Evaluation et déterminisme du bilan de l'azote en sols cultivés du centre Burkina Faso : étude par traçage isotopique 15N au cours d'essais en station et en milieu paysan

In the semi-arid and arid zones of Africa, nitrogen is a major limiting factor for crop production but, paradoxically, N can be in excess in the soil during sorne periods of the rainy season. The objective of this study is to have a better knowledge of N losses under maize and sorghum in order to reduce the N20 emission and to improve the efficiency of mineral N fertiliser in these regions. For that, we (i) quantified globally the N losses and distinguished those coming from N mineral fertiliser from N losses coming from other N sources, and, (ii) quantified the sources of N losses. In a second step, the obtained data were used to identify indicators, easily measurable, permitting to know the part of total N in the above ground plant components that is taken up from the soil (NdfS %). Our study was conducted in a research station and by onfarm research in Burkina Faso, and also in laboratories on a Luvisoil. 15N was used to determine N fertiliser efficiency. The results show sorne periods during the raining cycle with (i) a negative nitrification that we attribute to denitrification and, (ii) others events with an excess of N03-N in the soil as compared to the N demand of the cereals. Total N losses measured during ten years, appear to be linked positively with (i) the decrease of soil pH ( 20 kg ha-1 y-1 of N loss in soils that lost pH, but only 8 kg N ha-1 y-1 in soils receiving manure and showed no change in pH), and, (ii) with the amount of mineral or organic amendements applied to the soil: a) liming when the soil receives also manure (loss of 68 kg N ha-1 y-1 with manure + lime, but only 8 kg N ha-1 y-1 with manure alone); b) increase ofN fertiliser amount (loss of 63 vs.20 kg N ha-1 y-1 for high and low level of urea application respectively); c) increase in manure application (225 vs.8 kg N ha-1 y-1 with high and lower level respectively). Processes behind these N losses are leaching and denitrification; volatilisation concerns mainly N from the urea. Leaching under sorghum on the research station (mean annual rainfall 850 mm), does not exceed 10 kg N ha-1y -1 at the depth of 90 cm. Denitrification data, obtained in vitro, tend to show that (i) this risk exist in the field, (ii) is increased by manure application, and (iii) that this risk is positively related by the clay content of the soil. N losses from urea applied without and with manure are 31 % and 37 % respectively from which 7 and 0.4 % respectively is leached and 24 and 37 % goes to the atmosphere. Volatilisation, measured in laboratory, can reach 30% of the applied urea. NdfS% is high (between 60 and 80 %) and is linked positively to the efficiency of the N fertiliser. Increase of N fertiliser efficiency needs therefore a reduction of the N losses (by working the urea in the soil for example) and an increase of the pool of N in the soil that can be mineralised (for example, by increasing the symbiotic N2 fixation capacity of the cowpea). Statistical analysis shows also that, in the absence of mineral N fertiliser, the total N amount in the above ground plant components (Ntot plant) is a useful indicator for Ndts %, and that Ntot plant is a function of the levels in mineral N and exchangeable K in the soil before planting. (Résumé d'auteur)