Management and Dynamics of Potassium in a Humid Tropical Ultisol under a Rice‐Cowpea Rotation

Little is known about the role of K fertilization, stover management, and tillage methods on soil K availability as they affect rice ( Oryza sativa L.) and cowpea [ Vigina unguiculata (L.) Walp.] productivity on Ultisols of the humid tropics. The effects of five K rates (0‐120 kg K ha −1 ), returning or removing stover, and three tillage methods (no‐till, strip, and conventional) were evaluated during 12 crops of rice and cowpea grown for a 4‐yr period. Fertilizer K was applied to the first seven crops. The site was a recently cleared, 18‐yr‐old secondary forest in the Peruvian Amazon Basin. The soil was a fine‐loamy, siliceous, isohyperthermic Typic Paleudult. Soils samples were collected at each crop harvest to 90 em in 15‐cm increments. Potassium fertilizer always increased grain yields when stover was removed. Conversely, responses to K additions were seldom obtained when the stover was returned. The extractable K (Modified Olsen) critical level for both upland rice and cowpeas was calculated to be 0.10 cmol L −1 . Returning stover with no K fertilization maintained soil K concentrations above critical levels for both species up to the last crop of the rotation. Residual effects of fertilizer K were prolonged by returning the stover. When stover was returned, subsoil exchangeable K increased with increasing rate of K fertilization. Removal of stover resulted in greater increases in subsoil exchangeable K at the 40 kg K ha −1 rate than at 120 kg K ha −1 , apparently because the higher rate resulted in K fixation. Tillage methods did not affect crop yields.