Qualitative modelling of factors influencing the development of Black rot, for the prediction of damage to bunches

Abstract

Vines are one of the most pesticide-intensive crops in France, and reducing their use is a major challenge for both the environment and human health. Vines are subject to several diseases, including black rot (BR). This disease is causing more and more damage in South-Western France, and modeling of this damage has not yet been developed. Developing an Injury Profile Simulator (IPSIM) model appears to be an interesting tool for achieving effective BR management that guarantees vineyard productivity while limiting the dependence of cropping systems on fungicides. Indeed, this approach has proved its effectiveness on other crops such as wheat and sunflower. Based on expert knowledge and literature, we have developed IPSIM-VIGNE-Black rot. This hierarchical qualitative model evaluates the severity of BR attack on bunches of a field for a given year. The BR severity is simulated according to the disease’s health history, the winegrower’s cultivation practices, and the field’s environment. IPSIM-VIGNE-Black rot is a multi-attribute model comprising 46 attributes, including 29 input attributes and 17 aggregate attributes. In terms of the disease’s health history, the input attributes are the history of BR contamination at different levels: field, set of fields, and vineyard. The winegrower’s cultivation practices are divided into two main parts: the effectiveness of treatments against BR and the effectiveness of prophylactic measures. The effectiveness of treatments is determined by the type of sprayer and products used. Prophylactic measures include all winegrower’s cultivation practices that affect the development of BR without treatments. Finally, the field’s environment corresponds to the presence of abandoned vines adjacent to the field and the drying characteristics of the field, such as the presence of wind. Model results were compared with independent observations collected in conventional and organic vineyards in South-Western France over 12 years. IPSIM-VIGNE-Black rot showed encouraging predictive quality, with an accuracy of 48%. This model is promising, especially as improvements in its construction are envisaged to increase its accuracy. This model will be integrated into an IPSIM-VIGNE model including the 3 main diseases, powdery and downy mildew, and BR. Eventually, it could be used by the profession as a tool for diagnosing and understanding damages to limit its consequences and the use of pesticides.