Towards a single-point method for measuring phosphate sorption by soils

Abstract

Data from 4 separate studies using widely differing soils were used to assesswhether single-point estimates of the relative buffering could be used asestimates of buffering derived from fitted curves. The relative bufferingcapacity of soils for P was summarized using the ‘O&Svalue’. This is the increase in P sorption between concentrations of 0.25 and 0.35 mg P/L.Phosphate sorption curves were described by the modified Freundlich equation:S = acb – q, whereS ( g/g) is the measured sorption,c ( g/mL) the solution concentration, anda, b, and qare parameters. The fitted O&S value is given by:a (0.35 b – 0.25b ). If a single pair of values for sorption andconcentration is measured, rather than a sorption curve, thea value can be estimated from (S+ q)/cb . An estimate of the O&S value can beobtained from (S + q) (0.35b – 0.25 b)/c b .Because the b parameter occurs in both the numerator andthe denominator, the single-point estimate of the O&S value was notvery sensitive to the value allocated to b provided thevalue of c was in the range 2–5 mg P/L. Formost of the soils of low to moderate buffering, the P extracted by the Colwellreagent was a useful estimate of q. For these soils, thesingle-point estimates using this value of q weresimilar to the values found from the fitted curves. One group of exceptionswere some soils of low buffering which had received a single application ofphosphate at a high rate. For these soils, the Colwell reagent extracted largeamounts of phosphate. Another group of exceptions were soils of highbuffering. For these the Colwell reagent gave values much lower than thefitted value of q and was of little value in improvingthe calculation of single-point estimates. The single point estimates forthese soils were about 20% too low but nevertheless well correlatedwith the values from the fitted curves.