated to the year of cultivation. Hence, a single data set per field was enough. In order to facilitate the comparison among fields of the fallow system, the data analysed were those of the third year of cultivation after clearing. For the manured fields, the data were those of the last available year.

Results

Analysis of topsoil

In farming systems with no chemical fertilizers applied, organic matter and clay are determinant for the cation exchange capacity of the soil and therefore the key elements to appreciate the nutrient status of the soil. Organic matter and clay + silt contents were very low in all samples (Figure 1). These values are typical of the Sahelian sandy soils cultivated with pearl millet. In fields where fertility maintenance relies exclusively on fallowing, the topsoils of long fallow fields were different from those of short fallow fields by their higher clay + silt content, and to a lesser degree by their slightly higher organic matter content. Fields where manure was applied were split in two groups, the “old” fields being sandier than the “new” fields. Soil organic matter was not different in the two groups, probably because the quantities of manure applied yearly were equally low across all the fields.

Figure 1. Organic matter content and soil particles <20 µm content of topsoil (0-20 cm) of pearl millet fields, fertility management by long or short-term fallowing, with or without manure input, Banizoumbou, Niger

The well-known relationship between organic matter and clay + silt contents is usually interpreted as an indication of land degradation, because uncultivated soils exhibit the highest values, over cultivated soils the lowest and normally managed soils intermediate values. The difference with the literature is that the values observed in the study area were much lower. Data from fields after long and short fallow periods were located on a continuum because transitions from long to short fallow cycles occur. The reduction of fallow periods and the accumulation of rotations resulted in a gradual decline in soil organic matter from 3.7 g kg^-1 under long fallow management to 2.3 gkg ^-1 under short fallow, while the clay + silt content decreased from 107 g kg^-1 under long fallows to 57 gkg ^-1 soils under short fallows. Thus, the solid line in Figure 1 describes the loss in soil particles <20 µm and in soil organic matter as cultivation becomes more frequent.

In fields where manure is applied, the most frequently cultivated sites were also the poorest in soil particles <20 µm, but the organic matter content was not different. Thus manure application seems to make up for some of the organic matter reduction enforced by cultivation, but manure application cannot prevent the topsoil from becoming sandier with frequent cultivation.

Table 1 shows the chemical analyses of the soil samples and zebu dung, the principal input. These results confirm the overall very low fertility of millet fields. Rotation of short fallow and cultivation without manure application resulted in the poorest topsoils in N, P and exchangeable cations.

Micro relief and soil surface

In the absence of tillage other than the breaking of the surface crusts, the micro relief is always natural. The micro relief can be microdunes (favouring infiltration) or crusted micro heights and lows (generating runoff). Both kinds of micro relief were found in every millet field, but the repartition between infiltrating and impermeable micro relief was strongly correlated with cropping practices (Table 2).

In manured fields, the micro relief was mainly made of microdunes formed around any obstacle littering the soil surface. Their maximum height (up to 8 cm) depended on the degree these obstacles could trap aeolian sand. In fields of the fallow system, the micro relief was mostly related to trees and shrubs. The micro mounts were highest in fields cleared after a long fallow period because trees and shrubs were oldest. Each individual or cluster of trees and shrubs stood on its own steep-sided pedestal, often over 15 cm high. Fallow-managed fields carried an average of 700 (long cycles) and 480 (short cycles) woody plants ha^-1. Manured “new” fields had an average densities of 340 woody plantsha ^-1 but manured “old” fields carried only 80 plantsha ^-1. Being relicts of the fallow vegetation, these plants managed to grow over successive cultivation periods, but woody plants tend to disappear with frequent cultivation. They also disappear gradually from manured fields, possibly because of heavy grazing.

Table 1. Analytical data from topsoil (0-20 cm) of pearl millet fields, Banizoumbou, Niger, according to cropping system

¹ Zebu dung, collected from the soil surface in may 1994 where it had dried in situ
² 1 part dung to 4 parts of water

Table 2. Mean maximum micro relief and soil surface characteristics according to cropping system, Banizoumbou Niger. Observations in 25 m² plots, 5-15 days after the sowing of pearl millet

Generally, the higher the micro relief, the more the soil is protected against erosion because with increased surface roughness, the water stays longer on the soil surface and thus can infiltrate. Thisseems to account for the sand dunes, the more and the higher these sand masses were, the less signs of water erosion were observed. The opposite seems to be true for the micro mounts formed by woody plants. In contrast with the gentle slopes of the aeolian sand deposits, these slopes were steep and impermeable. Instead of favouring infiltration, they accelerated the circulation of water over the soil surface. Field cultivated after long fallow periods had the highest micro relief and the largest amount of crusted surface, also showed evidence of much superficial water movement: gullies, sand fans, depositional crusts and eroded areas were frequent. By contrast, in fields where microdunes were highest and occupied most of the cropped surface, i.e. in manured fields, these indications of water erosion were scarce.

Pearl millet production

Sowing requires little effort in terms of time and sowing seed, compared to weeding and thinning. At the onset of the rains, as much land as possible is planted and later preferential choices are made as to in what part of the field crop care will continue. Most often, where crop establishment is bad, cultivation will stop. However, the transects were weeded and harvested even when the plots were regarded as hopeless by the farmers.

Table 3. Pearl millet production according to cropping system, Banizoumbou Niger. With indication of the proportion of the field abandoned after sowing because of expected low yield

¹ Cultivation continued in the experimental plots
² Farmers’ practice

Fields of the fallow system had large parts of abandoned land, compared to manured fields where almost the entire planted area contributed to grain yield (Table 3). In long fallow fields, unproductive areas were associated with gullies, coarse sand deposits (sand fans) and large tracts of erosion crusts. In short fallow fields, the abandoned areas were patches of extremely sandy soil, too poor to sustain production. In fields of the manure system, 90% to 100% of the sown land contributed to grain yield because gullies, sand fans were largely absent and because patches of extremely poor sand were fertilized with dung. Considering the grain production under farmer’s management, that is the yield from that part of the field where cultivation continued after sowing, then grain yields were slightly under 400 kgha ^-1 in all cropping systems, except for the short fallow fields without manure input where grain yields did not exceed 200 kgha ^-1 (Table 3).

Discussion and conclusion

In the Sahel ecosystem, the soil particles <20 µm of the topsoil constitute a capital on which the long-term success of pearl millet cultivation depends. This study demonstrates that this capital can be lost, because, with time, cultivated soils loose part of their clay and silt fraction from the topsoil. Though the loss of fine particles results in less surface crusting and visible products of water erosion like gullies and sand fans, and though most of the soil surface can be sown to pearl millet, the final result is the development of extremely skeletal soils. This loss constitutes a real threat to pearl millet cultivation as can be seen by the very low yield obtained in the plots that have been cultivated for a long time in alternation with short fallow periods.

Long-term processes

During fallowing, the woody vegetation is allowed to grow. Part of the vegetation that dies at the end of the rainy season is worked by the macro soil fauna and buried into the soil, thus increasing the soil organic matter content and forming more stable aggregates (Feller et al., 1989). Airborne particles accumulate on obstacles like shrubs and trees. Some of this material gets incorporated into the soil via stemflow beneath the shrubs or trees, and some is washed down to lower parts of the land (Ambouta et al., 1996). With land clearing and the cutting of the vegetation, not only the input of biomass stops but also the dust-trap disappears. While mineralization rates have gone up because of increased exposure, reduced organic matter means loss of soil structure and subsequently less stable aggregates. Repeated weeding further disrupts the aggregates and clay + silt particles are subject to erosion (Valentin, 1995; Pieri, 1989; Feller & Beare, 1997). Particles <20 µm are liberated when unstable aggregates are disrupted by raindrop impacts, they clog the pores and form crusts. Cover by crusts is highest in fields cultivated afer a long period of fallow because the building material of crusts is relatively abundant, secondly during cultivation the reduction of organic matter from the topsoil is probably more rapid than the loss of fine earth. On the other hand, high organic matter content increases the structural stability of soils, preventing them from disintegrating. This makes newly cleared fields that are manured less prone to crusting. Despite the still considerable clay + silt fraction of the topsoil, the decomposed dung mixed with soil by organisms ameliorates its structure so aggregates become less sensitive to disintegration. A second argument is that the soil surface of manured fields is littered with dung, thus dung protects the soil against aggressive rains. The farmers in Banizoumbou are aware that the issue of crusting, typical of long fallow fields can be greatly reduced by the application of manure. The best option is to start applying manure even before clearing, in the last year of fallowing. In the history of land use in Banizoumbou, this was a general practice when opening new land for cultivation (Loireau, 1998) but this option is now only open to the few well-off farmers.

Rajot (2001), by studying a field and an adjacent fallow during the same storm event, demonstrated that vegetated sites accumulated dust whereas cultivated fields lost dust. Wind erosion from fallowed land was always very limited, but it could be very large from the millet fields. He assessed that the mass budget of wind erosion (erosion versus deposition) at the scale of a village territory was positive in Banizoumbou, about 150 kgha ^-1year ^-1 and calculated that a further clearing for cultivation of only 6% of presently vegetated land would lead to a budget of zero (Rajot, 2001). Both the above-mentioned studies and our data show that the dominant process driving the fallow system is wind erosion. The loss of fine particles is enhanced by manual cultivation and the subsequent loss of fertility can only be restored by the trapping of dust during a long-term fallowing. Manure can only partly replace the fallow as a means of sustaining fertility. Dung application can supply the necessary nutrients to provide an average of 400 kg of grain yield/ha^-1. Under the current forms of manuring, visible traces of runoff and water erosion disappear and the soil organic matter content stabilises at a low level. However, the clay + silt content of the topsoil keeps declining with time.

Time span

How much time is needed to reach such losses of clay, silt and organic matter and to change a relatively productive skeletal soil, into an unproductive soil without the constant application of manure? As this is a long-term process, no direct measurements are available.

Ga-koudi, department of Maradi, Central Niger has similar climate and soils but the pressure on arable land is much higher than in Banizoumbou (Micheau, 1994; Wango, 1995; Dosso et al., 1996). Two types of soils are cultivated with pearl millet, one called Jigawa, a very sandy soil (typical value 55 gkg ^-1 of <20 µm) and the other Hako, a less sandy soil (typical value 81 gkg ^-1 of <20 µm). Local farmers reported that the Jigawa soils, located close to the village, were formally Hako, and they estimated that the transformation took about 40 years of cultivation (Dosso et al., 1996). They further ascertained that the reverse was possible under long term fallowing. Hako soils are regularly returned to fallow for short periods and seldom manured. They produce an average grain yields of 250 kgha ^-1, similar to those obtained in Banizoumbou under short cycles. All the manure is applied on the Jigawa soils, where the average yields reach about 300 kgha^-1 (Dosso et al., 1996). In Ga-koudi, the practice of long-term fallowing has disappeared. Most of pear millet production depends entirely nowadays on the yearly application of manure. However, dung is becoming increasingly scarce because fallow land, formerly used as pasture, has been cleared for cultivation.

A second estimate comes from Banizoumbou farmers. Farmers crop long season pearl millet land races called Somno (120-130 days) exclusively in long fallow fields because they consider that Somno requires a heavier soil, and they plant the short season Heinkirey land races (90-100 days) in the other fields. Farmers observe that after cropping the same field for a certain period (generally over 40 years) they must shift from long- to short-season cultivars because the topsoil becomes sandier.

The use of long-season varieties has become less and less frequent with time. The growing popularity of short-season land races should be attributed to the spread of very sandy topsoils and this is due to the gradual loss of the clay + silt fraction by wind erosion from cultivated soils, losses that are no longer compensated by long-term fallowing.

References

Abdoulaye, T., Lowenberg-DeBoer, J., 2000. Intensification of Sahelian farming systems: evidence from Niger. Agricultural Systems, 64, 67-81.

Ahn, P.M. 1970. West African soils. Oxford University Press, London, U.K., 332 p.

Ambouta, J.M.K., Valentin, C, Laverdière, M.R., 1996. Jachères et croûtes d'erosion au Sahel. Sécheresse, 7, 269-275.

Bationo, A. and Mokwunye, A.U. 1991. Role of manures and crop residue in alleviating soil fertility constraints to crop production: with special reference to the Sahelian and Sudanian zones of West Africa. Fertilizer Research, 29, 217-225.

Bielders, C.L., Michels, K., Rajot, J.-L. 2000. On-farm evaluation of ridging and residue management practices to reduce wind erosion in Niger. Soil Science Society American Journal, 64, 1776-1785.

Bielders, C.L., Rajot, J.-L., Amadou, M. 2002. Transport of soil and nutrients by wind in bush fallow land and traditionally managed cultivated fields in the Sahel. Geoderma, 109, 19-39.

Buerkert, A. and Hiernaux, P. 1998. Nutrients in the West African Sudano-Sahelian zone: losses, transfers and role of external inputs. Zeitschrift Pflanzenernährung und Bodemkunde, 161, 65-383.

Casenave, A., and Valentin, C, 1992. A runoff capability classification system based on surface features criteria in the arid and semi-arid areas of West Africa. Journal of. Hydrology, 130, 231-249.

Chappell, A., Warren, A., Olivier, M.A., Charlton, M. 1998. The utility of 137Cs for measuring soil redistribution rates in Southwest Niger. Geoderma, 81, 313-337.

Collinet, J., and Valentin, C. 1985. Evaluation of factors influencing water erosion in West Africa using rainfall simulation. In: Challenges in African Hydrology and Water Resources. IAHS publication, 144, 451-461.

D’Herbès, J.-M. and Valentin, C. 1997. Land surface conditions of the Niamey region: ecological and hydrological implications. Journal of Hydrology, 188-189, 18-42.

Dosso, M., Michau, P., Wango, O. 1996. Diversité des sols et pratiques de gestion de leur fertilité, en zone sahélienne sableuse Mayahi (Niger). In: Jouve, P., ed. Gestion des terroirs et des ressources naturelles au Sahel. CNEARC, Montpellier, France, 15-27.

Feller, C, and Beare, M.H. 1997. Physical control of soil organic matter dynamics in the tropics. Geoderma, 79, 69-116.

Feller, C, Fritsch, E., Poss, R., Valentin, C. 1989. Effet de la structure sur le stockage et la dynamique des matières organiques dans quelques sols ferrugineux et ferrallitiques (Afrique de l’Ouest en particulier). Cahiers ORSTOM, série Pédologie, 26, 25-36.

Harris, F. 1999. Nutrient management strategies of small­holder farers in a short-fallow farming system in Northeast Nigeria. Geography Journal, 165, 275-285.

Herrmann, L., Hebel, A., Stahr, K. 1994. Influence of microvariability in sandy sahelian soils on millet growth. Zeitschrift Pflanzenernährung und Bodemkunde, 157, 1-5.

Herrmann, L. 1996. Staubdeposition auf Böden West-Afrikas. Eigenschaften und Herkunftsgebiete der Stäube und ihr Einfluss auf Boden und Standortseigenschaften. Hohenheim Bodenkundliche Hefte n°36, University of Hohenheim, Stuttgart, Germany.

Landais, E. and Lhoste, P. 1993. Systèmes d’élevage et transferts de fertilité dans la zone des savanes africaines. II. Les systèmes de gestion de la fumure animale et leur insertion dans les relations entre l’élevage et l’agriculture. Cahiers Agricultures, 2, 9-25.

Loireau, M. 1998. Espaces - Ressources - Usages: Spacialisation des interactions dynamiques entre les systèmes sociaux et les systèmes écologiques au Sahel nigérien. Doctoral thesis 12 December 1998, University Paul Valéry Montpellier III, France, 393 p.

Micheau, P. 1994. Caractérisation des ressources naturelles renouvelables de l’arrondissement de Mayahi au Niger. Dynamiques et modes de gestion. Msc thesis, CNEARC, Montpellier, France, 101 p.

Möberg, J.P, Esu, I.E., Malgwi, WB. 1991. Characteristics and constituent composition of Harmattan dust falling in Northern Nigeria. Geoderma, 48, 73-81.

Pieri, C. 1989. Fertilité des terres de savane. Bilan de trente ans de recherche et de développement agricoles au sud du Sahara. Ministère de la Coopération/Cirad, Paris, 444 p.

Powell, J.M., and Williams, T.O. 1993. Livestock, nutrient cycling, and sustainable agriculture in the West African Sahel. Gatekeeper Series SA37, IIED, London, UK, 15 p.

Powell, J.M., Fernàndez-Rivera, S., Hiernaux, P., Turner, M.D. 1996. Nutrient cycling in integrated rangeland/ cropland systems of the Sahel. Agricultural Systems, 52, 143-170.

Rajot, J.-L. 2001 Wind blown sediment mass budget of Sahelian village land units in Niger. Bulletin Société Géologie de France, 172, 523-531.

Rockström, J. and Rouw, A. de 1997. Water, nutrients and slope position in on-farm pearl millet cultivation in the Sahel. Plant and Soil, 195, 311-327.

de Rouw, A. and Rajot, J.-L. 2004a. Nutrient availability and pearl millet production in Sahelian farming systems based on manuring or fallowing. Agriculture, Ecosystems & Environment, 104, 249-262.

de Rouw, A. and Rajot, J.-L. 2004. Soil organic matter, surface crusting and erosion in Sahelian farming systems based on manuring or fallowing. Agriculture, Ecosystems & Environment, 104, 263-276.

Sivakumar, M.V.K. 1992. Climate change and implications for agriculture in Niger. Climatic Change, 20, 297-312.

Valentin, C, and Bresson, L.-M. 1992. Morphology, genesis and classification of surface crusts in loamy and sandy soils. Geoderma, 55, 225-245.

Valentin, C. 1995. Sealing, crusting and hardsetting soils in Sahelian agriculture. In So, H.B., Smith, G.D., Raine, S.R., Schafer, B.M., Loch, R.J. Eds., Sealing, crusting and hardsetting soils: productivity and conservation. Australian Society of Soil Science, University of Queensland, Brisbane, Australia, 53-76.

Wango, O. 1995 Distribution des sols à l’échelle du territoire villageois de Gakoudi et pratiques traditionnelles de gestion de la fertilité. Msc thesis, CNEARC, Montpellier, France, 69 p.

¹ IRD, BP 06 Vientiane, Lao PDR