The importance of developing an integrated data-driven modelling platform for herbicide resistance research: A review.

Herbicide-resistant weeds pose a global challenge, constraining agricultural practices worldwide. Despite efforts to establish an integrated, data-driven framework, understanding the varied risks of herbicide resistance (HR) across different agroecological zones remains elusive. This review paper advocates for an integrated approach that incorporates socioeconomic, environmental, adoption behavior, and physiological factors to uncover insights into HR drivers and develop tailored management strategies. HR not only escalates production costs but also necessitates alternative weed management tactics, highlighting the urgency for proactive environmental management to mitigate soil health degradation and biodiversity loss. While current initiatives prioritize integrated weed management (IWM) like crop rotation and herbicide mixtures, challenges persist in integrating socioeconomic factors into predictive models and promoting the universal adoption of sustainable practices. Advancements in big data analytics, spatial modeling, and remote sensing offer promising avenues for predicting and managing HR across landscapes. This study proposes a research framework to predict the emergence and management of HR in agri-food systems. Additionally, the study utilizes a novel text-mining technique to conduct a comprehensive literature review, highlighting gaps in the development of data-driven modeling platforms for predicting HR emergence. The text mining findings explored that while common terms like weeds, resistance, herbicides, crops, management, and control are prevalent, research often lacks focus on predictive data-driven approaches for HR. Therefore, urgent development of an integrated national-scale approach to predict HR emergence is imperative. Global cooperation is essential for sharing best practices, data, and responses to emerging resistance threats.