Perda acumulada de nitrogênio em fertilizantes aplicados ao cafeeiro: modelo von Bertalanffy com abordagem bayesiana e GAMLSS

Abstract

Brazil occupies a prominent position in the global coffee industry, being the world’s largest exporter and the second largest consumer of the beverage. To ensure high productivity levels, the application of fertilizers to the soil is essential. Urea, although widely used, presents an important agronomic limitation: high nitrogen losses in the form of ammonia when surfaceapplied under unfavorable climatic conditions. In this context, the general objective of this study was to use statistical approaches to model the dynamics of cumulative nitrogen loss from fertilizers applied to coffee plants. The first application consisted of comparing the performance of different nitrogen sources regarding ammonia volatilization losses using the nonlinear von Bertalanffy model under a Bayesian inference framework. The second application aimed to estimate centiles through GAMLSS modeling and use them to construct reference curves for cumulative nitrogen loss. The results showed that the adopted approaches were able to adequately describe the nonlinear dynamics of the process, as well as provide informative estimates for agronomic purposes and decision-making. Thus, the methods employed proved to be suitable for evaluating nitrogen losses associated with fertilizer application in coffee cultivation.