Grapevine anthracnose caused by Elsinoë ampelina is a serious threat in many vineyards, and its control requires repeated application of fungicides, usually on a calendar basis. A better understanding of the pathogen life cycle would help growers manage anthracnose more safely and effectively. After conducting a systematic literature search of grape anthracnose, we used the retrieved information and data to develop a mechanistic model based on systems analysis. The model simulates production and maturation of primary inoculum, infection caused by both primary and secondary conidia, and lesion formation and production of secondary inoculum. The model was validated for its ability to predict first seasonal onset of anthracnose lesions by using 8 years of data collected at Auckland, New Zealand, and disease progress during the season by using 3 years of data collected at Frelighsburg, Canada. Overall, the model provided accurate predictions of infection occurrence, with 0.96 accuracy, 0.91 sensitivity, and 0.97 specificity. The model also showed good accuracy for predicting disease progress, with a concordance correlation coefficient between observed and predicted disease severities of 0.92, a root mean square error of 0.14, and a coefficient of residual mass of 0.06. Although the model failed to predict 10 of 110 real infection periods, these missed infections led to only mild disease symptoms. We therefore conclude that the model is reliable and can be used to reduce the costs of anthracnose management by improving the timing of fungicide applications.
Keywords: disease control and pest management; disease modeling; epidemiology; fungal pathogens; model evaluation; modeling; prediction of infection; systems analysis.