High-throughput field phenotyping in cereals and implications in plant ecophysiology

Resumen

Global climate change effects on agroecosystems together with increasing world population is already threatening food security and endangering ecosystem stability. Meet global food demand with crops production under climate change scenario is the core challenge in plant research nowadays. Thus, there is an urgent need to better understand the underpinning mechanisms of plant acclimation to stress conditions contributing to obtain resilient crops. Also, it is essential to develop new methods in plant research that permit to better characterize non-destructively plant traits of interest. In this sense, the advance in plant phenotyping research by high throughput systems is key to overcome these challenges, while its verification in the field may clear doubts on its feasibility. To this aim, this thesis focused on wheat and secondarily on maize as study species as they make up the major staple crops worldwide. A large panoply of phenotyping methods was employed in these works, ranging from RGB and hyperspectral sensing to metabolomic characterization, besides of other more conventional traits. All research was performed with trials grown in the field and diverse stressor conditions representative of major constrains for plant growth and production were studied: water stress, nitrogen deficiency and disease stress. Our results demonstrated the great potential of leave-to-canopy color traits captured by RGB sensors for in-field phenotyping, as they were accurate and robust indicators of grain yield in wheat and maize under disease and nitrogen deficiency conditions and of leaf nitrogen concentration in maize. On the other hand, the characterization of the metabolome of wheat tissues contributed to elucidate the metabolic mechanisms triggered by water stress and their relationship with high yielding performance, providing some potential biomarkers for higher yields and stress adaptation. Spectroscopic studies in wheat highlighted that leaf dorsoventrality may affect more than water stress on the reflected spectrum and consequently the performance of the multispectral/hyperspectral approaches to assess yield or any other relevant phenotypic trait. Anatomy, pigments and water changes were responsible of reflectance differences and the existence of leaf-side-specific responses were discussed. Finally, the use of spectroscopy for the estimation of the metabolite profiles of wheat organs showed promising for many metabolites which could pave the way for a new generation phenotyping. We concluded that future phenotyping may benefit from these findings in both the low-cost and straightforward methods and the more complex and frontier technologies.