Relationships between soil test phosphorus, soluble phosphorus, and phosphorus saturation in Delaware soils

Methods to identify agricultural soils that contribute to nonpoint-source pollution of surface waters by P are of increasing importance, particularly in areas with high animal densities (animal units per hectare of cropland). Our objective was to determine the relationship between agronomic: soil test P (STP = Mehlich 1) and other soil P tests proposed to measure the potential for P loss by erosion, runoff, and leaching. We compared STP with soluble P, P in the "fast desorbing pool" (strip P), and soil P saturation for 127 soils (122 from Delaware and five from the Netherlands). Soil test P was significantly correlated with total P (r = 0.57***, significant at the 0.001 level), soluble P (r = 0.71***), strip P (r = 0.84***), and oxalate-extractable P (P-ox; r = 0.84***). Strip P was a better predictor of soluble P than STP (r(2) = 0.76***). The ratio of strip P/P-ox (the percentage of reversibly sorbed P in the fast desorbing pool) increased as P sorption capacity, estimated from oxalate-extractable Al and Fe (Al-ox + Fe-ox), decreased. We also determined the degree of P saturation (DPS) using three methods: Langmuir P sorption isotherms; oxalate extractions of P, Al, and Fe; and STP plus a single-point P sorption index (PSI). Soluble P, STP, and desorbable P increased for DPS values >30%, similar to upper DPS limits in the Netherlands and Belgium. Soils rated agronomically excessive in STP (>50 mg kg(-1)) had higher ratios of soluble P, strip P, and P-ox to total P than those in agronomically optimum or lower categories.