Does long-term low-rate organic and inorganic nitrogen management guarantee maize yield under semi-arid conditions of eastern Kenya?

Drier parts of Embu County endure high atmospheric heat, prolonged dry spells, declining soil fertility and erratic rainfall. Integrated soil fertility management (ISFM) technologies have been developed, and tested, with standardized application procedures and rates aimed at optimizing overall yield productivity. Despite their significant impacts, high variability in local soils and climate contributes to large variations and inconsistence in research results among replications of the expensive and limited (time/site/scenarios) experimental treatments. This occasions poor comparability of results within-and-without different agro-ecological zones (AEZs) due to the complex non-linear soil-climate-crop relations. Crop-growth simulation models suitably complement experimental research, to support decision making regarding soil fertility and water management. This study evaluated the performance of the Agricultural Production Systems Simulator (APSIM) model, parameterized and calibrated based on a rain-fed randomized complete block trial (2010-to-2013) at a research station in Machang’a, Embu County; and the long-term N management in maize production. The trials combined three and two level amendments each as follows; inorganic fertilizers (30 kg N ha-1, 60 kg N ha-1 and 90 kg N ha-1), goat manure, Lantana camara, Tithonia diversifolia (each at 30 kg N ha-1 and 60 kg N ha-1), Mucuna pruriens ( N depended on residue harvested; in a mirror rotation with maize) and control (TSP 60 kg ha-1); replicated three times. Half rate (30 kg N ha-1) applications were complimented by half rate inorganic fertilizer application. Model validation was based on goodness-of-fit between observed and simulated parameters derived from residual-errors statistics; root mean square error (RMSE), square of the correlation coefficient (R2), and model efficiency (EF). APSIM simulations adequately predicted observed maize crop-growth (Leaf Area Index; LAI, Grain yield, and biomass). Grain prediction across the treatments was good (R2=0.87 and EF>0.9) but biomass was slightly under-predicted (R2=67 and EF=0.87). Long-term (26 cropping seasons) simulations showed that moderate and low cost application of N (40 kg N ha”1 from combined manure and mineral fertilizer) improved both long-term average and the minimum guaranteed grain yield (2.5 Mg ha-1) and thus recommended for smallholder farmers especially in dry areas. These findings should be considered in conditions where P is added proportionally to N (P/ N in the range of 20 to 30%).