Meta-Analysis on Criteria and Forecasting Models for Potato Late Blight Pathosystem: Assessing Robustness and Temporal Consistency

Climate change, along with the pathogens adaptive potential, challenges the robustness of criteria, forecasting models, and decision support systems for late blight (Phytophthora infestans) control, the most destructive disease affecting potato crops worldwide. Under PRISMA criteria, this meta-analysis examined the criteria and forecasting models in potato late blight over the last 106 years in 25 countries. The evaluation groups a total of 271 trials in which 59 different models were used. The criteria and the forecasting models were categorized by three generation types (G1 to G3) based on their statistical methodology, and by three mechanism types based on their internal structure (Semi-Mechanistic, SM; Non-Mechanistic, NM; Mechanistic, M). For each one of these groups, the accuracy, fungicide reduction capacity, and temporal consistency were evaluated. The results indicated that Mechanistic models (integrate pathogen biological variables) outperform Non-Mechanistic models (only consider environmental variables). Therefore, the integration of pathogen life cycle dynamics in the context of climate variability is crucial to developing robust forecasting models. This study highlights the limitations of Non-Mechanistic models and underscores the need for forecasting models to be developed under criteria of ecological realism of plant-pathogen interaction and pathogens adaptive potential under climate change scenarios.