On-farm agrohydrological analysis of the Sahelian yield crisis : rainfall partitioning, soil nutrients and water use efficiency of pearl millet

Abstract: Sub-Saharan Africa is presently experiencing a steadily aggravated food security crisis. This crisis is a result of a rapidly growing population combined with insignificant yield increases of the major cereal crops maize, millet and sorghum, during the last decades. The situation is most severe in semi-arid regions where water is a major constraint on food production. The thesis addresses the human impact on crop yield development during the last century and the biophysical causes underlying the gap between extremely low on-farm yields and potential yields of pearl millet (Pennisetum glaucum (L.) Br.) in the semi-arid Sahel region in West Africa.Results are presented from two studies in Niger. The first analyses the driving forces behind the agricultural crisis in an agrarian system in the region of Zarmaganda. The second study analyses rainfall partitioning and crop water use efficiency in a farmer's field, located along a catena characteristic of the Sahel, in the Samadey watershed.Water scarcity in food production in semi-arid tropics is not necessarily a result of the low annual rainfall levels (ranging from 200 - 600 mm), but rather a result of a high rainfall variability between and within years and an extremely high evaporative demand of the atmosphere (600 - 900 mm during the 3-4 month rainy season). In addition to these two climatic modes of water scarcity, humans induce water scarcity with land degradation, which results in lowered infiltration and reduced soil water holding capacity.The Zarmaganda study shows that increased population pressure has forced a shift in the agrarian system during the last century from a system based on long bush fallows, to a "nutrient-mining" system with short (< 10 years) or abandoned fallows. This development is proposed to be the most significant structural transition of dryland agriculture in the Sahel region since its introduction.The results from the Samadey experiment indicate that low on-farm yields are caused by water and soil fertility constraints. Crop water scarcity is caused by rainfall erraticness, leading to episodic droughts even during average rainfall years, and runoff production as a result of soil surface crusting. Non-fertilised millet had a grain yield of 347 - 422 kg ha-1 compared to 526 - 697 kg ha-1 for the fertilised crop, for the 3 studied rainy seasons 1994-96. Substantial volumes of surface overland flow were measured. Observed inflow of sheet flow as runon from degraded upstream zones, corresponded to an additional 20 - 50 mm of water if distributed over the 8.5 ha field. Runoff production on a plot scale (15 x 6 m) amounted to 10 - 13 % of the annual rainfall. The study shows that runon and runoff flow can constitute significant components of the on-farm water balance.A systematic slope gradient was observed where yields decreased 35 - 40 % from the downslope to the upslope position along the 315 m gently sloping catena. This is explained by decreased soil water availability, caused by surface overland flow, and a low absolute but significant relative decrease in soil fertility, when moving upslope. The effect of this soil water gradient was that the upslope crop suffered more from episodic droughts, which hit the crop during panicle initiation in 1994, during grain filling in 1995, and during flowering in 1996. Spatial variability of soil water was high, with the mean infiltration around neutron probe access tubes ranging from 0.43 - 1.13 times rainfall.The variation of infiltration over time was high, due to rapid changes in crust coverage and the effect of rainfall intensity on runoff. This explains how a location in the field could function both as a runoff and a runon zone.Estimates from water balance modelling of seasonal water flow partitioning for runoff and runon producing zones in the field, indicate very low productive water losses by plant transpiration. These losses account for only 4 - 9 % of the available water (rainfall + runon) for the non-fertilised crop, and 7 - 24 % for the fertilised crop. This is due to the low leaf area observed in the farmer's field (maximum LAI = 0.44 m2 m-2 for the non-fertilised crop). Soil evaporation was high, amounting to 32 - 50 % of the available water, and conservative, presenting insignificant reductions with increased leaf area (LAI < 1 m2 m-2 for all treatments). Deep percolation was high, amounting to 140 - 250 mm. Large non-productive losses of water, together with low yields, resulted in low water use efficiencies (WUE), with evapotranspirational WUE = 6000 - 8000 m3 ton-1 grain, and rainfall WUE = 12,000 - 16,000 m3 ton-1 grain.Effects of rainfall erraticness, resulting in an unfavourable distribution of rainfall over time, explains why a crop that uses such a small proportion of the available water, in an environment with substantial deep percolation, still suffers from water scarcity.Large volumes of runon and the frequent dry spells, indicate the possibility of increasing yields with rainwater harvesting techniques for supplemental irrigation. The risk of crop failure would then be reduced, which possibly could increase the incentive for the farmer to invest in soil fertilisation. Such soil and water interactions would enable a long-term win-win solution for the farming system.The high spatial and temporal variability of soil nutrients and water availability should be considered as a non-negotiable part of the on-farm reality, which could constitute an opportunity, not only a constraint, for increased crop yields. Moreover, the development of site specific farming strategies, inspired by development in temperate regions, and cropping practices adapted to a variable toposequence availability of water, inspired by farmers in less arid tropics, could possibly be interesting in the semi-arid Sahelian savannah.