Assessing the interaction between cropping density, nitrogen, and soil moisture with a systems simulator: a case study for semi-arid Machakos District of Kenya

The APSIM model was used to evaluate a range of maize densities on a sandy-clay loam at KARI-Katumani research centre in Machakos and two contrasting farmers' fields situated on sandy and clay soil types. Each simulation was initialised with actual soil parameters as measured for the time of commencing the run. Simulations on the Katumani soil were done under five contrasting short rainy (SR) seasons of different rainfall regimes, namely SR1996 (185mm), SR1995 (270mm), SR1991 (328mm), SR1992 (720mm), and the El Nino season of SR1997 (958mm). Maize densities used were 1.0, 2.0, 3;0,4.0, 5.0, 6.0, 7.0, 8.0, 9.0, and 10.0 plants m'2. The model showed that maize grain yield declined as density increased under the poor 1996 season. Optimum crop density increased with seasonal rainfall, to reach 5.0 plants m'2 in the 1992 and 1997 heavy rainfall seasons. On the farmers' fields, APSIM showed that the clay soil out-yielded the sandy in high rainfall, while the reverse was true on poorer seasons. Simulations were done to determine the effects of varying nitrogen at constant crop density of 5.3 plants m-2. Nitrogen rates used were 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, and 120 kg ha'l. The response of maize to nitrogen fertiliser varied with seasonal rainfall. Grain yields were depressed when nitrogen was added to maize during the poorer SR1996 (185mm) and SR1995 (270mm) seasons. With higher seasonal rainfall, APSIM predicted a sharp increase in grain yield at low rates ofN. Simulations of the effects of nitrogen on maize yield on the three farmers' fields (clay and sandy soils) were done for SR1997 (13lOmm) and SR1998 (122mm) seasons, using N rates of 0, 10, 20,30,40,50,60,70,80,90, and 100 kg ha.l. The model predicted crop failure for the SR1998 season on the clay soil throughout all levels of N rates. In the wet SR 1997, grain yield was low at 0102030 kg N ha.I, and then increased sharply at levels greater than 30 kg N ha'.. The response to nitrogen was, therefore, highest in the clay soil under wet conditions. Grain yield responded well in the sandy soil during SR1997, though the rate of response was lower than in the clay soil. However, the sandy soil proved superior to clay in the poor season (SR1998). A gentle response to nitrogen was predicted in the sandy soil, up to 30 kg N ha'i. Sowing date was more important for sandy than clay soil type. Delayed sowing on the clay soil type did not adversely affect yield, whereas a week's delay on the sandy soils led to yield decline even in a favourable season (SR 1997)