Comparison of Modified Urea Fertilizers and Estimation of Their Availability Coefficient Using Quadratic Models

Abstract

AbstractA procedure to compare N fertilizers based on comparing complete response curves simultaneously represented by a quadratic multifertilizer response model (MRM) is discussed. A modified version of the MRM, called fertilizer testing model (FTM), is used to test differences among response curves. The FTM describes the response curve to a fertilizer selected as a standard as well as the additional effect of test fertilizers over the response to the standard. In case of different response curves, an approximate test to decide whether the curves have equal limiting yield is introduced. Finally, a procedure to estimate the availability coefficient from quadratic response curves having equal limiting yield is derived.These procedures were employed to analyze six field experiments conducted over a period of 3 years in Korea using three varieties of rice (Oryza sativa L. ‘Tongil,’ ‘Jinheung,’ and ‘Milyang 23’) and three fertilizers (urea and two sulfur‐coated ureas, or SCU) applied at four or five levels of N. Two out of six experiments showed a significant effect from SCU‐A (25% release/7 days) beyond that of split‐applied urea, whereas SCU‐B (14% release/7 days) showed significant additional effects in three out of five experiments. However, in four of these six experiments, the limiting yield of the response curves of SCU was significantly higher than that of urea, precluding a valid estimation of the availability coefficient. Possibly, the proper combination of N dose and timed release of N from SCU in these experiments translated into grain yields unmatched by split applications of urea.