Long-term measurements and model simulations of phosphorus leaching from a manured sandy soil

Cropping systems with high phosphorus (P) inputs may constitute a risk of P leaching, which contributes to eutrophication. The main objective of this study was to identify P leaching risks associated with three long-term fertilization regimes in separately tile-drained plots on a sandy soil in southwest Sweden. The three regimes resulted in different annual P surpluses of, on average, 16 kg P ha(-1) (14 lb P ac(-1)) in mineral form and 18 kg P ha(-1) (16 lb P ac(-1)) and 37 kg P ha(-1) (33 lb P ac(-1)) as pig slurry. The importance of different soil characteristics (soil P. iron, aluminum, and calcium content, and degree of P saturation [DPS]) and processes (water flow and P sorption/desorption) was examined using 15 years (1989 to 2003) of P leaching measurements and simulations with the ICECREAM model. Measurements of high soil P content and DPS values in the topsoil, in combination with high precipitation and rapid water flow, indicated a high potential for P losses, which was confirmed by the model simulations. However, the model considerably overestimated total P leaching by a factor of 5 to 9 since measured P leaching was small for all treatments. Measured mean annual total P leaching and total P concentration ranged respectively from 0.14 kg ha(-1) (0.12 lb ac(-1)) and 0.06 mg L-1 (3.75 x 10(-6) lb ft(-3)) at a high rate of slurry application to 0.20 kg ha(-1) (0.18 lb ac(-1)) and 0.08 mg L-1 (4.99 x 10(-6) lb ft(-3)) in the mineral P treatment. The differences in concentration were statistically significant (p < 0.001). A main conclusion from this 15-year study was that annual pig slurry application rates of 37 to 58 kg P ha(-1) (33 to 52 lb P ac(-1)) did not increase P leaching. High sorption capacity of the subsoil, caused by Fe, Al, and Ca, was obviously very important for controlling P losses. Thus, information on soil P content and fertilization must be supplemented with estimates of soil P sorption capacity when evaluating the risk of P leaching for different soils. This must also be considered in models used for assessment of P leaching from arable land. The current ICECREAM model does not include appropriate functions for describing P sorption/desorption processes in this type of soil and needs further development.