Selection of nonlinear models to describe nitrogen loss in stabilized fertilizers

Abstract

Brazil leads the world in coffee production. The use of new fertilizer technologies has notably boosted the increase in coffee productivity in Brazil. Urea is one of the main sources of conventional nitrogen used in Brazil and around the world, but it presents significant nitrogen losses. The pattern of nitrogen loss by volatilization presents a behavior similar to a sigmoid shape, showing a faster loss, followed by a gradual deceleration and tending to stability in later stages. The objective of this work was to compare the adjustment of the Logistic, Gompertz and von Bertalanffy nonlinear models on the accumulated nitrogen loss due to urea volatilization + NBPT applied to coffee plants. The data studied in this work correspond to the first application of urea + NBPT in the 2018/2019 harvest year, from an experiment carried out in Santo Antônio do Amparo-MG. The goodness-of-fit assessment was carried out based on the coefficient of determination, Akaike information criterion and residual standard deviation. The von Bertalanffy model was the one that presented the highest value of coefficient of determination and lowest values of Akaike information criterion and residual standard deviation. Therefore, this model presents a better fit quality. Therefore, the nonlinear von Bertalanffy model was the one that best adjusted the accumulated loss of nitrogen from urea + NBPT applied to the coffee plant.