Answers are grounded in this paper and cite their sources.
Web-based decision support systems (DSSs) are essential tools for precision crop protection, guiding farmers and advisors in implementing need-based control measures against pests, diseases, and weeds. These systems rely heavily on weather-driven models, where data accuracy and availability are critical. Leaf wetness is a key factor for infection and reproduction of many fungal plant pathogens, making it an important input in DSSs. However, the availability of leaf wetness data from weather services is variable, leading to the development of numerous estimation models without a universally accepted standard model. This study aimed to develop a robust mathematical model for estimating leaf wetness across diverse European climate zones and to integrate the model for enabling real-time leaf wetness estimates as inputs to web based DSSs. Hourly weather data, including leaf wetness, temperature, precipitation, relative humidity and wind speed were collected from automatic weather stations. Training data came from five Norwegian sites, while testing data covered 17 locations across Europe. Five machine learning based models (decision tree (DT), random forest, K-Nearest neighbour, multi-layer perception, long short-term memory (LSTM)) were trained and their performance compared with five existing empirical models (RH87, RH90, extRH, DPD, CART/SLD) from the literature. LSTM and DT achieved best performance, demonstrating strong robustness across climate zones. The LSTM model, required fewer data inputs and well suited for time-series data, was integrated into a Flask-based service for automatic use in forecasting models within the web-based IPM Decisions DSS platform, thereby enhancing the precision of this DSS.
No takes yet. Share an insight, caveat, or question.
A 2025 study studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: