Smart weeding: Ecologically guided precision management of arable plant biodiversity

Background

Current weed management on farms is typified by a ‘zero-tolerance’ approach to arable plant species (weeds), underpinned by significant herbicide use. Nevertheless, not all weeds are equally ‘bad’ or detrimental. Some provide benefits for agroecosystem health, including mutualistic or beneficial ecosystem services (e.g. nitrogen fixation, habitat for beneficial insects, green soil cover), while species rich weed communities are also often less competitive with the crop (1). Some weed species could therefore be tolerated at a lower abundance within diverse weed communities, contributing to on-farm biodiversity. The accelerating development of AI technologies for species-level weed detection and precision control, now means that engineering diverse, beneficial, and non-competitive weed communities is a realistic goal.

Objectives

This project integrates weed community ecology with recent technological innovation (e.g. AI weed recognition and precision weeding), to enable ecological weed management. Field-collected images will be analysed using existing weed recognition software to determine current capabilities for identification and quantification of weeds. Weed surveying and imposition of weed-free field plots will be undertaken on selected farmers’ fields and the Rothamsted long-term experimental sites, to map weed distribution and evaluate yield-biodiversity trade-offs. Simulations and plot-based trials will be used to determine selective weeding strategies that can optimise yield by removing detrimental weed species, while preserving acceptable levels of weed biodiversity.

Novelty & Timeliness

This approach seeks to combine AI-based technology with ecological insight, to transition away from ‘zero-tolerance’ and towards true ecological weed management (2). In doing so, this project aims to foster an ecologically balanced, biodiverse weed community that supports broader ecosystem services. This represents a new perspective for sustainable weed management, balancing productivity with agroecosystem health. The successful candidate will benefit from working alongside a strong research team with current projects in both weed recognition and fundamental weed ecology. The student will be part of collaborative, international partnership bringing together institutions in the UK and Denmark, and travel between these countries will be facilitated as part of the studentship. The candidate will receive training in both agroecological and technological approaches to weed identification and control, providing a sound basis for a career in agroecology and sustainable weed management.

References

1: https://doi.org/10.1111/wre.12310

2: https://doi.org/10.1016/j.agee.2023.108861