Abstract

The effect of 3 successive yearly applications of single superphosphate (SSP) to pastures on bicarbonate- extractable soil phosphorus (Olsen P) was measured. The soil was a duplex derived from basalt and the pastures, based on perennial ryegrass and subterranean clover, were continuously stocked with sheep. Six levels of SSP were compared at 3 stocking rates. The amount of P applied annually varied from 0 to 100 kg/ha. These data were used to create an empirical model which used the current value for Olsen P (Olsen P_n), the amount of P applied as fertiliser that year (fert P_n), and a lower limit for Olsen P for basalt-derived duplex soils (Olsen P_low) to predict the Olsen P for the following year (Olsen P_n+1). The model had the form: Olsen P_n+1 = Olsen P low + afert P_n + b(Olsen P_n – Olsen P_low). Olsen P low was fixed at 3 mg P/kg soil, and the coefficients a and b were 0.0995 and 0.8020. The model accounted for 96.6% of the variance in Olsen P_n+1. This model was tested at the same site at 2 other periods: when fertiliser was withheld for 3 years and again after applications of SSP were resumed. The model was also tested against data from another experiment conducted on a similar soil. The model can estimate the amount of fertiliser required to maintain the P status of the soil and predicts that to increase Olsen P by 1 unit in the following year it is necessary to apply 10 kg P/ha in excess of soil maintenance requirement.