Modelling crop response to phosphorus inputs and phosphorus use efficiency in a crop rotation

Phosphorus (P) is one of the macro nutrients required by crops, and is deficient in many agricultural soils. Due to the low recovery and residual effects of applied P, long-term measurements are required to quantify crop response to P inputs. A modelling approach enables integration of available data to investigate response of crops to P additions in different soils. The APSIM agricultural production systems model has an ability to simulate crop response to soil P. However, the model has only been tested against a few datasets for maize, sorghum and bean crops outside Australia. It has not been applied to Australian conditions, nor against other crops. In this paper we derived the model parameters for wheat, soybean and peanut based on literature and measurement data. The parameterised model was tested for simulation of biomass growth, grain yield, P uptake and grain P contents of four crops (wheat, maize, soybean and peanut) in response to different levels of P addition in a 6-year rotation on a Red Ferrosol soil at Kingaroy, Queensland, Australia. Overall, the model could explain 87% of the variation in observed total aboveground biomass, 84% in grain yield, 75% in crop P uptake and 73% in P content in grain. It was also able to reflect the differential P recovery of crops in the rotation under the 5 different P input levels. In addition, a close relationship was found between modelled labile P and measured bicarbonate extractable P (Colwell P), which may help to initialise APSIM using Colwell P measurements in future studies. It is noted that much of the uncertainties in modelling crop response to phosphorus input arise from the uncertainties in input variables, particularly those describing soil P sorption capacities. Unless P sorption characteristics of a soil in a particular region are quantified, it remains difficult to conduct meaningful modelling research or develop model applications.