EFFECT OF EXTRACTABLE SOIL SURFACE PHOSPHORUS ON RUNOFF WATER-QUALITY

Phosphorus (P) additions to surface water from agricultural nonpoint s ources are of concern, because P often limits eutrophication of surfac e waters. Numerous sources of runoff P exist: indigenous soil and plan t material, land-applied manure and sludge, and commercial fertilizer. In many soils receiving commercial fertilizer and manure, concentrati ons of P at the soil surface have been steadily rising due to either l ong-term or excessive applications of P. Critical levels of soil surfa ce P may exist, above which runoff may promote eutrophication. Methods for rationally identifying these critical levels are needed to manage losses of P, which implies the need for accurate methods of relating soil surface P concentration (P(s)) to runoff P concentration. A study was conducted on both pasture and tilled plots (with and without resi due) to evaluate the relationship between P(s) and dissolved reactive P in runoff (P(R)) using simulated rainfall. The data indicated that e ven for comparable storms, P(s) alone was not a satisfactory estimator of P(R). A model describing the kinetics of P release from surface so il to runoff was used to include additional variables in predicting P( R). When used with uncalibrated parameters, the model explained a sign ificant proportion of the variation in observed P(R) values for pastur e plots (r2 = 0.43) but was less successful in predicting PR for tille d plots (with and without residue, r2 = 0.13). Calibration of (adjustm ents to) the extraction coefficients resulted in an overall coefficien t of determination between observed and predicted P(R) values of 0.73. While the model was successful in describing how P(R) and the indepen dent variables are related for the pasture plots, the extraction coeff icients should be calibrated to obtain best estimates of P(R). When us ed with calibrated extraction coefficients, the model provided realist ic estimates of P(R) over the range of experimental conditions.