Comparison of phosphorus fertilization diagnostic methods for wheat under no-tillage

Wheat (Triticum aestivum L.) production under no-tillage in the southeast of Buenos Aires province (Argentina) has increased markedly over recent years. It is therefore necessary to assess some of the criteria used in diagnosing the response of wheat crops under no-tillage to phosphorus (P) application: extraction method and soil sampling depth. Available P was determined on 25 farm fields by the Bray 1 and Mehlich 3 methods in samples taken to a soil depth of 20 cm and segmented into 0–5, 5–10 and 10–20 cm. Wheat yield response was expressed as yield difference and relative yield and then related to available P by using quadratic, linear-plateau and quadratic-plateau models for the depths of 0–5, 0–10 and 0–20 cm. Available P concentrations using Bray 1 and Mehlich 3 methods were positively correlated (r = 0.95). In slightly acid soils, both methods extracted similar concentrations of P (15.9 and 16.4 mg P kg−1 for Bray 1 and Mehlich 3, respectively). In one site with soil pH > 7, up to four times more P was extracted by Mehlich 3 than by Bray 1. Relative yield was better than yield difference as expression of crop yield response (greater R2). The quadratic and linear-plateau models were better at describing the relationship between crop yield response and available P than the quadratic-plateau model, which was not adequate for some combinations of sampling depths and extracting solutions. Shallower soil samples did not improve the R2 values in the models assessed, and samples from 0 to 5 cm depth showed the lowest adjustments (R2 = 0.35–0.56 and 0.57–0.72 for 0–5 and 0–20 cm, respectively). For estimating P fertilizer requirements for wheat crops under no-tillage, it is recommendable to sample soils before sowing to 20 cm depth. There was a strong positive correlation between Bray 1 and Mehlich 3 procedures, thus either could be used to quantify available P in soils with moderate to slightly acidic pH values.