Modelagem de perdas de produtividade do cafeeiro relacionada à ferrugem

Abstract

Coffee rust disease (Hemileia vastatrix) can cause losses of 30% or even more in coffee (Coffea arabica L.) crops. However, such losses were often attributed to empirical observations and evaluated in isolated experiments, with no repetitions over time and with a mathematical and statistical approach that was not very well suitable to this area of knowledge. Therefore, the objective of this work was to quantify coffee productivity losses due to rust disease. The data collected from 19 experiments, which were carried out from 2014 to 2018, in coffee plantations, cultivar Catuaí Vermelho IAC 99 susceptible to rust aged between five and eight years, spacing of 3.8 x 0.6 m between plants and conducted according to the technical recommendations for the crop. The experiment designed in randomized blocks with four replicates, varying the number of treatments with fungicide formulations, at different doses, obtaining different disease intensities. The experimental plot consisted of 10 plants and the 6 central plants considered useful plot. Evaluation of rust incidence and leafing occurred monthly. The data obtained were integrated into Area Under the Disease Progression Curve for Rust Incidence (AUDPCRI) and Leafing (AUPCL). The harvest was performed between the months of May and July in all experiments, on each year evaluated. For the statistical analysis, the AUDPCRI, AUPCL and productivity variables were distributed in class intervals and submitted to the normality test. Linear Critical Point, Multiple Point, Integral and Exponential Nonlinear Models were fitted using the software “R” version 3.4.4. The choice of model adjustment followed the criteria of higher determination coefficient (R2 ) and lower Akaike Information Criterion (AIC). The Exponential Nonlinear Model was the one that obtained the best fit to describe the occasional loss of productivity as a function of the rust. The Exponential Nonlinear Critical of Model obtained a best fit to describe the productivity losses as a function of the disease in specific months. The Linear Model of Critical Point was the best described the occasional losses as a function of leafing.