Abstract

In recent decades, the Brazilian field crops industry has been expanding into the Cerrado region, which has poor sandy oxisoils. In addition to heavy lime and phosphorus (P) applications, potassium (K) requirements are also difficult to meet. The objectives of this long-term (2005/06-2014/15) study were to assess application methods, timing, and doses of K applied to notillage soybean-maize rotation systems, and generate information supporting the establishment of new criteria for K fertilization on light soils in Western Bahia. The results shared here refer to the soybean crop cultivated during the 2014/15 harvest season. The Research Findings experiment included eight treatments that were applied on plots throughout the nine years, as follows: non-fertilized control; P fertilized control; low, basal K dose; farmers’ practice (N-P-K, 2-15-20); high, basal K; high, top-dressed K; high, split K dose; and farmers’ practice and additional top-dressed K dose, with seasonal K doses of 0, 0, 60, 83, 120, 120, 120, and 203 kg K₂O ha^-1, respectively. All treatments, excluding the non-fertilized control, received a basal P dose of 62.3 (farmers’ practices) or 96 kg P₂O₅ ha^-1. Soybean yields from the controls varied between 750-900 kg ha^-1, whereas K-applied treatments yielded 3,300- 3,650 kg ha^-1, with no significant differences between application regimes or doses. It is concluded that K supply is essential for sustainable soybean production, as poor sandy oxisoils cannot meet soybean K demands. Degrading straw residues alone fails to support K crop requirements for high yields. When a high K dose was applied as basal, top-dress or split to two applications, K uptake remained constant at 60-70 kg K₂O ha^-1, K retrieval from the applied dose was less than 40-45 kg K₂O ha^-1, and the rest was wasted. One suggestion is to consider splitting K application when higher doses are used in order to benefit from higher pH, OM and K₂O soil content, and lower Al+H.