Current Trends and Perspectives on Predictive Models for Mildew Diseases in Vineyards

Abstract

Environmental and economic costs demand a rapid transition to more sustainable farming systems, which are still heavily dependent on chemicals for crop protection. Despite their widespread application, powdery mildew (PM) and downy mildew (DM) continue to generate serious economic penalties for grape and wine production. To reduce these losses and minimize environmental impacts, it is important to predict infections with high confidence and accuracy, allowing timely and efficient intervention. This review provides an appraisal of the predictive tools for PM and DM in a vineyard, a specialized farming system characterized by high crop protection cost and increasing adoption of precision agriculture techniques. Different methodological approaches, from traditional mechanistic or statistic models to machine and deep learning, are outlined with their main features, potential, and constraints. Our analysis indicated that strategies are being continuously developed to achieve the required goals of ease of monitoring and timely prediction of diseases. We also discuss that scientific and technological advances (e.g., in weather data, omics, digital solutions, sensing devices, data science) still need to be fully harnessed, not only for modelling plant–pathogen interaction but also to develop novel, integrated, and robust predictive systems and related applied technologies. We conclude by identifying key challenges and perspectives for predictive modelling of phytopathogenic disease in vineyards.