Recommandations specifiques d'engrais: calibration et validation du module phosphors du modele numass

摘要

Fertiliser recommendations in Mali as well as in many other countries of West Africa were made according to Chaminade's method. For socio-economic reasons, correcting deficiency rates of major nutrients were vulgarized. These blanket recommendation,when applied, lead to a continuous soil nutrient mining. The objective of this research was to calibrate and validate the P module of NuMass. Laboratory incubations were conducted to calibrate the P buffering coefficients used by the P module. Rates ofN, P and lime predicted by NuMass model, considered as specific recommendations, were compared to the blanket recommendations in the field in order to validate them. The buffer coefficients were inversely proportional to the clay content. The buffering coefficient of sandy soils of Cinzana/Mali (0.73) and Kollo/Niger P3 (0.63) were higher compared to the clayey soils of Longorola-bf (0.22) and Kollo/Niger (0.21). Generally, buffering coefficients obtained by laboratory incubation (0.60) tended to be lower than the estimated coefficient by the P module NuMass (0.67). The range of the used soil texture (1.5-54.6 percent clay) showed that the buffering coefficients estimated by the P module of NuMass were correct for flooded soils. Grain yield of different trials and tests do not indicate the expected higher performance of specific recommendations of fertilisers from NuMass model compared to the used blanket recommendation (1624 and 1582 kg ha~(-1) of maize; respectively).