Phosphorus fertilization: A meta-analysis of 80 years of research in Finland

In order to integrate Finnish phosphorus (P) fertilization trials, we performed a meta-analysis on the relationship between P rates (6–100 kg ha−1) and the yield increase of agricultural crops, and studied the source of variation in yield responses (soil group, soil test P, soil pH, plant group, cereal species, productivity without P fertilization, and cultivation zone). Our database consisted of over 400 short- and long-term experiments conducted in Finland over 80 years, between 1927 and 2007. The crops studied were spring-sown barley, oats, spring and autumn-sown wheat and rye, potato, turnip, pea, and perennial grass mixtures. The experiments had been carried out all over the country, on clay, coarse-textured mineral and organic soils. The meta-analysis demonstrated that P fertilization significantly increased crop yields (by 11%) compared to the control (fertilized with nitrogen and potassium). However, the yield response to P fertilization was highly dependent on the soil group: on clay soils it was only 5%, while on coarse-textured mineral and organic soils it was 10% and 15%, respectively. On clay and organic soils, the lowest P rates (6–15 kg ha−1, mean 13 kg ha−1) were enough to gain the maximum yield increases of cereals, while coarse-textured mineral soils needed higher P rates (16–30 kg ha−1, mean 25 kg ha−1). In the case of perennial grass mixtures, maximum yield increases were obtained in all soil groups with the lowest P rates (6–15 kg ha−1, mean 13 kg ha−1). Yield responses to P fertilization were apparent on coarse-textured mineral and organic soils with low soil test P, measured by the acid ammonium acetate method (<10 mg PAc l−1, coarse-textured mineral; <8 mg PAc l−1, organic), or on soils with low productivity without P fertilization (<2000 kg ha−1). The responses were negligible or low on soils with medium or high soil test P (>10 mg PAc l−1, coarse-textured mineral soils; >8 mg PAc l−1, organic soils), or with medium (2000–3500 kg ha−1) or high (3500–5000 kg ha−1) productivity. On clay soils, P fertilization increased yields to the same extent regardless of soil test P level or productivity. Finally, neither soil pH (3.2–6.6) nor cultivation zone (with growth periods varying between 100 and 175 days) had any effect on yield responses in any soil group. For cereals and perennial grass mixtures, the P rates justified by the meta-analysis were only about half of the maximum values allowed by the third Finnish Agri-Environmental Program, or those applied in practice on Finnish livestock farms. Thus excessive P application can be further markedly reduced without measurable yield decrease.