Residual value of superphosphate for oat and barley grown on a very sandy,phosphorus leaching soil

In a field experiment in a Mediterranean climate (474 mm annual rainfall, 325 mm (69%) falling in the May to October growing season) on a deep sandy soil near Kojaneerup, south-western Australia, the residual value of superphosphate was measured relative to freshly-applied superphosphate. The grain yield of five successive crops (1988–1992) was used to measure the residual value: barley (Hordeum vulgare), barley, oat (Avena sativa), lupin (Lupinus angustifolius), and barley. There was no significant yield response to superphosphate applied to the first crop (barley, cv. Moondyne). There were no results for the second crop (barley) due to weeds or the fourth crop (lupin) due to severe wind erosion which damaged the crop. The residual value of superphosphate was measured using grain yields of the third crop (oat, cv. Mortlock) for superphosphate applied one and two years previously, and the fifth crop (barley, cv. Onslow) for superphosphate applied one, two, three and four years previously. In February 1992, before sowing the fifth crop, soil samples were collected to measure bicarbonate-extractable phosphorus (P) (soil test P) which was related to the subsequent grain yields of that crop. This relationship is the soil test P calibration used to estimate the current P status of soils when providing P fertilizer recommendations.The residual value of superphosphate declined markedly. For the third crop (oat), it was 6% as effective as freshly-applied superphosphate one year after application, and 2% as effective two years after application. For the fifth crop (barley), relative to freshly-applied superphosphate, the residual value of superphosphate in successive years after application was 46%, 6%, 3% and 2% as effective. The soil has a very low capacity to sorb P, and P was found to leach down the soil profile. The largest yield for P applied one and two years previously in 1990, and two, three and four years previously in 1992, was 35 to 50% lower than the maximum yield for freshly-applied P.Soil test P was very variable (coefficient of variation was 32%) and mostly less than 8µg P/g soil. The calibration relating yield (y axis) to soil test P (x axis) differed for soil treated with superphosphate one year previously compared with soil treated two, three and four years previously. The top 10 cm of soil was used for soil P testing, the standard depth. P was leached below this depth but some of the P leached below 10 cm may still have been taken up by plant roots. Consequently soil test P underestimated the P available to plants in the soil profile. The soil test P calibration therefore provided a very crude estimate of the current P status of the soil.