Deciphering the complex impact of yellow rust on wheat health and nitrogen assessment through high-throughput phenotyping
Plant Phenomics
High-throughput in-field phenotyping has pioneered new pathways for evaluating crop stress, which, while traditionally centered on isolated stresses, often falls short in adequately addressing the intricate web of simultaneous stresses encountered in field conditions. Recent research has advanced in identifying and quantifying stresses using innovative methods like spectral data analysis and machine learning. Nevertheless, formidable challenges remain, particularly in the task of differentiating between similar symptoms and deciphering the complex interactions of multiple stresses. As a result, the current research gap lies in effectively discerning and managing the compounded effects of multiple stresses on crops.
In September 2023, Plant Phenomics published a research article entitled by “To What Extent Does Yellow Rust Infestation Affect Remotely Sensed Nitrogen Status? ”.
In this study, researchers employed a multi-sensor mobile platform to capture RGB and multispectral images of wheat under varying conditions of yellow rust and fertilization-fungicide treatment over two years. The SegVeg method, combining U-NET architecture with a pixel-wise classifier, was used to identify damaged and healthy leaf areas. The results showed that disease pressure varied between years, with the EfficientNetB2 and XGBoost models enabling precise segmentation of soil, green elements, and damaged regions. The SegVeg approach was particularly adept at identifying disease-induced damage and its effect on grain yield, with the proportion of damage correlating strongly with yield losses. Furthermore, the study uncovered that yellow rust's impact extends beyond the visibly damaged leaf areas to the reflectance of healthy portions as well, indicating a more pervasive disruption of the plant's nitrogen status. This was mitigated by focusing analysis on healthy leaf portions or adjusting for damage in the models. Reflectance measurements unveiled a significant impact of diseases on Bidirectional Reflectance Factors (BRFs), with the 680-nm wavelength being particularly telling, signaling discernible changes in plant health. The study also observed strong correlations between variations in BRFs and disease intensity. Furthermore, diseases influenced the green area's BRFs, affecting biophysical and biochemical properties and thus the plant's overall nitrogen status. This suggests that disease presence complicates nitrogen status assessments using spectral data. In the broader scope of nitrogen status analysis and modeling under varying treatments, the study found significant impacts on various nitrogen status variables, with the fungicide influencing some variables more than others. The inclusion of disease indicators, like the proportion of damage from RGB imagery, enhanced model performances, especially under high disease pressure.
In conclusion, the study underscores the complexities of accurately assessing crop health and nitrogen status in the presence of diseases like yellow rust. It not only underscores the critical necessity of integrating disease impact into reflectance-based decision support tools but also illuminates the promising potential of sophisticated, multi-sensor approaches to unravel the complex effects of multiple stresses on crops. The findings also point to the need for further research to refine these methods and understand the varying impacts of different diseases on plant health and nutrient dynamics.
###
References
Authors
Alexis Carlier1*, Sebastien Dandrifosse1, Benjamin Dumont2†, and Benoît Mercatoris1†
†These authors contributed equally to this work.
Affiliations
1Biosystems Dynamics and Exchanges, TERRA Teaching and Research Center, Gembloux Agro-Bio Tech, University of Liège, 5030 Gembloux, Belgium.
2Plant Sciences, TERRA Teaching and Research Center, Gembloux Agro-Bio Tech, University of Liège, 5030 Gembloux, Belgium.
About Alexis Carlier
He is currently a PhD at the Center for Biosystems Dynamics and Exchanges at the University of Liège.
Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.