Logistic models to predict olive anthracnose under field conditions

Olive anthracnose (OA), caused by Colletotrichum spp., is the most important disease affecting olive fruit. Key elements of OA epidemiology are known, but no tools are available for predicting OA development in the orchard as influenced by agronomic factors and environmental conditions. In this work, a long-term dataset (covering 12 years from 2002 to 2013) on anthracnose incidence on olive fruits (OAI) representing 73 cases (13 locations with nine olive cultivars differing in OA susceptibility) was used to study the quantitative relationships between OAI and 84 weather variables (monthly average values, from January to December, of daily temperatures, relative humidities, and rain). The OAI in December was correlated with OAI in the previous December (Spearman correlation rho = 0.803; P < 0.001), the OA-susceptibility category (rho = 0.366; P = 0.003), and 14 weather variables measured between April and November (during olive maturation). Binary logistic models were developed for predicting conditions leading to OAI >0, 1, and 5%. OA-susceptibility category and some of the monthly weather variables (Tmax in April, Tmin in May, rain in October, and Tmax and Tmin in November) were included in the models, which had an overall accuracy of 81, 86, and 85% for OAI>0, 1, and 5%, respectively. Spring temperatures (during flowering and fruit set) predicted OAI>0%, whereas autumn temperatures and rain (during fruit ripening) supported the prediction of OAI>1% and >5%. The identification of factors associated with OAI will improve the ability to predict and control the disease.