Ability of different soil extraction methods to predict potassium release from soil in ley over three consecutive years

To avoid over‐fertilization of potassium (K) and thereby a mineral composition in the grass crop not optimal for animal health, estimation of K release from soil is important. The analytical methods should therefore predict the total K release. Furthermore, to minimize costs for the farmers they should provide information which remains valid over a period of several years. The relationship between different soil extraction procedures for K and K uptake in ley for three subsequent years after soil sampling was studied in 19 field experiments on a range of mineral soil types in Norway. Potassium determined with solutions that extracted exchangeable K or parts of exchangeable K (0.01 M CaCl₂, 0.5 M NaHCO₃, 1 M NH₄oAc, or ammonium acetate lactate) was significantly (p < 0.05) related to the K yield only in the 1st yr after soil sampling. Potassium extracted with boiling in 1 M or 2 M HNO₃ was significantly related to the K yield only in the 2nd and 3rd yr. Potassium extracted with cold 2 M HCl, boiling 0.1 M HNO₃ or 0.5 M HNO₃ was significantly related to the K yield in all 3 yr after soil sampling. Among these extractants, 0.1 M and 0.5 M HNO₃‐extractable K were better predictors of K uptake than 2 M HCl‐extractable K. These three extractants release some non‐exchangeable K in addition to exchangeable K. The fraction of 1 M HNO₃‐K extractable with 0.1 M HNO₃ varied from 4% to 45%, whereas from 15% to 78% of 1 M HNO₃‐K was extractable with 0.5 M HNO₃. Consequently, the more easily releasable fraction of K extracted by boiling with 1 M HNO₃ varied considerably between different sites.