• Citer

A win–win combination in plant disease prediction: field data to update model estimations, estimations to drive field data collection

par Benhamouche, Ouassim ;Rossini, Luca ;Bono Rossello, Nicolas ;Pezzutto, Matthias ;Turco, Silvia;Garone, Emanuele
Référence European journal of plant pathology
Publication Publié, 2025-12-01

Article révisé par les pairs

• Accès en ligne
• Détails
• Contenu
• Statistiques

Résumé :

Are mathematical models truly ready to predict plant epidemics as soon as they have been validated? The answer to this question is, generally, no. A key limitation lies in estimating the spatiotemporal dynamics of the disease, notably the time and location of pathogen introduction in cultivated fields. Model reliability is further reduced by biological and environmental variability and by unaccounted factors, which together decrease the accuracy of open-loop simulations and require data-based corrections. However, field data are often noisy, expensive to obtain, and provide only retrospective information. In other scientific fields, integrating models with measured data has greatly improved predictive accuracy. This study explores whether plant pathology can benefit from such integration by introducing an enhanced modelling framework that combines an epidemic model, a sensing model, an estimator that combines the two sources of information, and an optimisation process guiding data collection. The approach is demonstrated through the model-driven first detection of infected plants, a resource allocation problem aimed at maximising early detection efficiency. Results highlight the potential of estimators to improve prediction and optimise measurement strategies for more effective epidemic monitoring.