The availability to plants of fertilizer phosphorus (P) applied to soil, as measured by chemical extraction, is used to estimate P fertilizer needs. We studied the availability of P, applied as monocalcium phosphate (MCP) powder, ordinary superphosphate (OSP) granules and diammonium phosphate (DAP) granules in 24 calcareous Vertisols and Inceptisols of Andalusia, Spain, by using laboratory incubation techniques. The soils differed widely in their P adsorption- and Ca-phosphate precipitation-related properties. For MCP, availability (defined as the proportion of added P that is recovered by extraction with NaHCO3 or is isotopically exchangeable) decreased markedly with incubation time and increasing addition rate. The mean recoveries after 180 d of incubation at field capacity at a rate of 246 mg P kg−1 soil were 17% for Olsen P, 38% for Colwell P, and 16% for isotopically exchangeable P (IEP). Increasing the application rate to 2460 mg kg−1 resulted in recoveries of 6% for Olsen P, 25% for Colwell P, and 4% for IEP. While IEP-based recovery was not significantly correlated to any soil property, that based on Olsen P (and, to a lesser extent, Colwell P) decreased sharply with increase in the ratio of clay (or Fe oxides) to total (or active) calcium carbonate equivalent. Accordingly, Olsen P might overestimate P availability in those soils relatively rich in carbonate and poor in clay and Fe oxides. On the other hand, recovery of applied P from soils containing more clay and Fe oxides, by a sequential extraction (with H2O, two 0.5M NaHCO3 treatments, 0.5M HCl), was lower than 100%, thereby suggesting phosphate occlusion by Fe oxides or clay.
Read full abstract