Management of Soil Fertility in a Long-Term Field Trial of Semi-arid Kenya

Abstract

The long-term experiment was initiated in 1989 on a P-deficient soil in semi-arid Kenya where sustainability of the cereal/legume intercropping was assessed by monitoring trends in grain yield, soil organic C (SOC) and soil extractable P (Olsen method). Goat manure was applied annually at 0, 5 and 10 t ha−1 and trends in grain yield were not identifiable because of season-to-season variations. SOC and Olsen P increased for the first 7 years of manure application and then remained constant. The residual effect of manure applied for 4 years only lasted another 7–8 years when assessed by yield, SOC and Olsen P. Mineral fertilizers provided the same annual rates of N and P as in 5 t ha−1 manure and initially gave the same yield as manure, declining after 9 years to about 80%. There was a wide agreement between observed and predicted dry matter yields, soil labile P, but less for organic C using the Agricultural Production Simulator (APSIM) model. Soil nitrate was highest at the onset of the season and was highest where manure had been applied and was subject to loss through leaching or runoff