Fate of phosphorus in soil during a long‐term fertilization experiment in Finland

Accumulation and depletion of soil phosphorus (P) was studied in a long‐term (37 y) field experiment in Southern Finland. The loam soil had a high pH (7.5–7.7) due to an earlier liming. Spring barley, spring wheat, oat, and ryegrass, grown in rotation, were annually fertilized with 0, 32, or 67 kg P ha −1 y −1 (P 0 , P 1 , and P 2 K) and sufficient N. The average dry matter grain yield 2,600 kg ha −1 of the P 0 plots increased by about 500 kg ha −1 at P 1 treatment and another 600 kg ha −1 by P 2 K. Soil samples were collected in 1978 (beginning), 1995, 2005, and 2015. According to the Chang and Jackson sequential extraction, the P 2 K and P 1 treatments increased the inorganic soil P by 732 and 32 kg P ha −1 in 37 years, respectively, while the P 0 plots were depleted by –459 kg P ha −1 . The P 2 K treatment increased all four P fractions, extracted with NH 4 Cl (easily soluble), NH 4 F (Al‐P), NaOH (Fe‐P), and H 2 SO 4 (Ca‐P). Continuous depletion (P 0 ) decreased the NH 4 Cl‐P and NH 4 F‐P pools, NaOH‐P and H 2 SO 4 ‐P pools remaining stable. None of the P pools changed significantly at P 1 . The remarkable gap between the measured change and the balance for the P 2 K and P 1 treatments cannot be explained solely by lateral soil movement, meaning that a significant proportion of the applied P was lost either in surface runoff or transported below the investigated depth of 40 cm. Despite large P applications, the degree of P saturation reached only 20% in the P 2 K topsoil, assuming a 50% reactivity of Fe and Al oxides. As derived from sorption isotherms, a high EPC 0 ( i.e ., equilibrium P concentration at zero net P sorption or desorption) of 1.30 mg L −1 had been built up in the P 2 K treatment, while in the P 1 treatment EPC 0 (0.33 mg L −1 ) had remained unchanged and P depletion (P 0 ) had caused a decrease to 0.12 mg L −1 . These results demonstrate that P sorption and desorption properties respond strongly to both P fertilization and null fertilization treatments and that in a long‐term field experiment only a low proportion of the residual fertilizer P can be recovered from soil.