Identification and characterization of phosphorus-rich sandy soils

For the identification of phosphorus-rich and phosphorus-leaky soils, various extraction methods have been proposed as indicators to estimate the risk of phosphorus (P) losses. In this study we compare and discuss the extractability of P from sandy soils, using various modifications of mild extractants (water and CaCl2), FeO-impregnated filter-paper strips (Pi-test), and ammonium oxalate (P-ox). In the soils studies, the amounts of water- and CaCl2-extractable P were (much) smaller than the amounts of P extracted with FeO-strips or with ammonium oxalate. With the waterbased extraction method Pw (1:60 v:v soil:solution ratio), end-over-end shaking (Pw-2) compared with reciprocative shaking (Pw-1) resulted in increased amounts of extractable P in all soils. Because in the original Pw procedure reciprocative shaking and end-over-end shaking were implicitly treated as being exchangeable, the shaking method needs further standardization. Exponential relationships were found between P-1:2 (water-extractable P at a 1:2 w:v soil:solution ratio; roughly comparable with P in the soil solution) on the one hand, and Pw-1 (R2 = 0.99; P < 0.001) and Pw-2 (R2 = 0.96; P < 0.001) on the other. The non-linearity of such relationships must be taken into account if Pw is used as an indicator to estimate the risk of P leaching, i.e., at high values of Pw, leaching may be underestimated if a linear instead of a non-linear relationship is used. A similar exponential relationship was observed between P-1:2 and the P saturation degree (PSD) for soils sampled from the same site (R2 = 1.00; P < 0.001). The relationship was not valid for soils from other sites, suggesting that the PSD as such cannot be used as an indicator for the risk of leaching dissolved P. Based on both PSD and P-1:2, a simple alternative scheme is proposed to provide guidance for selecting P-leaky soils and for evaluating the effectiveness of measures to restore such soils