lexicon
FREN

How good are farmers’ weather stations for modelling the risk of potato late blight?

2025 · ACTA · eng

One objective of phytopathology research is to model the development of plant pathogens to help farmers make informed decisions about crop protection. Farmers may install their own weather stations in the field to provide more local input for these models. However, it is unclear whether the quality of data from farmers’ weather stations (FWS) is sufficient given the sensitivity of certain agrometeorological models. The objective of this study is to assess the uncertainty of FWS based on a 3-year field trial and to estimate its impact on potato late blight (PLB) modelling. A Guntz & Divoux model (GD) will be used to assess the risk of PLB using hourly weather data. The outputs of interest (e.g. the date of the third PLB generation) will be selected and an acceptable error will be assigned to them. To investigate the relevance of FWS, a straightforward approach involves running the GD model over three years of recorded weather data and comparing the simulations of FWS against a reference weather station. However, this approach is of limited interest due to the short historical record: only three seasons are available for analysis. In this study, we propose modelling the temporal structure of errors on FWS data to generate a larger dataset (e.g. 1000 samples) that simulates FWS behaviour. Therefore, a more statistically robust approach based on uncertainty analysis will be conducted. The generated FWS data will be used to feed the GD model. The distributions of the outputs of interest will be analysed in relation to the acceptable errors defined earlier. This analysis will determine whether the quality of FWS data is sufficient for use in PLB monitoring. The detailed methods and results will be presented at the congress.

JSON

Source : Plateforme R&D Agricole (ACTA) — CC BY-NC-SA 4.0. Usage non commercial uniquement.