Spatial variability and farmer resource allocation in millet production in Niger

The Sahel of West Africa is the agro-ecological zone located between 12 <sup>o</SUP>and 16 <sup>o</SUP>N, with an annual rainfall of between 300 and 1000 mm. Crops are grown in a subsistence type of agriculture during the 75 to 125 days growing period between May and September. Major crops are millet, sorghum, cowpea, and groundnut, sown usually in mixed-cropping. Crop production suffers from frequent droughts and poor soil fertility, which cause low yields and repeated food shortages. Traditional farming, in the western part of Niger, relies on the use of livestock to combat soil nutrient deficiencies, and other risk minimizing strategies, to produce an average 350 kg ha <sup>-1</SUP>year <sup>-1</SUP>of millet grain. This is the context within which this research was conducted in four localities in western Niger in 1995, 1996, and 1997.The overall aim was to investigate soil and crop variability at various scales and the rightness of traditional land management in millet cropping subject to extremely variable weather conditions. The main objectives were:to test and refine the hill scoring technique for characterizing within field soil and crop yield variability in 3 localities of western Nigerto identify soil properties explaining yield variability in space and timeto evaluate the effects on millet yield of site specific management of manure and fallow in a Fulani farmer's fieldto compare management systems used by farmers in four different landscape positionsto describe farmers methods of land evaluation and management options, based on farmers' expert knowledge of the environmentWithin-field soil and crop variability are discussed in chapters 2 and 3. Grain and straw yields varied considerably within-field at each of the three sites, and also between sites and between years. In 1996, grain yields varied from 8-383 kg ha <sup>-1</SUP>, 2-1343 kg ha <sup>-1</SUP>, and 7-815 kg ha <sup>-1</SUP>at the three sites along a 400 km N-S gradient. The coefficient of variation at the three sites was 61, 55, and 53% respectively. Only 5 to 28% of the yield variability could be explained by soil chemical properties. Other factors such as micro-topography and water redistribution also made yields vary over short distances according to several studies in the region. A millet hill scoring method was tested and refined to obtain yield estimates before crop harvest. Scoring studies have remained at the row or plot level as opposed to the individual hill level. The method was simple and was able to explain up to 67% of the yield variation. Yield variation between years was least at the drier northern site with poorer soil and yields. Variation in yields among years was most closely correlated with soil pH and Al in the northern and central sites, whereas P was the best in explaining grain yield differences in the southern site.Chapter 4 deals with land management at the individual household level. Soil and crop management were investigated in two fields owned by a Fulani household. Management practices created man-made variability in the fields. Soil nutrient contents were improved in patches through the use of field corralling of livestock and fallow. Quantities of manure applied at specific sites varied from 1500 to 17000 kg ha <sup>-1</SUP>. Corralling increased yields from 500 to 1100 kg ha <sup>-1</SUP>, but amounts of manure applied were too high in some spots and losses were likely to occur through leaching. Nutrients and pH changes in the soil profile were found to change appreciably over time following the application of manure or of fallow. Based on the results obtained, an improvement in the farmer's management can be achieved by spreading manure to an area 3 to 5 times larger than the area actually used.Chapter 5 focuses on farmers' land management practices according to landscape position. Contrary to common belief, soils on the plateau were richer than soils on other landscape positions, which were cropped continuously and had a sandier texture. Plateaus, however, showed limitations such as low infiltration of rainfall and were located relatively far away from villages. Fields on undulating terraces and valleys had the highest number of different management practices, among which the use of livestock corralling and fallow were the most important.Chapter 6 considers social aspects of soil management, through the use of ethno-pedology by farmers in ordinary cropping activities. Farmers' knowledge of the land, plants and weather is described as well as the use of this knowledge by farmers in millet production. Simple soil descriptors such as texture, color, surface condition and indicator plants have been used by farmers to evaluate the land. Management practices rely on local knowledge regarding variability in soil fertility and water availability as a guide to implement crop management decisions and to design risk reducing strategies. These compensate in part for the lack of external inputs. The value of ethno-pedology for modern research is that it can serve to better identify farmers' needs and, as a result, improve research quality and the effective communication of results obtained. However, limitations in the use of ethno-pedology exist and they include its inability to quantify factors and the confusion between causes and effects for processes not well understood by farmers