Indicator kriging modeling epidemiology of common bean anthracnose

Abstract

The present work aims to characterize the spatial variability of common bean (Phaseolus vulgaris L.) anthracnose (Colletotrichum lindemuthianum (Sacc. & Magnus) Briosi & Cavara) using indicator kriging systems and nonlinear regression models. Seeds inoculated with a hydric restriction technique at −1 MPa were sowed in the center of plots as a pointwise inoculum source during a rainy and a drought season. The disease incidence was evaluated 26, 33, 40, 47, 54, and 61 days after sowing, in both considered seasons. Spacewise patterns of the disease were influenced by weather conditions and sprinkler irrigation. Based on spherical semivariogram models and kriging probability maps, the disease was disseminated around the primary inoculum source presenting strongly aggregated pattern in the rainy season. Disease presents higher aggressiveness in the drought period, spreading with secondary inoculum sources that became coalescent over time. Power model describes the increase of disease range variation over time in the rainy and drought seasons, with higher range values in the drought period.